Home > POP / PPV / Redirect > Follow-along Campaigns

---

There Is No Spoon (58)

---

Hey all.

I've known about STM for a long time now and I have friends that have been part of the forum, whose information I've tried to soak like a sponge. I've been at affiliate marketing for the past 7 - 8 months and was attempting stuff with Clickbank and FB ads for maybe 3 - 4 months before that.

I don't have a lot of money so it took a long time to convince myself to pay for the membership, even though I know now (and knew before) that it's totally worth it. I'm super excited to be here and to absorb all this info and contribute to this common goal of ours, even if we're doing it individually.

I just joined a couple days ago and I've been devouring as much as the info and posts as I possibly can. Total goldmine here, guys. Well done .

The follow-alongs that are most on the mind are Jaybot's Boring Thread, which totally made me fall asleep. BECAUSE I just don't want to stop going through that Odyssey since it's LOADED with valuable info and I fell asleep reading and re-reading it. It seems to be the most recent and updated follow-along on here. I even printed out parts of it, annotated it, then shredded it to keep the sanctity of this awesome forum.

I've gone through diplomat's Starting from the beginning one recently too, which had posts that I really resonated with. I'm personally at a point right now where I'm generating about $300 revenue daily, but I'm sitting at between $50 - $250 loss, daily. I have one offer right now that if I ran that, alone, would bring in $200 rev, ~$40-50, profit daily. Then the rest of the loss comes from testing new offers and attempting to scale, which so far has been a failure of a mission.

I'm fighting myself mentally right now because I'm thinking of all of the ways how this money I'm losing this money right now could help me in other ways of my life, but I'm also trying to convince myself that it's an education cost, like going to school. Did that, and didn't get much out of it. I think this will be more valuable. At least that's what I'm telling myself.



Hoping to change that.

I'm telling myself that this is something I have to overcome, mentally. There is no spoon. I know that there is great potential in this and I just have to develop my process and find the right offers. I just haven't found it yet. Against all of my internal alarms, I keep spending, hoping that I'll find the offers that bite and then I can take off running with those. That's my hope.

A few years back, I made a choice to live my life in a way that would bring me the most happiness instead of putting my head down and grinding for the grind's sake. In doing that, I moved to Hawaii and started working on boats. Now I sail everyday for a day job and life is pretty freaking awesome. But expensive. I have a few side hustles to help out, but I'd love to be able to get affiliate marketing to a point where I can do that for the job, along with the other side hustles, and work the sailing job for fun.

I'm starting this in the hope of helping those in the same boat and, hopefully , being helped.

My process:

For new campaigns, I'm starting on Pop traffic on PropellerAds.
I make multiple identical bids, separating them by the different targetting.

Ex:
1) Mobile only, 3g traffic, 1/24, $0.50 SmartCPM
2) Mobile only, wifi traffic, 1/24 $0.50 SmartCPM

Then I take those campaigns and clone them with staggered bids, from the beginning.

3) Mobile only, 3g traffic, 1/24, $1.00 SmartCPM
4) Mobile only, wifi traffic, 1/24 $1.00 SmartCPM
5) Mobile only, 3g traffic, 1/24, $1.50 SmartCPM
6) Mobile only, wifi traffic, 1/24 $1.50 SmartCPM
7) Mobile only, 3g traffic, 1/24, $2.00 SmartCPM
8) Mobile only, wifi traffic, 1/24 $2.00 SmartCPM

And then finally I make corresponding WL campaigns for each of those and run them them when they have 2+ conversions each, profitably.

9) - 16) WL campaigns, initially paused until I find good placements and then I start them with placements. I do this to offset costs and hopefully make some profit on top of that, while finding good placements.

I tend to run these campaigns with lower daily budgets until the start to do something, $10 - $20, and I try to focus on sweeps SOI offers that have a payout of $1.20+.

My attempts at running CC offers have pretty much been fruitless. I've run SOI offers and offers that are easier to convert, then I take the most converting placements and make new campaigns, changing only the easily-converting offers to the CC offers.

The few CC conversions I've gotten have lulled me into a false sense of hope. I get one conversion, I increase the budget or make more campaigns and then I, so far, have ended up losing 3x - 5x+ in costs of whatever I've earned with the one or few previous conversions.


I'll officially start a new post with data and where I am in this journey soon. Thanks guys!

---

Awesome post man! Glad to have you on here!

I don't know much about pops and push so those experts can chime in with that stuff, but I wanted to give a massive thumbs in agreement on the "spending for education" theme. The digital media buying/copywriting nexus is by far the most valuable skill in the modern economy in my mind, and when you factor in the fact that success at it can potentially let you earn more than doctors and lawyers who paid half a million dollars and years on end for their education, I think it more than justifies spending money to learn, even if your at a loss early on (which everyone is at first).

It took me well into negative 5 figures and several months to learn natives (which are obviously a lot more pricey than pops though so don't let that scare you) but it was the best investment I've ever made ROI-wise.

Very much look forward to seeing your progress and keep up the great work man!

---

To be fair, I've been well into negative 5 figures on push and pops, so it's doable! Don't worry, unless you are as dumb as I am, you'll never let it get that bad.

Also, I love @jack_l and his posts. Always good insight. But one day, I will finally go into Native and become his greatest nemesis

---

Good post! It's also helpful to make some affiliate marketing friends so you can talk about all your ups and downs. It makes it easier mentally to handle the losses.

I think masterminds are quite helpful for this.

Also this almost made me spit out my food

Sent from my YAL-L21 using STM Forums mobile app

---

Lol sounds great to me man!

---

Welcome to the forum, you've got to right place! I see a lot of people, especially juniors in this industry, thinking that they can scale from $XX daily to $XXXX or more overnight - it doesn't work that way.
You have the right mindset, your work should be fulfilling and give you the life YOU want, and you're already on your way.
I wish you good luck! Keep up the hard work, you'll see results soon.

---

So stoked, thanks for all of the comments guys!

Thanks man! Makes me feel a lot better to read this . Totally agree on the doctor sentiment too. Sister dearest is an ER doc and she pretty much just got out of school at 34 years old. Yeesh.

I can definitely see the potential profit in this online world. Makes total sense.

I've seen his posts and your posts all over the place here. You guys are just littering this place with gold.

Also, I have no doubt you're smarter than me. Only thing I might be is a little more stubborn. I made a bet when I was a kid to see who could not eat pizza the longest. Fifteen years later, no pizza . Well, maybe twice .

Thanks for the reply, Jay. Your Follow-along is the big reason I started this Follow-along .

Looking forward to that followalong

Don't I know it, man. I definitely see these Follow-alongs as masterminds. Up to this point, I've taken this journey solo for the most part, other than talking to a few people here and there and reading the free stuff on the interweb. Having other people to bounce ideas off of makes a world difference. I don't think I know anyone in the physical world that is into AM. Bores the crap out of the girlfriend. Everyone I know now are surfers and beach bums . Glad to know you guys

Thank you, sir!

On to the meat and potatoes! Or tofu and potatoes for those veganoids out there.

Yesterday's Stats:

Profit:


The top green campaign is the one I've tried to scale to the other TS with no success. It's a WL campaign with data taken over the course of a couple months and only consists of profitable placements overall for that GEO.

A few of the other campaigns are new tests in different GEOs, all sweeps offers because that's what I'm focusing on right now.


Red Sea:


One of the campaigns in the top bottom Red tends to be one of the ones in the top green. That campaign has shown to be pretty volatile, but up to this point, it is profitable.

The bottom three red are one of the CC campaigns I was referring to in the original post. They all have had a conversion or two in the past few days. At the moment, those campaigns have one lander and three potential offers, just rotating through the different device types, each of which have gotten a conversion and not always in the flow of the offer.

iPhone lander > Samsung offer > conversion
iPhone lander > iPhone offer > conversion
Samsung lander > Samsung offer > conversion

So I'm running more traffic toward it, trying to narrow it down, as much as that red and loss kills me. I feel like CC offers can turn around that amount of loss quickly, so at the moment I'm just waiting and seeing.

Can't remember but I think diplomat was the one who posted this. This has totally been me lately:



The campaign setup from the original post points all of those with the identical targeting aside from the bids to one campaign in Binom. So in that original example, those 16 campaigns on PropellerAds point to 2 different campaigns on Binom, separated into Wifi campaign and 3G campaign. That way, if the bid makes a difference in how well a placement converts, I can see that reflected in the campaign ID-Placement drill down, but I can still see which zones are doing well when just looking at the placements.

Also, side note, I love Binom. The features are great and their customer service is amazing. Vladimir and Co. have helped me so much along the way, even up to setting up my new server when I got it. There was an issue with the OS, so they reinstalled it for me and then transferred over my tracker, on my time, which was the middle of the night for them. I always feel bad contacting them because Hawaiian time is the exact opposite of their time on the other side of the world. Regardless, they always seem to help me out. Couldn't be happier with Binom.

One of my AMs is awesome and helps me find landers, but I do have access to anstrex push. I had Adplexity mobile at one point, but I've found just as good landers on anstrex (I think) and they were always easier to clean for me. Whenever I used Adplexity to download .zip's, there was always some kind of problem that I couldn't fix just by removing code. I would actually have to change the code. Anstrex works great for me. I figure, don't mess with a good thing.

I don't have Adplexity anymore. I also have a decent stash of landers built up right now but if I find good ones, I try to add them to my arsenal.

Problems I'm running into or have run into:

1) Recently, I went through about 5 different campaigns because they were getting flagged by Kaspersky as Malware. Now I'm not a hacker and I definitely have a hard time writing code, but I can read it with a good amount of understanding. I checked and re-checked the code and couldn't find anything out of the ordinary, other than my base64 scripts to obfuscated it on the most basic level.

PropellerAds didn't like that, so they suspended all of my running campaigns pointing to that domain, which at the time, were all of them. So I switched it to a new one. *.xyz domains are cheap, so cost wasn't a problem. It's just annoying setting them up and going through the Cloudflare process and changing the nameservers.

Set it up again, got campaigns running again and everything was peachy. Then flagged and all were suspended again.

Changed it, Flagged, Suspended. Times Five.

What I ended up finding was I had recently started trying to scale that campaign, so I spread it to desktop and on another Traffic Source. Since then, the domains were flagged. I can only assume that since most of the campaigns I've run up that point were mobile, when I added a desktop campaign, someone's Kaspersky Antivirus flagged either the offer as Adware or my base64 code, flagging my domain.

Original flagged domain on VirusTotal:


Newest domain on Virus Total, only running it on Mobile and PropellerAds for the past few days:


2) I have a somewhat profitable campaign on PropellerAds, so I wanted to scale it to other networks, both pop and push. So far, I've only had luck on one other network's for a push campaign. All of the others ended badly for me, at least in the initial tests. Maybe 1-3 conversions and around 20x + loss.

Because of this, I'm testing new offers in different GEOs in the same vertical, so I can focus my goal instead of having my mind so spread out. Same vertical, different GEOs, focus on testing and optimization.

3) Zeropark. I'm terrified to start a campaign on there. I think I read a post by twinaxe saying that he ran an offer on there once and his AM gave him a bump, just to run Zeropark traffic because the quality was so good.

My first experience a few months back with Zeropark was losing $80 in 20 minutes. First big mistake. No source or target limit.

I want that good traffic, but I don't think I'm at the point where I can throw down that loss if it comes, which in my mind it probably will. If anyone knows of a Zeropark case study or Follow-along, that would be awesome . I imagine the process isn't too terribly different than any other network. But I'd love to see a success story on that, even if it's just to mentally re-assure me.

---

A new day. Yikes.

Not any better today. There's something soothing about having some green on the board, even if the overall numbers aren't actually better. At least I had a single campaign yesterday that had $30 in profit. Man, this doesn't feel good.

Anyway, here they are.

Most profitable:


Red:


Pretty much no green. My main offer hit a snag.

Every day, I've tried to focus on ramping up traffic to one new offer to test. Yesterday, that was the least profitable campaign, the -$60 one.

Bright side? At least one conversion

Shit.

I want to be as transparent as I can with this follow-along, without jeopardizing my offers. Ideally, I don't want people snagging this offer out from under me since it's the only offer keeping my AM journey alive right now, but I know I'll get more out of this, the more info I give because there's more insight you guys will have. And hopefully, people can learn from my disastrous mistakes.

And I'd be lying if I said that when I saw a couple follow-alongs on here, that I didn't start a campaign similar to the one in the follow-along, to see if I could replicate the success. Or improve on its potential success . That was more the goal than making money with it. Right now, I want to systematize (is that a word?) and perfectamitize (definitely a word) a process.

Haha, anyway I doubt anyone will want to copy this "success".

I put a message into my clickdealer AM to see if my offer is on its way out and waiting to hear from him. Whatever it is, my WL campaigns suddenly stopped converting. I opened up the WL campaign on PropellerAds and it went to this:



That seems like that would be a great amount of traffic, so I don't think that's the issue. This offer has a $1.50+ payout and I had SmartCPM WL campaigns with bids in the range of $1 - $4 cpm. I tested a campaign last night on those whitelists at $6 and ended up at 3 conversions,

leads: 3
rev: $4.92
cost: $31.08
(no)profit: -$26.16

At the same time, I had a CPAGoal 2.0 that I started a while ago that wasn't spending much of its budget until yesterday, when it started spending its budget on its own. It brought in a few conversions but it was also unprofitable at around -70%.

I'm going to burn some more money on it, raise my bid with a lower budget and I'll report back on that. Thought process is higher bid, higher quality clicks, more conversions.

Its recommended bid is ~$23 cpm, which was ~$14 cpm not too long ago. At $23, it would have to have at least 15 conversions per 1000 , 1.5% cvr just to break even. Up to this point, it's been something like 0.1% cvr on targeted pop traffic. Seems like a stretch, even if quality is better. It would need to be 15x higher quality.

My guess is that ZA traffic started getting more expensive for some reason, so my bids that work working well, started underbidding and I lost my good converting placements and clicks because of it.

I was teetering a line before and there wasn't much of a profitable margin. I feel like this is the breaking point to where I can't make this offer work.

Same story, different day.

I'm testing sweeps offers in different GEOs. Right now, aside from my main offer, I'm testing sweeps in HU, PT, JO, AU, TR, DE, UK, CZ, IT, PH, CY, and PS. I've had at least 2 conversions for each of these GEOs. I have also noticed similarities in CTR and conversions across the different GEOs in some of the same placements on PropellerAds. I've been adding global WL campaigns with these placements where I try to test new GEOs and offers, even if I've never run that GEO before.

I've narrowed down my WL using affiliatecase's rules. I found their website a few months ago and I was trying to use their method to go from SOI/Pin offers to CC offers. Looks like they posted here too. After running an SOI offer for a month and some change, I took the data from that and filtered it according to the rules in step 2.



Then, I tried to send it to a CC offer with the exact same targeting, using only the placements that matched those rules. I asked my CD AM for recommendations on CC offers and he gave me some options so I started sending traffic to those. Got a few conversions but the SOI offer still ended up being more profitable in the 3-day test that I did. But with that data and those placements, I was generating $40+ profit daily for a week or two with that SOI campaign.

Sometimes, just to make myself feel better, I filter my view to my profitable placements so I can look at a wall of green instead of sea of red. Been doing that a lot lately.

I've also been going through a bunch of older posts. I skimmed through the 40-day guide, which I pretty much did a 90-day scattered version of before I joined STM based off of scattered info online and youtube.

I also found vortex's placement killer excel sheet and then massiv biz's sheet and I spend a few hours incorporating that into my system. I don't understand statistics or probability at all, but the F probability function from the excel sheet seems to come up with great placements, however that blackbox comes up with its "Kill" or "Whitelist" answer.

Found it here.

I just want to say you guys are awesome here. Not only because of all of this valuable info on this forum, but how you help each other here too .

I'm thinking about modifying my strategy. All of these tests are draining my bank. I've noticed more and more that there are a handful of placements that show up on PropellerAds that get decent CTR and usually are among the zones to get conversions, whether the best zones or just converting ones.

I think when I make my tests from now on, I'll start these scouting campaigns only with those zones and if it bears any fruit, then start full fledged ones.

Any thoughts/comments welcome! As much as I'm afraid to lose more money, I will for a test if anyone has some ideas . Otherwise, I'm just going to keep on keeping on.

---

Great intro, great thread

In the very beginning it´s not really needed to run staggered bids right away.

When it doesn´t fly then you just lose multiple times.

One big mistake that many beginners make is that they want to save money and thus run too low bids.
Result is that they recive lower quality traffic so that tests often show bad results.

But to see if a campaign has potential one competitive bid is enough, when you see that it works you can still test different bids.
Important then is that you don´t too low so that you receive good quality traffic.

Yup, making money needs investments but with the right mindset and when you work smart to reduce the losses it´s really worth it.

Think longterm and not for some quick bucks.

Building a business tales time but running a good business also lasts for long time then.

Thanks, such reminders are always good when there´s so much bullshit and false promises out there.

Keep it realistic, be busy and work for your success and your chances too succeed will increase alot compared to a "I want to make 5 figures a day with low payout offers on pops in 30 days" mentality.

Disappointment level will also be a lot lower then

Same here, I am like an alien in my family and among my friends

The only affiliate people I met in real are people I know from internet before.

About your screenshots:

I see that you are running many different things at once.

Propeller pops, Propeller push, Richpush, Pushground, Popads, Adsterra, Zeropark, Adcash, Selfadvertiser, Advertizer.

You don´t need that many sources to run good campaigns, 2-3 big sources with good quality are enough.

Check my thread HERE, there the most traffic was from only 2 sources.

The more you run on a source the more data you receive.

Then all further tests there will become faster and cheaper = more effective.

It´s good to have more sources up your sleeve so that you can scale good campaigns better.

But when you use too many sources for testing already you risk that you learn a bit from each source but don´t really crack just one of them.

Yup, I met the Binom team several times already and they are really fantastic people.

Shout out to Wowa and Romanski Binomski

It doesn´t have to be real malitious code.

Sometimes even harmless stuff gets flagged very fast.

It can also depend on where you run campaigns.

For example I had normal quiz landers in a specific vertical banned again and again when I ran them in ZA.

Yes, domains are cheap so I always register several domains in bulk and add them to Cloudflare already and set the nameservers.

When then a domain gets flagged the next one is ready to use already and it only takes few minutes until campaigns cn continue.

Yes, ZP can be very good but you also can lose much money there because they have lots of volume.

Consider them rather as a source for scaling and not for testing.

To be honest but ZA traffic got very expensive few years ago already.

Don´t test on too many sources and don´t test with staggered bids from the beginning.

First test for potential, when it looks good try toget it running on higher volume and different bids on the original source.

When everything is working good test the winning funnel on other sources.

---

Definitely having technical difficulties. I think I just deleted my last post. If I broke some rule and a moderator deleted it, I'm sorry I re-posted it. Here it was:

Thanks for chiming in, twinaxe . You're definitely a master at your craft and I'm lucky to have your input!

That's definitely true. I definitely do end up losing more than I think I need to.

I think from some other posts, I saw that you recommended global WL and then testing RON with BL if something works out. That is, if I have a WL for that GEO and that Traffic Source already.

WL for quality, BL for quantity/scale.

Where do you recommended setting the bid for quality placements? For example on PropellerAds?

Also, how do you decide at what increment to stagger bids, once you find a good offer?

I sent a message in to my PropAds manager but I never heard back. Part of living in Hawaii and living on the other side of the world from everyone else.


I have done this time and time again. Pretty much every time I've tested something on US. It was also one of my first campaigns I started, sometime within the last year. I figured:

"I speak english. This will be easy!"

Start Campaign.

$20 dollars gone.

Hey, at least is wasn't $80 like my first attempt at Zeropark.

Anyway, didn't have proper funnel built up for that at all, but since then, most of the attempts I've made at US traffic, my instinct is to go low bid so that doesn't happen again.

Clearly, a bad idea.

Understood. I got the main offer turned into a private offer about a month ago, along with a bump and my AM asked if I could send it more volume. That's when I signed up to these other sources. I couldn't get any of them to work though, for whatever reason. Maybe I wasn't bidding high enough. Exact same funnel and targeting that was working on PropAds, but wasn't working anywhere else.

I tried CPA on Adsterra, Adcash and Advertizer, but they weren't working. Adsterra came up with one placement that had 10 conversions, so I set up a WL campaign for that one placement. It worked for a while, but that has since fizzled out too though. Advertizer was working out on a non-CPA BL campaign, but it's getting less profitable lately too.

AM got back to me and said the offer was still going strong but more people were advertising it, which probably explains part of the hike in traffic cost and my conversions going down, as well as my ROI decreasing (down to the negative lately).


Makes total sense!


Yeesh. Must be the price to pay for those ZA offers.



I think that's the way to go! Exactly what I did. Now I have 5 backups, ready to go.


It'll be a while I think before I dig into Zeropark.

Question on ZP's RBO optimization:

Do you like/use ZP's RBO automatic rules? It seems to me that it's less of an algorithm to pick and choose good placements, like how I understand PropAds' CPAGoal 2.0 works, but moreso an automation of placement optimization, which we tend to do manually usually.

For example:

say payout is $1 for sake of the example (although I seem to remember that there was a minimum when it came to ZP's RBO and I feel like it was more than $1)

target: smelly-feet spends more than 3x payout, pause smelly-feet
target: bad-hairdo gets conversion so increase bid to get more traffic

Considering how much traffic ZP tends to get, that could get really expensive before it starts to take off.

Also, is it pretty standard across POP traffic sources that higher bid == higher quality? I've done that one PropellerAds with some success, but enough to see that bid is proportional to quality.


I was surprised that I got it running in the first place. I had tested some US and other Tier 1 GEOs before I tested this ZA offer and I couldn't get anything to stick. They were pretty short tests though.

Somehow, I got lucky and I got this ZA offer to work. Looking at the cost of all of the traffic for PropAds' traffic estimator, I was surprised that ZA's average and max CPM cost was more than US's, at least at the time.


Golden nugget of the day . Maybe that'll help stop the bleeding. Thanks twinaxe .

Another Question/Scenario:

I had a campaign at one point that had a single placement at somewhere around $800 spend, ~$200 profit. This was over the course of a month.

When setting my filtering, I set it to look at all time data. At one point, the placement stopped converting well and starting getting negative ROI, I assume because it became a popular placement and the bid price shot up. It was still converting but it just had a negative ROI. When looking at the data for 3-days back, it was negative and had reached -3X payout, and I would have normally paused it at that point.

Letting it continue spending, I was losing roughly $5 per day. But when looking at the all time data, since it still looked profitable because it was green and it was still getting conversions. I eventually caught it, but had lost ~$50 of the profit, down to ~$150 profit from ~$200, by the time I did.

Up until this point, I always just looked at all time data; if it was profitable, run. If it was negative profit, cut. I started looking at all-time, 7-day, 3-day and 1-day after this, but that's super time consuming.

How far back do you usually look when considering blocking placements? Is it too aggressive to look at 1-day data? Do you factor in past history for individual placements when deciding when to cut them? For example, if they have performed well in the past, and you're cutting all other placements in that campaign by -3X, would you wait until -5X to see if this once performing placement bounces back?

For TheOptimizer, how does it handle overall profitable placements when they start to get negative ROI after a certain point? Does it log stats and compare profit against previous points in time?


TLDR


Where do you recommended setting the bid for quality placements? For example on PropellerAds?


Also, how do you decide at what increment to stagger bids, once you find a good offer?


Where do you recommended setting the bid for quality placements? For example on PropellerAds?


Do you like/use ZP's RBO automatic rules? It seems to me that it's less of an algorithm to pick and choose good placements, like how I understand PropAds' CPAGoal 2.0 works, but more so an automation of placement optimization, which we tend to do manually usually.


Also, is it pretty standard across POP traffic sources that higher bid == higher quality? It seems that with PropellerAds, bid is proportional to quality.


How far back do you usually look when considering blocking placements? Is it too aggressive to look at 1-day data? Do you factor in past history for individual placements when deciding when to cut them? For example, if they have performed well in the past, and you're cutting all other placements in that campaign by -3X, would you wait until -5X to see if this once performing placement bounces back?


For TheOptimizer, how does it handle overall profitable placements when they start to get negative ROI after a certain point? Does it log stats and compare profit against previous points in time?


Here's yesterday's stats:


Profit:


Red:


Some of my recent tests panned out (a bit) and I'm only down -$75 . Compared to yesterday's -$200, I'm happy. Plus there's some progress on potential new offers, maybe.

I woke up to 3 CC conversions, one in PH and two in TR. The PH one was a delayed conversion from push. I tried it once before a few months ago and I totally forgot that push kept sending clicks up to two weeks after I stopped the campaign. Guess this click was a lucky one. TR ones aren't huge payout and I really only started testing it because I found a translated lander for it, plus the traffic was cheap. I'm a little in the red for that offer, but only slightly and it wasn't optimized so I have some hopes for it.

I like PropellerAds and I usually do the bulk of my testing on there, both Pop and Push, but usually starting out with Pop.

Most of the meat of my campaigns happen when I'm sleeping because a lot of the GEOs I'm testing are on the other side of the world for me. I have a script that attempts to keep an eye on my campaigns while I'm sleeping, checking my Binom for placements and killing the bad ones, but obviously it's not working how I'd like right now.

I'd really love to get my IT SOI offer up and running because my AM told me that he could probably raise my cap on my preferred offer if I could run the same advertiser's IT offer too. I'm testing a giftbox lander (which I just learned may have come from the illustrious @twinaxe, so FREAKING thank you man) and a few survey landers. So far, only a couple conversions in more ad spend than I would like. I want to make that offer work though because it could help my out on my good offer.

I'm also running a Monetizer push sub on all of my landers, but no back button. I'm making ~$15 per day with that, but I mentally don't consider that just so I don't come to rely on it.

At one point, on my main offer, I split test Monetizer back button with '#' back button, to keep users on the page. What I found with that test was that '#'

CTR:
Monetizer BB: 2%
# BB: 3.5%

That extra 1.5% clickthrough turned into 1.5x the conversions over the monetizer bb landers, and that extra 1.5% way made up for a slightly larger monetizer revenue.

It may be an isolated case, as it seems a lot of people are throwing monetizer on their back button, but it worked out better for me this way in this case.

I'll test it out with other offers in the future. I also plan to look into ways to add the bb url to a parameter in the url, so I don't have to make separate dedicated landers for '#' and monetizer bb, or even bb redirect to other offers or landers. It's possible. I just don't have it set up yet.


That's totally ironic and weird to me. I would've thought you were raised in some secret AM community, where they were rearing you to be their leader, with you being the AM maestro that you are.

I'd also believe you were an alien, with such advanced knowledge.

Sent from my SM-G9550 using Tapatalk

---

Holy moly, your posts are novels

There are different ways to do it.

Testing on a good WL is mostly good because there you know that you buy quality and can eliminate a good bit of guesswork.

Nonetheless a good WL is also very good for scaling because there you don´t have to pay for non converting placements.

This alone saves some money already and you can also bid higher then for the good placements = more converting traffic.

BL are good because there you often can get more volume for lower bids.

Downside is that you also buy non converting traffic.

But one big advantage of BL campaigns is that you can find new placements there that you wouldn´t find in WL campaigns.

There is no general recommendation.

It depends on the placements itself.

Check here



These are all good converting placements in same campaign but the highest possible CPM bids to run on profit range from $0.39-$18.19

By dynamic rules

For example when original bid is $1 then the next higher bid could be $2, then $3 and when you see that the sweetspot is between 2 bids you can adjust accordingly.

But when the original bid is only $0.20 then I wouldn´t set the next bid $1 higher.
There I would rather do something like $0.40 or $0.50 and then continue in similar steps.

In the end there are no fixed values for it because it also depends on the original bid and the general traffic price for your campaign.

Just try to increase in steps that are in good relation to your original bid.

Yup

But don´t worry, testing US because it seems easy and bidding too low are very common mistakes.

Did you run the WL campaign also on CPA?

In the end it´s just a tool where you set metrics to pause placements when they are $x.xx in loss or where you can let it increase/decrease the bids.

I use TheOptimizer for such stuff but I only use it to pause placements or campaigns.

For different bids I run different campaigns.

Yes, the effect higher bid = higher quality is much bigger on pops compared to push.

Good placements can have bad days as well.

So when a good converting placement shows bad performance for a day keep an eye on it but keep it running to see if it gets better again.

When performance stays bad for 3 days or so in a row just pause it for a week or so and test it later again.

Important is to take care that the bad performance days don´t eat up the profits from the good days.

It´s also a difference between going from 50% ROI to -10% ROI or going from 50% ROI to -100% ROI.

Mostly I just set rules like "If placements is at 1 payout in loss in the last 90 days then pause it".

When I see big discrepancies I also check it manually and act accordingly.

You would laugh when you know how I started affiliate marketing.

In 2005 I had my first PC and internet access at home.

I was 24 years but I was never interested in any compuuter or internet stuff at all so I also had no knowledge about anything related to it.

Well, then i had my PC and internet and just said to myself that I want to make money online and so I did.

15 years later and I am still doing it

---

Haha, sorry man. I think I was just happy to get an answer from you. I'll tone it down.

Thanks man . The past few days, I've been testing new offers/campaigns with one campaign RON with BL at a somewhat higher bid for quality and one WL campaign at slightly higher than recommended bid with 2-5 placements that I've noticed most of my GEOs have had good metrics with, whether it was CTR or CVR. Those are my testing campaigns lately.


I feel like that would be a CPA campaign with how the bids in such a big range.

I noticed that on my Adsterra CPA campaign. Some bids were at $3.00, some were at $100.00. The one big placement that converted 10 times was at $30 CPM. I was fairly new to the network when I ran this CPA campaign, so when I made a separate CPM campaign with that one placement only and set the bid at $30, which would have still ended up profitable if it was converting like it was in the CPA campaign, my TS manager sent me a message on her day off, asking if I did that intentionally because the approval staff didn't want to approve it, thinking it was a mistake.

I set up another clone of that new campaign at $3.00 to see if it would get any traffic and it didn't, but the $30.00 brings in some leads. It's not quite break even. I set up another campaign at 1/2 that, $15.00 CPM and I'm going to test that one next.

At this stage, my main goal isn't so much to make the big bucks, but more find the process so that when I'm running campaigns down the line, I can have a checklist to loosely follow.

I want to see what works.

Eventually, I'd like to have a set up with a bunch of campaigns running so that if some don't work, the others cover the loss and then some. That's the goal right now .


Makes sense! I read somewhere on here that vortex recommended staggering at $1, so as not to start a bidding war with yourself, but it also seems like it could be overkill to have $1 staggering if, for example, the max bid from the traffic estimator is set at $2.

Gold .


Didn't, I didn't see that as an option on Adsterra, but maybe I'm missing something. CPA with BL was an option that I saw. I'll shoot a message over to my manager.

I did recently clone the CPA campaign and set the bid to 50% payout instead of 80%, like it was before and I'm running a test now.

Problem with this offer is that it has a dual payout; $1.60 for certain converting traffic and $4.00 for other. I've wondered if that messed with the CPA algorithm at all. I set the CPA bid in relation to the lower payout.


I think I remember reading somewhere on here higher bid == higher quality for pops, higher bid == more scale for push. I usually only set my quality to high on PropAds Push. One test I did a while back that grouped them all into the same campaign gave me stats of like:

High => 10 leads
Medium => 2 leads
Low => 0 leads

Again, one instance, but it kind of gave me a picture, along with what I saw people on here saying, quality setting makes a big difference. I haven't had a good enough converting campaign yet to merit testing the Medium and Low settings for scale again.

Roger. I have a placement or two that's doing that right now. I think I'll pause them and see if they turn around.


Got it. That would pretty much be the equivalent of an all-time filter for me. At this point.


And now you're the twinaxe .

Legend of twinaxe. That should be the next Netflix series.

Thanks for all of help and insights, man .



Me covered in red...




Here's yesterday's stats:

Profit:



Red:


~-$80

Got lucky with my Advertizer campaign. I really don't touch it any more and I just let it do its thing. Not a CPA one, normal CPC. For a while, I was testing higher bids, incrementing by $0.005. I had 3 campaigns set at incrementally higher bids, but they neither converted nor made profit and they just cut into the profit of these lower bid campaigns, so I stopped them. Right now, I have 5 campaigns on Advertizer pointing to that one Binom campaign, differing by bid. They started out with exact same placements and I removed placements as they reached -1X payout loss, so they have slightly different WLs.

The Richpush campaign has the auto-rules on it, but I also haven't touched it in a minute either. It's a WL campaign for all the placements that had 2+ conversions profitably in a different corresponding BL campaign. It stopped sending traffic by itself a while ago until yesterday when it jump-started itself.

When you see Megapush, Hilltop, RichPops and SelfAdvertiser on my dashboard, those campaigns are just old campaigns that are sending in delayed clicks. All previous attempts to scale my main offer but ended up losing ~20x payout with 0-5 conversions. Exact same lander+offer setup for what was working for me before on PropellerAds.

I started up three new Adsterra campaigns yesterday to test out a few of the mobile offers that got a conversion on PropellerAds. There wasn't a lot of traffic on PropellerAds for CY so I put it on Adsterra which seemed like it had a little more to test.

Another campaign was a CPA campaign testing a TR CC offer, but I may have jumped the gun on it. Got a few conversions on PropellerAds SmartCPM over the past week and it's still running there, but no conversions there since, at least as of yesterday and today. Testing the CPA campaign at 50% payout and I'll see where that goes.

The worst campaign from yesterday was my best campaign up until that point. PropellerAds ZA campaign was bringing in ~$20 profit daily, now down ~-$20. Shucks. Best converting placements just stopped converting, so I paused them for now. They were slowing down a day before too. Maybe in a week, I'll start them up to see if they've changed.

I think the bidding has changed since it seems like more people are pushing this offer, according to my AM, so if ZA is getting more competitive, I would think that bidding higher might fix the problem. With that in mind, I set up a CPAGoal 2.0 WL campaign with all of the previous profitable placements, setting the bid at 50% of the lower payout and will see if that helps. I'll report back.

Read a post by jaybot and convinced myself to try an iPhone offer in US again. With super narrow targeting, at least at first. At the moment, extremely narrow targeting with WL of 5 placements and bid of $1 and $4 on PropAds, each with total budget of 20.

2 leads so far. $0.82 profit.

Things actively running:

JO, CY, CZ, ZA, PS


Total things running now, tests and other:

JO, CY, CZ, UK, US, ZA, HU, PS, TR, DE, PH, FI, IT, AU, PT


I'm trying to stay focused but I want to have a plan, ready to go for if and when my main offer dies.


I'm thinking of putting all of the details of one of my new potential campaigns that I'm running right now, in case anyone wants to run it with me. I got a good chunk of the legwork done and so far, I've gotten 35 conversions. I would post data, landers and offers could maybe tackle like a mastermind.

So much for toning down my long posts :/. To be continued.

---

You’re testing so much right now, it’s awesome. I have no doubt of your future success.

I don’t need anymore information than that to say: Vodacom Pin Submit is the higher CPA. The rest of the carriers is the lower CPA. You’re best off separating it into two separate camps with strict targeting on 3G only vodacom. And another for the rest (3G+WiFi).

Just be prepared for when that offer dies out suddenly, as it always does. Until then, milk that motherfucker

---

haha man, this stuff is exhausting, isn't it?

You got me .

I actually did that from the beginning. Since CD lumped both offers into my one my private link, I've been directing all of my ZA traffic into the one link with explicit rules set in my tracker, Binom, to separate all Vodacom traffic. Weird thing is that even Wifi traffic is getting the Vodacom payout every once in a while and that's where the problem comes in.

I chalked it up maybe somehow the offer was picking up that some of the Wifi users use Vodacom in their cellular, non-wifi lives. I have no idea how the detection stuff works on the advertiser's end, so I just ran with that assumption. Not complaining that I'm getting a higher payout for a conversion when it technically should have been a lower payout.

I also brought it up to the AM and he pretty much gave me an electronic shrug and said, "Hmm. Better payout for you. Nice." So, okay.

After some other offer tests, I found a different Vodacom offer that converted better than this one in my private offer, so I send all vodacom traffic to to that offer, then all other traffic to my private link.

Do you think the benefits of a CPA campaign outweigh the benefits of these phantom higher-wifi-payouts enough that I should push my AM for some kind of separated link? Or restructuring of that link on their end? I really don't know how any of that works.

Game on

---

Hey Twinaxe

I wonder if you treat the zones on Push exactly the same?

Have a FANTASTIC weekend!

---

I went into hibernation for a bit, but I was still running campaigns. Not a whole lot different though.

I'm curious about that too .


I had one anomalous day where my campaign started acting like they did a few months ago and I got $50 profit along with my monetizer revenue jumping from the average ~$20 daily to ~$50, then it was gone the next day and every day after that, back to negative ~$100.44

ZA has gotten expensive on PropellerAds lately and that's been taking a huge toll on my campaigns.

It got to the point where my campaigns on PropellerAds weren't bringing more than 1000 impressions, when they were bringing in 50,000+ before. I thought it had something to do with my domains getting flagged too many times, so I contacted PropellerAds support to see if they had put some kind of lock on my account. They got back to me saying that I was bidding too low and showed me a screenshot of my measly $6 CPM bid versus their recommended $46 CPM bid.

I started a CPAGoal campaign for ZA and it spent $20 with no conversions and the same setup as the original converting campaigns. When I talked to my AM, he said that the Sales Manager saw my page on spy tools and didn't like my "Congratulations" copy, and asked me to change it. I changed it but since then, my campaign that brought in ~70+ conversions daily has brought in ~1 - 5 daily. Could be the copy change or maybe the advertiser is throttling me. I was running a separate offer alongside this ZA offer in the same campaign, split by rules, and it has the same EPC as before, so I don't think it's the quality of the traffic.

I also noticed there's roughly a 25% dropoff in clicks between my Binom tracker and what's showing up on my AM dashboard. Last I checked it was ~5400 LP clicks from my tracker and ~4000 clicks to my CD dashboard. I sent a message over to my AM at CD and contacted Binom Support, but no real answer yet.

I'm testing a CZ campaign that was bringing in ~$20 profit, but I think it's the reason I've been going through so many domains. It's getting my domains flagged as GSB Phishing or Social Engineering and have been running it on desktop and mobile, but I split test domains between the Deskop and Mobile and I think it's getting flagged because of the desktop. Went through 6 domains this last week.

Looks like Jaybot found Twinaxe's time machine and used it to look into the future of my campaigns. Because they're dying.

So right now, I'm pretty much at square one. I'm testing more sweeps offers in different GEOs, leaving my ZA campaign alive at low daily spend in case it decides to say hello again.

Testing Sweeps campaigns in:

ZA, CY, CZ, GR, JO, MX, FI, DE, PS, NL, TH

Using Giftbox, Spinner landers and Survey landers scraped from Adplexity and Anstrex along with others found by AM.

---

Sorry, somehow missed the thread.

More or less yes.

On Push I often have several/many different campaigns running for the same funnel/offer.

These campaigns share the same BL but the single placements itself can perform different in the different campaigns so I mostly don't check their performance on a per campaigns base but rather check the stats from all matching campaigns combined.

I hope it makes sense, if not let me know to explain it better.

Not lately, ZA got very expensive long time ago ready.

Few years ago ZA WA was one of the hottest geos to run because there were many very good converting real 1 clock offers with payouts up to $5 or so.
There was also huge volume and at this time cheap prices.

Then everyone and his dog started running stuff there and ZA got saturated.

When you want start there now you often have to compete with people who have campaigns there running for long time already and who secured spots at the good placements and can bid much higher.

Theoretically it could be both or maybe something else.

When you already received a notification about the copy and the changes in performance happened after the changes then it obviously seems that it could be the reason.

On the other hand it also could be that the y were not too happy with your copy or so that they shave more.

It's really hard to tell what the reason is but when you didn't change anything else on the funnel and the performance doesn't catch up again don't spend too much time finding the exact reason.

It surely can help when you can identify the issue so that you can solve it.

But sometimes it's better to just accept the fact that the campaign isn't working that good anymore and move on.

Maybe you can see the total hits on Binom but only unique or filtered hits on Clickdealer.

That could explain such a difference although 25% indeed sounds pretty high from tracker to CPA network.

---

I'm just happy that you stop in at all! Always good to virtually see you, Twin .

I know this is kind of getting into the nitty-gritty and probably too much to think about until I actually have a good funnel flow going, but if your differently-bidded campaigns have the same blacklists, what's the point in testing bids? Couldn't a badly converting placement in a lower bid campaign be the rockstar placement with a higher bid?

I've run into this problem that at $1 CPM that Zone-X doesn't convert profitably, but at $3 CPM, Zone-X is at 300% ROI. It's blacklisted in the lower campaign, and therefore blacklisted in the higher campaign, wouldn't you be losing out? Or is that just part of the acceptable 80/20?

I've been considering trying out TheOptimizer a lot, just so I can see how it handles these different campaigns. It doesn't take your campaigns like a black box and grow them, right? It is just a system for automating rules, like pausing and starting campaigns, right?

I've built up my own personal optimizer over the past few months and at the moment it only has functionality in PropellerAds and partial functionality in a couple others. If it sees a bad placement, then it pauses the placement. I have it set up to detect Pop or Push campaign and it dictates its rules based on the average payout of all of the offers in each Binom campaign and whether it's a Pop or Push campaign. It pauses based on 3X payout for Pop and 1X payout on Push and I recently added the Placement Killer's algorithm from here to include those killzones that it finds in the placement pausing, if there are additional zones that the algorithm picks up. I also have been trying to do the calculus behind the multi-variate bayesian testing so I can incorporate that to auto-pause badly performing landers and offers during initial testing, but that's a mess on its own. I've been able program it to do up to 4 different choices max. Eventually, I want it to do as many as exist in the campaign. I'm putting it on hold because it's not the priority.

In its eyes, each campaign on the traffic source is its own individual campaign, but it can pull data from the tracker's overall campaign and separate each traffic source campaign by drilling down the different levels, then optimize from there.

I've kind of put continuing that entire project on hold while I try to manually optimize different campaigns, just to see if I can get them working. After all, how's my code supposed to do it if I can't?

I was checking the changes manually to make sure the code was doing what I would do, as if I were the one making the changes.

It was kind of trying to prepare that future day, when I have 60 campaigns running and it's just too much for me to handle myself. If it's positive, it'll keep it running. You mentioned it somewhere that if it's bringing in profit, you keep it running, even if it's €0.18 (paraphrasing ). Any profit is good profit in my mind, as long as the capital exists to fund the traffic.

I'm losing that battle. I think I originally hopped in during a fluctuation and the CPM pricing was just high. Now the cost is starting to be ultra-high.

They did ask me to keep going, just to change the copy. They approved the new copy too. It's not working as well, but if I was still paying for traffic what I was paying before, it still might have been doable.

I tried but I can't get the traffic for a sustainable price. I feel like people have to be running crypto or something on it now, which how much they're paying for that traffic. Super high payout offers, or something. $46 CPM? Those are push numbers, without the targeting of a push funnel.

That's exactly what I'm going to do.


A few of the offers I was promoting were flagging as adware, so maybe they're getting caught by adblockers. I've been testing Smart Meta Refreshing in binom, ever since my AM brought up the idea to cloak a little more so that my landers don't show up on spy tools. That seems to be slowing the offer redirect down a little bit, but I wonder if it's enough to drop off that much. I haven't gone back in the past and checked what the drop off rate was before. I'll check that and report back.


Thanks Twinaxe .




So with that, I'm checking out new offers.

With the recent thread on Sweeps offers being rigged, it's hard for me to convince myself to dig into more sweeps SOI offers.

A lot of the offers I'm testing right are Pin offers, MT/MO. Been testing those because contrary to what my first thought was, you can run them on Desktop too, even though they're mobile offers. So there's potential for more volume, but less likely to be shaved in my mind because the advertiser is getting data that's less likely to be fraudulent. SMS info, I imagine, is much more valuable than an email address, which are so easily created.

Anyone with some python knowledge could write a script that would create email accounts from behind different proxies and go to a page to fill in that email. I've never done that for CPA, but I may have done it for my sister to win an online popularity contest . I want to get AM going the right way, so that's not something I would do.

But, from that mindset, it would be much harder to and more expensive to have a bunch of different phone numbers and respond to different texts.

Point being, I would think MO/MT offers are safer in this rigged sweepstakes game.

I'll post screenshots on my next post when my computer isn't acting up. Yesterday, I ended around -~150 loss, trying to hold out for some hope on that ZA offer. Changed the color of red for that text because that bright red hurt my eyes too much. Maybe it was just the amount.

I let it run much longer than I should have, hoping it would change around and at least that I could glean some experience and knowledge from the ordeal. The lesson, don't let campaigns bleed you dry. Time will probably show that I haven't learned my lesson.

Jaybot, Twinaxe, want to check the future for me in your time machines to see if that's true?

I don't expect much better results in the very near future with the tests I'm going to be doing, but hopefully there will be some green in there.

My basic plan right now is:

- run SOI offers, ideally at some profit
- collect data for those GEOs, finding good placements and weeding out bad
- after enough data collected, swap out offers for higher payout offers like Pin or CC offers

Before my last good day with the ZA campaign, I tested CPM bidding on High quality traffic at the average profitable bid from my all-time data from that campaign and that campaign gave me the ~$50 profit day. Then, the following day and everyday since, everything dropped off to losses.

I've also been testing push CPAgoal2 on PropellerAds. I contacted support to get it enabled for my account. I never had luck with CPAgoal2 on Pop yet, but CPAgoal2 on Push seems to be breaking even or at a slight loss for a JO mobile content offer. Maybe it'll optimize after a little bit of time to profit.

Otherwise, I mainly run SmartCPM campaigns on Pop and CPC campaigns on Push. I get creatives and push copy from Anstrex Push.

I've gotten some conversions in DE with the mediamarkt voucher offers. A few on the pringles party ones too. If I can optimize those, that could be good, but I know I'm jumping from one difficult GEO to another.

I've gotten a few CC leads in MX and AR, some very scarce ones in PH and even scarcer ones in ZA. Push seems like it would be much better for CC offers, but AffiliateCase has a few posts on STM where they have had great success using POP.

Also running (testing and campaigns that alive at some level):

DE, CY, CZ, JO, HU, NL, NO, FI, GR, MX, AR, AU, PS, TR, (ZA, residual)

All Sweeps, all some kind of phone offer.

---

Yes, TheOptimizer is a set of rules that you configure + reporting. It will not magically improve your campaign. Lots of great options though. It can cut placements, but also increase / decrease bids at campaign or placement level.

For something more advanced, you can look at @kintura Machine Learning algos
https://kintura.com/docs/machine-learning

---

Hey man .

That could definitely come in handy, being able to step bids like that. I'm trying to wrap my head around how PropellerAds SmartCPM works, since their algorithm does bid-altering behind the scenes anyway. But in my mind, it seems like it works in a similar way to their CPM model, just because it won't bid higher than a threshhold you set. In my mind:

SCPM bid: $2.00
Actual bid: $1.50

So why not just set a CPM bid of $1.50 at high and leave it at that? Traffic costs fluctuate with time, so SmartCPM changes up the bids when they do. I guess I just answered my own question, maybe.

I've noticed in my campaigns, the SmartCPM bid rarely gets even close to the actual bid, unless my campaign is in a GEO with tons of traffic like US. For example, SmartCPM bid of $1 on a WL will have an actual bid of $1.
Otherwise, there's usually like a $0.30+ gap between those two bids. All depends on the targeting and the GEO, I assume.

Interesting! It's awesome that they created a system like that, but I wonder how Kintura actually applies its ML to some of these traffic networks though. It factors in all of those things, but the ability to manipulate a single campaign on those different levels doesn't exist on a lot of these traffic networks. I would think it would need the ability to create tons of different campaigns and then set them according to what it finds works the best.

I think RichAds/RichPush has microbidding where you can change up bidding for certain targets. PropellerAds is a little more limited. You can change bids for placements individually, or pause them altogether. I personally don't like changing up my campaigns' bids. I just make a new campaign. It keeps my data separate, but I guess SmartCPM is going to have different bids anyway, so that wouldn't make a difference. Maybe that's a reason I'm having problems?

I would think with how the customization of targeting is limited in many of the Pop and Push networks, TheOptimizer and Kintura would do a pretty similar job, even if Kintura's algorithms are potentially that much more advanced.

Haha for me, all I care about is green. I want green. It just comes down to how can I maximize my green. Or I guess right now, how can I minimize my red .

I don't even want to think about bid manipulation and optimizing on a placement level, especially on that level, aside from different bids on different campaigns, until I can get my campaigns green.

That's the 20% of my 80/20 rule.

Or am I underthinking this?

---

Kintura is focused on Native mostly if i remember correctly. I don't think they have push / pop networks integration.

As far as the gap between Smart Bid and real bid, there are lots of reasons, but i don't feel motivated enough to explain in depth. You can search for "Real Time Biding algos" in Google and you will lots of research papers explaining. But that is not necessary to run campaigns ;-)

---

Would love to spread to native at some point, even if only to have another option for scale. For the time being, I want to focus on pop and push so I don't spread myself too thin. Obviously, I have room to grow there :/.

It looks like kintura supports propellerads at least, but I would guess with a more basic version of their advanced AI features.

I'll have to look into RTB at some point. I'm sure it couldn't hurt having a more thorough understanding of how it works. Thanks for the guidance .

Sent from Maui using Tapatalk

---

Yes It makes sense But a few questions come up.

How do you take care of that?

How does this impacts your decision making?

How do you do that? On most trackers you don't have the option to combine reports.

---

I'd like to take a shot at some of those questions and maybe Twinaxe can confirm or shoot my answers down.

Important is to take care that the bad performance days don´t eat up the profits from the good days.

I'm trying to figure this one out too. The only way I've thought to do this is to take a look at the placement's performance from by time. For example, look at past performance in last 14-day, 7-day, 3-day, and 1-day and look at if the trend of that given placement is a positive one or a negative one. You could also set it to look at specific days in the past leading up to the current day (24-hour performance for: {7-days ago, 6 days ago, 5 days ago, 4 days ago, ... , yesterday, today})

If it's positive trend, meaning the profit/roi is increasing every day, then keep it running or give it more budget/bid so that you can get more good traffic from that placement.

If the trend is staying the same and it has a positve profit/roi, then just hold it steady and keep it going.

If it's negative, cut it or pause it temporarily and re-start it later to see if it has gotten better.

But doing that for each placement is pretty ridiculous without having some kind of automated system in place to do it for you. And it still might be overkill even with some kind of system.


It´s also a difference between going from 50% ROI to -10% ROI or going from 50% ROI to -100% ROI.


I think he had mentioned pausing the badly-performing placements temporarily, especially if they had be good before that. You can then restart them sometime later, like a week, to see if they're good again or still bad.

So, maybe just keep an eye on the campaigns so it doesn't get to that point?



These campaigns share the same BL but the single placements itself can perform different in the different campaigns so I mostly don't check their performance on a per campaigns base but rather check the stats from all matching campaigns combined.


I'm guessing he points his campaigns with from a given traffic source to a single Binom campaign, and then separates via the drilldown function from there. In Binom, you can split a given campaign by its campaign id, just how you can split a campaign by its zones.

From what I understand, he combines all of the campaigns with the same targeting aside from bid into one tracking campaign. I've tried to model mine that way too.

I feel like I understand what he's saying there, but then I'm just curious about those good placements that perform well at higher bids, that don't perform well at the lower bids. How would you change your setup to allow for those good placements in the higher bid campaigns?

In my auto-optimizer, I don't have it set up yet, but I intend to put in some kind of code that if there's a badly-performing placement in 30% of the campaigns pointing to a binom tracking campaign, then include that in the comprehensive blacklist that each of those campaigns share and block. Then, it'll check on a campaign level if it has been profitable in any of the campaigns and if it has, exclude it from the blacklist for those specific campaigns only. So it'll only allow that placement on campaigns where it is profitable.

If it does prove to be a bad placement for specific campaigns eventually, based on the normal cutting guidelines (1X Push or 3X for Pop), then it'll cut it when it hits that mark.

---

[QUOTE]if your differently-bidded campaigns have the same blacklists, what's the point in testing bids?/QUOTE]

In the reply I was talking about push campaigns because that´s what roiter123 asked for.

In some points pops and push are very similar but in others they are pretty different.

For both it´s same that a higher bid can result in more volume and can also open up more and better quality placements.

But on pops the bid has also a bigger impact on the bid position and that can result in other differences in performance.

So yes, there can be rather big differences in performance between the different bids but especially in the beginning it´s better to work as effective as possible and to keep things as simple as possible.

This means to rather start with global BL instead of specific BLs for every geo you run.

Same for bids, when you combine stats from several "same" campaigns you can grow the lists much faster compared to building lists for every single campaign separated.

Of course a BL usually is better the more specific it is but it also takes more time, more tests, more budget to build such lists.

Not really, it doesn´t happen often that a placement that doesn´t convert at all on one bid suddenly takes off on another bid.

It happens all the time that placements convert different on different bids.

Sometimes it happens that you have placements that convert very good but are in loss on a higher bid.
Then a lower bid could help to bring them into profit when they still convert good enough there.

On the other hand you can have placements that convert on a lower bid but not good enough to be profitable.
Then a higher bid could help to increase the quality for a better performance.

And then there are placements that are converting good already on a lower bid and convert even better and receive more volume on a higher bid.

They all have in common that a placement should convert at least somehow, then it´s worth it to test it on other bids as well.

But when a placement doesn´t convert at all then it probably also wont fly on a different bid.

See above, such stuff can happen.

But as you said, the placement also converts at $1 but not enough to be profitable.

You have good questions

I think in the end everyone has to find it´s own way how to deal with it.

For me it´s important to work effective and with a reduced risk to lose money in the beginning.

So I rather recommend to create lists with bad placements as fast as possible, even when you have to scarifice some more or less good placements then.

This approach give you results faster and it´s better to (maybe) cut 1-2 good placements to save money instead of trying to build very specific lists for each and every campaign where you have to spend much more time and often lose much more money to get (very) similar results.

There is no need to test bad placements again just to see if they perform different in the next campaign.
This will only cost you money again and again but will bring same results.

And you should keep in your mind, these lists are never final.

When you have a very good running campaign with a strong funnel then you can always re-test placements to see if they work in that campaign.

Setting it in relation to the mentioned 80/20 rule: When you receive 80% the same results with a rather general list that you would receive with specific lists then I would go with the 80% because it saves lots of time and money.

When I see good results then at least I know that it couls be worth it to spend some more dollars testing placements that could have potential.

Optimizer is basically only for pausing or restarting placements, campaigns and creatives (for push).

It doesn´t help with splittesting landers and offers.

To get the most use from Optimizer you should have a working funnel already.

This is very interesting.

I was thinking about doing something like this just for my very own campagns myself.

Basically as some kind of trainings project to start learning coding.

May I ask you what language you used for it?

Very good decision.

In the end you only can program stuf like this when you know how to run the stuff manually.

I sometimes have hundreds of campaigns running and as a solo affiliate without a team it wouldn´t be possible without a platform like Optimizer.

And sure, even if it´s only $0.10 profit per day, when it doesn´t need any manual work and no maintenance I just keep it running.

Profit is profit and enough of such small campaigns also add up.

Important is to always do a quick cost-benefit analysis.

Profit only $10/day but it takes big investment and lots of time to maintain it = it´s not worth it.
Profit only $1/day but absolutely no time needed to maintain it = keep it running, many of these campaigns and you have good money.

But it´s worth to mention that I don´t look for such small campaigns and I don´t actively try to find them.

Many of them are just abandoned campaigns that were doing much better before and then slowed down so that I just keep them running.

Small but converting volume is still better than just stopping them.

Sure, that´s also possible.

When offer domains get flagged and the users only see the red warning screen they mostly also don´t continue to the offer and won´t show up in the networks stats.

It won´t keep a good spyservice from still finding and indexing your landers.

For this you would need a real cloaker.

But really man, don´t waste too much time trying to hide from Adplexity or so.

Ask yourself, what do you expect from it?

Are you big enough to be considered as one of the key players on the spytools?

Or do you run unique landers that no one has and that are converting extremely good?

As long as you are just a normal affiliate running the same landers as everyone else (probably you got them from Adplexity yourself) you shouldn´t worry too much about that stuff.

On PropellerAds you can also micro bid ona per placement basis.

To be honest, I don´t have a real or strict approach for it.

When these are high volume placements that have big impact on the campaign I check it more in detail but apart from that I mostly just create rules in Optimizer like "IF Tracker ROI IS Less or equal THAN 10%" in the last 90 days then pause the placement.

Of course it can happen then that I have placements that were running at 1000% ROI (exaggerated) for few weeks and then suddenly they tank.

When I would stop them sooner I could save more of the profits but it would also mean lots of additional work to go through all the stuff.

With such a stop loss I at least avoid that such placements turn into negative.

And as long as the whole campaign isn´t affected that much by this one placement I just don´t care about it.

Basic explanation: When there is a "normal" difference in performance it can just happen, that´s normal and just by itself doesn´t tell much.

But when something very suddenly turns into -100% ROI = no performance at all anymore I would look closer to it because then probably something happened that should be investigated.

I can only talk for Binom and there you can.

Check my reply and when you still have questions about it just post them

---

I mostly run Push myself, but this is a very insightful post Thanks Twinaxe!

About cloaking, I wonder, whats considered to be big enough to have a justification of using a cloaker?

For example, on one of my offers I got the same landing page from Anstrex, but am I big enough? I don't know. I spend 400$ a day on average right now, on Push.

---

That´s something that everyone has to decide himself because there can be different reasons to use a cloaker.

You can use a cloaker even whitehat to filter out bad traffic and send cleaner traffic to to the offers.

Then you can use a cloaker to hide from Spytools or to avoid that your domains get flagged.

And then of course you can use a cloaker to hide from the trafficsources.

In the end it´s up to you, most important is that you already earn enough to justify the costs for the cloaker.

When you want to run aggressive stuff or so that could give your campaigns a better performance so that you earn more money then it could be worth to think about a cloaker.

But using a cloaker just to use from spytools is not worth it in my opinion.

Exception is when you run real unique stuff or some black magic

But I get all my landers from spytools myself and it´s nothing special, they are already in the spytools anyway.

---

I'm looking forward to when twinaxe shares all his secret, secret theoptimizer rules

---

I recently slowed my ZA campaigns almost to a stop because of how they were dragging me down, but I left the few most profitable overall ones running on lower bid and they have been bringing in roughly $10 profit. They were running on a BL, not WL. That made a huge different because my WL had a recommended bid of $46, more than 10x of what I wanted my CPM bid to be.


Originally, when I cut a placement that had met it's kill point (3X loss for pops or 1X loss for push) that was the end of it. Blocked the placement and forgot about it.

Now I try to take into account if that placement had any conversions, even if it had 3X loss.

With my plan, how I want to transition from lower payout offers to higher payout offers, if a placement is converting well, but just maybe at a loss for the lower payout offer, it could be a great candidate for testing on a higher payout offer where the margin is better.



Haha, obviously I haven't asked enough questions yet to show youhow bad my questions can be




I've noticed this and I think this is one of the golden nuggets of the day. Sure, you can build good/bad, but I think its important to keep in mind that these placements are websites and the quality of traffic coming from them are going to completely depend on any changes to the websites. If things improve on a website, it could increase the quality of the traffic coming from the website for you.


I think this is very important too. I think vortex mentions this a lot as well, that a good chunk of your results come from a smaller percentage of your traffic. Same coin, different side says a lot of your work goes into optimizing a small chunk of your results.

You can spend your time tweaking that extra little bit of revenue out of your campaigns, or you can focus that time on your other campaigns that could bring 5x what you would get out of those little tweaks.

Of course, you would never know what does well unless you test it, so it's hard to tell a lot of times what could be worth spending time on and optimizing.


Python. All the way. Anything you can do on a computer, you can program python to do for you. I've had social media codes running and managing 15 accounts at a given point. I've had email sign-ups automated so that my sister could win an online contest amongst her friends.

I also have codes that clean 70% of landers for me, swapping out links, removing other people's push scripts, adding my own personalize scripts, changing *.html files to *.php on my server so that it can use Binom's lp cloaking script, minimizing scripts. I still have to go in and check stuff, but it makes the process so much easier for me.

Quick side note for other people, not Twinaxe, being able to bulk clean lander like that can be a problem itself because then you'll probably have all of these landers you want to test. Could end up spending more on testing landers than you need to, when it's probably somewhat apparent what's working well, just by looking at how many times you see the lander on Spy Tools. It's probably a good idea just to test different types of landers in the beginning and get a feel for what's working. Like Survey vs Giftbox vs Spinner. I think I've read that as a recommendation by twinaxe on some other thread. Then, maybe drill down deeper and test variations of the different versions of the specific types.

Anyway, there's so much you can do with it.

From the beginning of this AM journey, I knew that there could be a point where I could get a code to manage a good chunk of this stuff for me, so I tirelessly worked on it from the beginning alongside my campaigns.

I'm not a coder by any means either, anyone can do this.

There are people on here that you can tell know the ins and outs of programming (ahem, jeremie). If he saw my code, he would be disgusted. Commented lines all over the place. I'm kind of like a homeless person when it comes to coding. My virtual cardboard box thrown all over the place and super disorganized. It's a mess.

But it does, for the most part, do what I want it to do. And if it doesn't, there's a way to change/optimize it. Just gotta work at it, like working on the campaigns. And have Google up and ready to try to find the answer to the questions and problems that come up .


Hopefully it doesn't take too long .


Everything you say makes so much sense. It's just getting to the point where your advice matches up to my current situation. That's the hard part.

There have been so many times I've asked myself "What would Twinaxe do?"

It took me ~11 domains and a lot of heartache to learn that lesson. I didn't know that they were seeing that message until CTR dropped to nothing and my traffic campaigns were getting suspended.

Also, sidenote for anyone using Binom, you can add your domains to the dashboard to have them get checked if Google has flagged them. I definitely recommend doing that. Could save you a lot of money, if you're running a lander and you have no idea why your CTR just disappeared. I would think other trackers would have that feature, but you'd have to check.


I was thinking this, but I don't like telling my AM that I think he's wrong. I want his advice and input to keep flowing, so I usually just bite my tongue . All of my landers have come from spy tools for the most part anyway.

I do have a pringles giftbox lander that I modified from your giftbox lander in German that I was using to test some of those pringle SOI offer. Pringles pop out instead of phone. Got a couple conversions





See if you can find it on spy tools

In all honesty, Twinaxes OG giftbox lander probably converts better.



Quick question on this, partially off topic: Doesn't setting a bid on a placement basis on PropellerAds change it to fixed CPM? So if you were running a SmartCPM campaign at say $1, and you set the placement bid $2, then essentially that placement acts as if it were in a normal CPM campaign? This is specifically for PropellerAds.

I've asked my manager there, but I have probably 18 questions stacked up in our Skype conversation, waiting for him to answer because I'm on the other side of the world for him. He's asleep while I'm awake. I guess I can't blame him, but I wish I would get some answers because I do have PropellerAds Gold. I message their dashboard's chat support pretty often.



That's an answer to a question I had that's been bothering me for a while. Of course, optimizing to squeeze more out of a campaign is ideal, but there's a point where you try to drill down into the specifics so much that this all becomes too exhausting and tedious. I would be more than happy if at the pinnacle of my AM career, that my campaigns worked like how I imagine yours work, Twinaxe. And I think that is key info.

I imagine it would be much more beneficial to put the energy that you would have spent on that one problem placement on another offer or campaign that has potential to do better. Unless that placement is a make or break placement, of course.


I would think that if you got the lander from Anstrex, it probably wouldn't be worth it. It's going to be on some spytool somewhere. The only reason I would think you would want to cloak at that level is if you were hiding your link, specifically, from someone like traffic source or your advertiser, which I think hiding from your advertiser is a bad idea by the way since they're the ones who are ultimately paying you.

For a long time, I had been looking for this one DE mediamarkt lander because I originally got it but I didn't download all of the images on it; they were linked to another domain. That domain went down, so my lander stopped working right.

Eventually it popped back up on spytools again, either same affiliate or a new one who had actually cleaned it right and I was able to snag it again, this time the right way.




I'll work on him if you do



Update on my campaigns soon. Although, not a whole lot to update.

---

I'm so happy I came across your post. I'm in the exact same boat as you. I truly believe that anything worthwhile requires persistence, and a good mindset is the way to achieve that persistence. It's nice to know that other people are exactly where I am. Your post has served as fuel for my fire!

---

So overall, my campaigns are and have been in a loss, usually up to ~-$90 loss per day lately.

Yesterday Profit:



Yesterday Loss:


Says -$87 but it's more like ~-$107 because Adsterra doesn't pass costs back to the tracker.


Maybe I'm jumping around too much to different GEOs. Not collecting enough data on one specific GEO.

I haven't found an SOI or pin offer yet that seems to be good enough to be profitable for me. At least one that I've had any luck scaling.

I have a code that compares my affiliate network's database of offers against the current cost for pop traffic for each geo and spits out offers, ranking them per the amount of potential profit each offer could provide if all of the impressions converted at 0.1% CVR. I also give it limits like minimum lower payout of $1, and max upper payout of $9. I also have been setting it to look for more easily converting flows like SOI, so its easier to get conversion data from the tests.

Obviously, all of the impressions converting is not going to happen, but it seemed like a good way to visualize how much potential an offer could have if it's able to convert. I've been choosing from that pool for offers to test.

Often times, a bunch of those offers match up with the ones my AM recommends. Sometimes they don't and I've found an offer that converts about once every two days or so, but the payout is ~$9 and it's profitable so far. Having a hard time scaling it because there's a lack of traffic for that GEO. I set it up on PropellerAds, PopAds and PopCash, just to try to get more traffic. On most of my campaigns, I usually just try to test on PropellerAds.

I checked facebook and it has 1,000,000 potential impressions for that GEO, untargeted and if I hop over to facebook for traffic, that'll be the first offer I run. I'm sticking to Pop and Push for now though and I'll collect my $5 profit every two days :/. So far, it's converted best on Push. I put it on EvaDav Push too. Take what I can get right now.

Any new campaigns I've tested lately start out on a Global WL that I've put together of zones that have done well or converted across different GEOs. I also run a RON campaign at a bid lower than PropellerAds recommended bid and a lower daily budget, usually around $10 - $15, depending on the offers and the GEO.

Then, if those pan out, then I'll either increase the daily budget or maybe clone the campaigns with a slightly increased bid to get more traffic and data.

My goal right now is to consistently run CC submits, and a lot of those are for phone offers, so I'd like to stick to testing SOI offers, or at least phone Pin submits, so the transition is a smoother one.

I am testing a couple voucher offers, main one being in DE. I have survey-to-giftbox landers that have the worst CTR of all the different variations I test (giftbox, survey, spinner) but it has the highest conversion rate amongst them, or at least out of the landers I've tested so far. I'm trying to think if there's a good way to increase the CTR on that specific lander to boost its performance without sacrificing performance.

I may introduce some other survey landers that I've gotten recently. I've found that survey type landers have worked better for me in the past for desktop, and I'm running both a desktop campaign and a mobile campaign. Right now, the Desktop campaign is outperforming the mobile one. Those are also the two main campaigns I'm running on Adsterra too, and they're doing better on Adsterra than they are on PropellerAds for me.

I speak a little german, but I don't think I speak it well enough to try to change up the copy. Maybe enough to change up the questions and/or answers. I think I noticed a grammatical error in the german lander above, haha.

At the moment, I have its back button set so that clicking the back button just keeps them on the page. In the past, that's increased the CTR for me. I'm trying to think of what I could change that to increase the CTR or conversion rate. Possibly a redirect to another lander, maybe more aggressive? I've noticed that can bypass traffic sources' compliance problems. PropellerAds, or any other network for that matter, has never suspended one of my campaigns for an overly aggressive back button lander. They have for an overly aggressive main lander though.

I haven't had luck with changing the back button redirect to the offer itself in the past, but I may try that again.

The lander from the previous post is one I'm testing alongside the survey+giftbox lander, and it has had some success. I haven't reached a stat sig yet to decidedly choose one lander over the other, but the survey+giftbox lander seems to be doing better for now. Just not good enough to justify cutting the other lander. I'm also testing two different offers, but both are pretty much converting equally.

I'm running all of my GEOs on PropellerAds, both Pop and Push. I also have a few campaigns on Adsterra, and one or two on PopAds and PopCash.

Since stopping most of my ZA traffic, my monetizer revenue has gone from an average ~$18 profit daily down to ~$8 profit.

---

And your post fuels my fire! . Welcome to this thread . I hope you can learn from my mistakes, haha.

---

@bennimen @twinaxe

You keep talking about using a cloaker to hide a landing page. What about hiding the offer or the trafficsource I'm running it at? To try and decrease competition?

---

Might be something to think about if you're running high numbers and you're really trying to hide your process.

But one major thing to think about is the cost of a cloaker, if you're going to use a service. Trafficarmor, for example, is a popular one and their smallest plan costs $129 per month for 32,500 clicks. If you're running pop, that's 32,000 impressions. Push would make a lot more sense than pop.

But, you'd have to be running super high payout offers, aggressively, to make that anywhere near worth it. Their highest service on their website is 750,000 for $600 daily, so 20-30 CC submits will get you breaking even for just the cloaking service. Not saying you couldn't do it. There seem to be guys on here that are pulling in $XX,XXX per day, those bastards .

If you're not going to use a service, you're going to have to make your own system. Everything I've heard about that, it's nowhere near easy to upkeep. I think one of the biggest advantages to using a paid service is that they're are super diligent keeping up to date on different ip's you'd want to block like google's, facebook's and different sources like spy tools that you may want to hide your pages from. From what I've heard, you need a proprietary system, an entire team, or both. I've heard that some paid established services out there whose whole business is about keeping track of those ips, they slip up or miss one address and people using the cloaker still get exposed in one way or another.

Relatively speaking too, I feel like it's pretty common for those guys to add to their IP address pool.

My two cents, you don't want to try to do that yourself.

If cost didn't matter, I don't know of any disadvantage to using a cloaker. Maybe speed? But I feel like cost would factor in for most people.

---

No secret stuff there at all.

Really man, you probably would be surprised how basic and simple some of my setup is.

I guess I put way less work in quite some stuff than most of you

For me it´s a huge difference if a placements doesn´t convert at all or if it converts but it´s not profitable.

When it doesn´t convert at all it´s mostly just a bad placement but when it is converting and it´s just not profitable then it means that it could be profitable under different conditions.

About your plan, I can definitely follow your logic but would like to add that often higher payouts also mean a more complicated flow.

When a placement would show exactly the same performance on higher payout offers then maybe the higher payout could bring it into profit.

But often a more complicated flow results in lower performance so that the advanateg of the higher payouts can be gone then.

Exactly, it´s important to focus on tasks that can have a big impact on your performance.

Often it´s just not worth it to waste time on stuff that could only result in 3% better ROI when you then miss chances that could make you 10 times the money.

Yes, I guess a good part of it is pure experience and the only way to gain experience is to run enough campaigns and learn from it.

Thanks alot, was considering Python as my main choice anyway so probably I wasn´t completely wrong with it

Learning by doing, the more you test the faster you learn.

And of course a good understanding of all the stuff can help as well

I really feel flattered by your kind words but do yourself a favor and don´t focus too much on how others would do it.

It´s absolutely understandable that everyone who starts something new needs some kind of orientation.

The problem is that probably every longtime affiliate has it´s very own workflow and workstyle.

I had more than 15 years already to develop my style and although this workstyle works very good for me it can be a total fail for others.

In the end I even would strongly recommend against some of the stuff how I run it because it´s deep connected to my own personality and my own mindset.

That´s why I always try to not explain how I run stuff or how I do it but I rather try to explain the system itself.
When you then understand it it´s much easier to find a workstyle that matches best to your very own personality and that works best just for yourself

Some things like techincal stuff or so are the same for everyone of us but when it comes to doing research, testing, optimizing and scaling there can be huge differences depending on who you ask.

10 experts, 10 opinions

When you then mix stuff from different peoples workstyles together it could result in a complete fail although all workstyles on their own can be very successful for the affiliates.

Yup, that´s a very good feature.

Yes, it changes to fixed CPM then.

But this is no problem because when you want to set a specific bid for that placement it means that you already know how it performs.

In such case fixed CPM is very good for scaling.

It´s basically like running a WL campaign with fixed CPM for specific placements inside a BL campagn on SmartCPM.

Of course you try to get most of your campaign but it´s really bad to spend time on the wrong stuff.

Then it can either happen that you waste lots of time on stuff that doesn´t bring a big enough increase in performance to be worth it or you optimize your campaign to death so that you maybe have a higher ROI but next to no volume left.

I run 99% of my campaigns like this:

Create campaign, set targeting, run it and only optimization I do is cutting placements

I don´t check browsers, browserversions, device brands, OS versions and such stuff.

It´s just not worth it.

When you have to "optimize" your campaign on such a micro level then it´s mostly not worth it.

Can you post some more info about all the campaigns?

What trafficsources, what geos, what verticals and so on.

That way it´s easier to see if you spread yourself too thin or how you could improve your campaigns.

Ask @jaybot about CC submits.

I would recommend to stick with easier flows for some time

Btw, what CPA networks do you use?

This is probably one of the main reasons to use a cloaker and it definitely can give you an advantage over other affiliates.

But also keep in your mind that you can lose your account in case you still get caught so you should be 100% confident that it´s worth it.

Apart from that there are several trafficsources where you can run aggressive and even very aggressive stuff when you ask them.
Some sources have special feeds for such aggressive stuff.

Often you need some history with them, spend already a good amount there.

But then it´s possible to run that stuff even without breaking the trafficsources rules.

Don´t go for cloakers who bill you on a per click base or so.

As you said, it´s idiotic for pops where you run on high volumes.

Apart from that there are very good cloakers with a fixed price like $200/month for unlimited volume so I don´t see a real benefit from using cloakers with limited clicks anyway.

I built my own cloaker in pure PHP many years ago, I think I started with it in 2009 or 2010.

It was pretty basic but it worked very good at these times.

Nowadays it wouldn´t work that good anymore because there are more variables to consider and the other side also improved their technical knowledge

---

Fucking CC submits. They convert and they're fun when they do. Too damn volatile in performance to run for me though. All it takes is one good/bad hour.

The best way to think of them is that, at scale, CC submits are the same as SOI, just way more expensive to test and run.

Not worth it unless you have a bunch of cash sitting around.

If you do have a bunch of cash sitting around, and a killer funnel, then go for it

---

That's an important thing I really learned from testing. Just because it's not profitable on one bid, doesn't mean it can't be profitable on another. I definitely test those non-profitable converting placements again on different bids.


Make total sense. It's a lot easier for a visitor to fill an email address than to whip out the credit card and enter that. Not to mention that they would have to trust the website is going to bring them something in return (free iPhone, whatever).

My question is: how do you change your setup to accommodate for this kind of flow?

I would think one option is that cloaking can come in here, make more aggressive landers. Match the appearance of your lander to the offer website make it seem more trustworthy.

I've asked some of my AMs about this and they've shown me landers that look exactly the same as the ones for SOI offers.

Sometimes they have the price difference like instead of "cost to you is free", it might be "cost to you is $2"

In the past when I've split test these landers, exactly the same but one with free, I've noticed that I have gotten a worse CTR with the $2 one.

I've also noticed though that my conversion rate has historically been better for CC offers when using the $2 version, despite a lower CTR.

Less of a shock for those people who submit the card info when the offer page asks for it, maybe. Which would make sense to me.



Maybe at some point I can make a post about doing something like this. In the meantime, look up "Jupyter" and "Pandas". Juypter is an interface that makes running and handling python code a ton easier. Pandas is a library that handles databases and a bunch of info. Once you have data in pandas, you can set up rules and conditional statements to manipulate that data. My code is pretty much just a bunch of loops that collect data, manipulate it, separate it and then shoot some commands over to propellerads (or the other networks I currently have it set for).

Also, make sure you have your api key right and you double check it if you change your password on Binom . If your code calls the wrong api key too many times, it'll lock you out of binom for a while and that could be disastrous if you're running things. There have been a ton of times I've bothered the awesome binom support and there was one time I messaged them frantically trying to get access back to my tracker because my code locked me out, haha.


Can't blame me for wanting to emulate your success . But I get it. Everybody has their own strengths and weaknesses. At the end of the day, we do this to better our lives in some way, so you gotta do what works for you.


One of my best days I had with my ZA campaign, I had just made a WL camp with the best performing zones that had more than 50 conversions profitably. I stuck it in a fixed CPM campaign on high with an average bid of what my tracker showed. This camp was on propellerads. I wanted to see what would happen even though I read somewhere on STM that a lot of people stay away from Fixed CPM. It gave me $80 profit with that camp, alongside all of the old ones I had been running before it.

The next day it tanked and the cpm bid went up though.

I notice a handful of guys using push on here clone their campaigns daily because it seems to refresh the quality. I thought about doing that for this specific campaign because they both had the "high" quality targeting in common, but it wasn't worth it at that point to do it because the cost went up so much.

And I know the cost of ZA is super expensive .



That's an important thing for me to really internalize. I see some conversions and a glimmer of hope because of them, so I try to optimize the campaign to get it profitable, but a good chunk of the time, I just keep spending more time and money trying to get it to work.

One good that has come out of it at least for me is that it keeps my clickdealer payments on net-7 (payment every 7 days), which is pretty important to me. It would be a lot harder to sustain my campaigns if I was back on net-30, or payment every 30 days. There's certain level of momentum I'm trying to keep, but I realize it might not make financial sense to keep going like this.


Sure thing, I'll make a new post so this one isn't as busy.


I guess I probably should do that.


Main is Clickdealer. Love my AM over there. He's awesome. I also have a pretty good system set up there where I don't have to manually add offers to my tracker, they automatically add themselves. I'm all about the automation, haha .

I also have accounts at a handful of other networks that I rarely use, but I do try to find corresponding offers to ones that seem to have potential for the ones I test from Clickdealer. One big benefit is that I get my payments from one source and don't have to worry about checking that I got my affiliate commissions from a bunch of different places.

But I have an account with YepAds, Mobidea, MaxBounty, CPAGrip, AdMobo, I think I'm supposed to have one at Gotzha but they never sent me the info to my account. I bugged my Gotzha contact like twice a week for probably a month and half to see if he would get my account access, because he said he approved me. Maybe he just got annoyed with me. I stopped because Clickdealer had so many potential offers. I also joined Haka because of your other thread, but I'm really gungho on figuring out sweeps.

I started out in the beginning testing Sweeps made ~$150 profit with a $0.30 payout TH offer, then jumped over to dating to try, didn't have as much luck. Then tried crypto for a hot minute, got maybe 3 conversions worth $200 each, but I ended up at maybe $500 loss getting them. It's hard getting data for good placements when the payouts are so high and there are so few conversions to base the data off of.

I recently was split testing some phone CC offers, four different ones from CD and one from YepAds and I got a lottery conversion from the YepAds offer. I was testing it on my super WL of placements with 50+ conversions. Could have been a fluke lottery conversion or maybe the flow is just better from that lander to that offer. It's still running now, and I'm not running it at a super high CPM or daily budget.

I'm kind of thinking since CD is so big, a lot of the offers may be saturated and may not convert as well as some of the other, smaller networks.

I also wonder how much that applies to CC submits, referencing the recent Sweepstakes-rigged thread. I wouldn't think CC leads would get shaved nearly as often as SOI leads, even if only for the fact that there's actual value to having a credit card lead, versus just having email info that needs to be further monetized with SOI leads. Which is another reason why I wanted to head in the CC offers direction. It's not just the payout.

I really would prefer to stick to ClickDealer mainly though, if I can get their offers to work.


I didn't know this! I noticed from spy tools, a lot of people are running pretty aggressively with sweeps on EvaDav. Amazon and Google Logos all over the place. I had just stayed away from being that aggressive for the most part. I read a recent story how Amazon's coming after affiliates using their logos. There's one about facebook ads and cloaking lately too.


I remember reading that somewhere . I consider you a one-man team.

Thanks Twinaxe


Hah, contrary to what my daily losses might make you think, I don't have a ton of cash lying around.

How have you tested them? Do you usually use the same lander for SOI for your CC offers?

Starting from the most basic level, it makes sense people are more likely to enter easier info like email than CC info. For example's sake, for every 20 people who are willing to enter their email for a chance to win an iphone, there's 1 person willing to put in the CC info too.

So even at higher numbers, you're going to get many less CC leads per SOI leads, if split with the exact same quality traffic, exact same landing page. Unless you somehow change up the flow (landing page?) in a way that targets CC submits more. Right?

Or how do you account for the more complex flow? Garner more trust so that those impressions/clicks are more likely to enter CC info? Make the page more targeted to users, letting them know they're going to have to pay something in order to get their new iPhone?

I would think finding good placements is just a start because those are just websites that have been proven through your own tests and data to be websites whose users are more likely to engage with your own lander and offer. Once you find those, then you need to test different landers, specifically targeting for CC info.

I would think Desktop would have better CVR for CC offers over Mobile. And maybe Mobile Wifi better than Mobile 3G, just because you're not going to be entering your CC info on your phone in between texting your girlfriend while you're driving. Although I know you can just generalize and assume that.

---

So ROI is roughly the same as yesterday, but since I've been stopping some of my test campaigns, my losses aren't as high. I did get two CC leads today in ZA after replacing an offer that stopped converting for me.

I mainly test sweeps offers, trying to focus on phone sweeps specifically.

I had been testing offers in the past two weeks in these GEOs:

ZA, HU, JO, CY, CZ, PL, DE, PS, MX, AR, FI, NO, NL

These passed initial tests (optimization has started):

ZA, HU, JO, PS, NL

These were testing GEOs (not cutting placements yet; narrowing down landers and offers and waiting for more conversions):

SOI & PIN: CY, CZ, PL, DE, FI, NO
&
CC: MX, AR (Initially tested SOI and used converting placements to test CC offers)




ZA Sweeps DOI/Pin phone:

TS: PropAds Pop mainly, along with PropAds Push, ClickAdu, Advertizer, EvaDav, HillTop, Adsterra, Ad-Maven, AdCash, PushGround, SelfAdvertiser
These were the campaigns that I tried to scale to other networks when it was doing well. Scaling to other networks didn't work for me. At one point I tried to run my working flow on PropellerAds CPA, Adsterra Pops CPA, Adcash Pops CPA and Advertizer CPA, but the only one that turned out remotely profitable was Adsterra CPA.

That was only because I took placements from the CPA campaign that converted and made a normal CPM campaign out of it. It also stopped converting after a while. The Adsterra CPA campaign itself wasn't profitable.

Current Targeting:
1) Pop, Mobile, Android, Chrome, Wifi
2) Pop, Mobile, Android, Chrome, 3G

ZA has room to scale, but it's a nuthouse and expensive. Also the cause of most of my heartache lately. I ran it mainly on mobile campaigns. Tested some Desktop campaigns and that was the start of my domain flagging problems so, I stopped and hadn't tested those much since. I've also stopped the majority of these campaigns because the offer wasn't converting for me for a couple days and was losing $40 loss daily by the time I stopped it. I left the three most profitable campaigns and those are bringing in roughly $12 profit average in the past two days, but some of those are CC leads.


HU sweeps Click2SMS voucher:
TS: PropellerAds Pops
Targeting:
1) Pop, Mobile, Android, Chrome, Wifi
2) Pop, Mobile, Android, Chrome, 3G

Using the giftbox lander. Recently tested a newer version of the same offer at 100/20 and it seems to convert better, so I split the traffic 50/50

I need to stop this one though. One zone is profitable, the rest are not. I have way overspent on this one. This was a campaign I kept running because it was bringing in conversions, so I thought I'd eventually whittle out the bad placements to make it profitable but I have ~$180 loss on this campaign for a €1.2 payout offer. Bad. -37% ROI overall over the course of 1.5 months.

My lesson here:
Be diligent in checking overall campaign status.


JO Sweeps Pin phone:
TS: PropellerAds Pop and Push
Targeting:
1) Pop, Mobile, Android, Chrome, Arabic, Wifi (Stopped)
2) Pop, Mobile, Android, Chrome, Arabic, 3G
3) Push, Mobile, Android, Chrome, Arabic, Wifi
4) Push, Mobile, Android, Chrome, Arabic, 3G
5) PopCash-Pops, Mobile, Android, All Browsers, Arabic, 3G (All browsers because chrome mobile traffic didn't have enough volume)

Landers: 2x Giftbox, 2x Spinner, Survey > 5 landers total > narrowed to 1 lander now for pops and push

Been testing JO for around for a couple weeks, hovering break-even and loss. Started with Pop campaigns only. At first, it was converting well. I split test 5 different landers and by the time three of them had multiple conversions, the other two had zero, so I cut those. I didn't check it with any stat sig calculator though.

I initially set my bid at around $2.00 CPM, max on the traffic estimator was around $1.00 for 60,000 impressions. I was hoping to get data fast. For a zone that had a recommended bid of $0.40, a $2.00 seemed like a sweet deal, so I bid high, hoping to get good traffic to test the offers. I targeted the three cell services that the offers allowed from their descriptions.

This is an example, although it's changed a little bit since I first made the campaign:



When I made the Wifi campaign, I figured in my head that a good chunk of those wifi impressions used those cell services and if they converted, it would be worth it. If not, then that was the test.

Ran it on PropAds Pops with separate wifi and 3g campaigns and PropAds Push with separate wifi and 3g campaigns. Pops-wifi underperformed and I stopped it.

Retested landers for Push campaigns and found that a giftbox lander worked better for Push than the survey lander, which worked for better for Pops.

Also have a PopCash campaign, although I have no idea why the CTR is so high. It's the same survey lander that was converting for PropAds Pops. The PopCash campaign is targeting more than just chrome though. Maybe that's an effect of the other misc browsers. I usually target Chrome when testing my new campaigns.

Last 3 days:



Just stopped Push-wifi today because no conversions in last 3 days.



PS Sweeps SOI phone:
TS: PropellerAds Pop
Landers: 6 Initial (2 Spinners, 2 Surveys, Giftbox) > 3 Landers ( Spinner, Survey, Giftbox) > 1 (Survey)

PS has been around for a few months, small payout but profitable. Somewhere around ~$100 profit over its lifespan, total revenue ~$700, but the offer recently stopped and its replacement doesn't convert as well so it's probably going to be dropped. Lately, about $1 loss per day. This was a set it and forget about it campaign. Also, don't pay attention to the [15022020] date in my campaigns. I usually clone campaigns and forget to change that first term.




Testing:

NL Sweeps SOI phone:
TS: PropellerAds Pops
Landers: 1 (survey+giftbox) > 2 (survey+giftbox, spinner) > 1 (survey+giftbox) > 3 (survey+giftbox, 2 survey) 7/14

Started this test for an offer at $3.00 payout on a Global WL I had been using. Got 1 conversion in first 20 minutes, and 3 total that first day on WL campaign, so was $9 rev, ~$6 profit that first day.



2020-07-09 had a high CTR because I had gotten a new spinner lander that I split tested and it had time-dependent redirects built in that I didn't catch. After 3 minutes, it would auto-redirect to the offer. That messed up the stats.

After initial conversions on the WL, I started a RON campaign:

SmartCPM
Bid: $1.80
Targeting: Mobile, Android, Wifi
Daily: $20
Total: $40 (and then extended)
Landers:
Tested 3, (Survey+giftbox, survey, spinner) > Narrowed to 1 (survey+giftbox)
CTR isn't great with Survey+giftbox, but CVR is better than any of the other landers

Impression rate: 1/24
Traffic Estimator available Impressions ~200,000 for above targeting (although that's from checking the current campaign stats)

I generally try to target more broad when there's less than ~40,000 expected impressions for the initial tests. Maybe combine Wifi with 3G.

This campaign has roughly $20 loss daily. I've kept an eye on this one, but 3X payout is $9.00 so it's taking a bigger budget to whittle the placements down on this one. Only one zone so far cut because of 3X payout, and it was a zone that got a conversion. Also, so far, most of the conversions I have gotten for this one are ones from my Global WL zones, whether or not they're specifically in that initial WL camp or in the RON camp with BL and those zones are just converting well there.

Average ROI is -70% overall, -9% on global WL.

all-time screen shot of campaigns:



The campaign with 5 conversions is the initial Global WL. The two campaigns above it are RON campaigns with added BLs as placements are paused for exceeding 3X payout.
The campaigns with 1 conversion each are global WLs with different bids; one fixed CPM bid on high quality at $2.35 and the other a SmartCPM campaign with a bid at $2.60 testing one single placement, which was the placement that got 3 conversions that first day.

Usually, if I got these stats, my instinct would be to stop these campaigns and bring it to another traffic source to test if it would do okay over there. It got some conversions, so I would think that it could perform better on a different traffic source.

After reading through some posts on here, I've seen some Twinaxe recommend not to take campaigns/offers to another traffic source if it's not profitable on the original.

My understanding of this recommendation is:

1) I've already spent money getting data for this traffic source, so just jumping to another traffic source just gives a smaller amount of data on a bunch of sources vs getting closer to mastering one traffic source which would be more beneficial in the long run.

2) It hasn't performed well enough on PropellerAds, so why would it perform well on another source?

Thoughts on NL campaign:

• testing only two unproven landers may not have been enough. There could be another better performing lander. reintroduced 2 survey landers into the mix
• doesn't work well on pop (?)
• bidding not competitive enough

DE Sweeps SOI voucher:

TS: PropellerAds Pop, Adsterra Pop
Landers: 3 Landers (giftbox, survey, survey+giftbox)
1) Pop, Mobile, Android, Chrome, German, wifi+3g
2) Pop, Desktop, Android, Chrome, German, wifi+3g (I've found in past tests that 3g on desktop doesn't have a lot of volume, but has better conversion)

Obviously one of the most competitive GEOs. I had considered with how much I've been losing anyway lately, stopping most campaigns in other GEOs to focus on this one. Or at least slowing down on the other ones. There's a good amount of traffic so it could be scalable. That way I could at least focus on one, but I don't know that doing that would make much sense because it's not like I could try to optimize this one GEO while I'm waiting for it to get its traffic.

CTR is low on DE, usually somewhere around 0.4% CTR, average pops. I'm used to ~2% on Pops.

Desktop seems to work better than mobile for me.

Planning on testing on push, and possibly pausing Pop campaigns temporarily during. Push seems to be a more trustworthy flow and more likely to get results.



I mainly use PropellerAds, but if there doesn't seem like there's enough traffic, I've been using Adsterra. I also had some campaigns on PopAds and Popcash. I also have a residual campaigns in EvaDav and Pushground. Been trying to stick to a few networks per Twinaxe recommendations, so I can get more data from one network instead of a small amount of data from a bunch of networks. I've been leaving the ones that are already running though because they're at least somewhat profitable.

I cut badly performing placements at -3X payout for Pops, -1X payout for Push on a traffic source campaign level.

This post is too long as it is so I'm going to wrap it up.

FI and NO both sweeps offers, trying to feel out the traffic situation. There are good potential CC offers to transition to eventually, which is a big reason why I started testing these SOI offers. Not a lot of data/traffic for either, so I've just been letting them run. Got a few conversions for both. Requested WLs from Zeropark support for FI and ran them on a low bid and low budget to see if I could get more traffic, but stopped those after I overspent on the whitelists. I need to be in a different place before I tackle Zeropark. For the FI lander, I had the giftbox lander translated by someone on fiverr. I might want to revisit it later but it's not a focus right now. I just let traffic trickle in on the low volume/low bid on PropellerAds in case of lottery conversion. Both payouts are around $3.00 and SOI offers.

CC: MX, AR (Initially tested SOI and used converting placements to test CC offers)

For this, I was testing the theory of just switching out the offers from SOI to CC when finding good converting zones. I ran SOI offers at $0.25 payout, found zones that converted and then just created a new campaign with the WL of zones with 50+ conversions. The way I ran the SOI offer, I ended up at around -30% ROI, but I ran that campaign completely with the intention of doing the SOI-CC swap. The Lander stayed the same.

I have tested two different CC submits so far: offer #1 has payout of $7.20 and offer #2 has a payout of $20. Offer #1 has three conversions combined in my Pop/Push campaigns and Offer #2 has 1 conversion. Right now, it's not profitable and I don't know that there's any way to further optimize this, other than drilling down device type and other factors like that. I could look back at the SOI data and investigate more on those conversions to see if there are any correlations amongst the devices, browsers, times, os versions. But other than that, I would just drive more traffic and see and test. But I probably won't be doing that right now or soon. And I feel like that might be a more reckless approach.

Sorry about the long post. I don't expect a response to most of it and I'm happy getting one at all.

---

Yes.

---

First thing, I personally would not run it on pops.

Apart from that the flow is basically the same but for push creatives you would use something like "Get an iPhone 11 for $1" instead of "Click here to win an iPhone 11".

Same with the landers, I use the same landers as I use for SOI offers and just change the wording a bit.

Instead of "Enter your info on the next page" I use something like "On the next page enter your info and pay shipping".

But as far as I remember jaybot uses even exactly the same landers for SOI and CC submits without making any changes.

That´s normal, people prefer to get something for free and not pay for it

Possible.

When they already see that it has a price and they still click through chances are that they are already somewhat prepared for the next step.

Would be great when you could post a bit about it.

I installed Jupyter Notebook and Anaconda now and will try to don´t mess it up completely

My trainings project will be some kind of optimizer but only for Binom.

Basically when I splittest offers or landers that the script then stops specific elements based on given rules.

For pausing placements I already use Optimizer but for tracker stuff an own script would be nice and Binoms API is pretty easy to use so I hope I can get it done.

Man, we really should meet for some cold beer and talk about some cool affiliate scripts.

I am also a big friend of automation but so far my lacking coding skills kept me away from doing more stuff.

Yup, one man team aka the Lone Wolf Affiliate

About the stuff that you ran already

Why did you target only chrome?

Did you set the bid to $2 SmartCPM or $2 Fixed CPM?

When you set it to $2 SmartCPM then you probably still pay way less and that can affect the volume you receive.

Yep

Often the reason why it converts on one source and not on another is the quality of the BL/WL you have for the sources.

On a source where you have lists with thousands of conversions it´s much easier compared to a source where you have to start from scratch and run RON campaigns.

On the other hand it also happens that source A works great for affiliate A and source B sucks completely when at the same time source B is killing it for affiliate B whereas source A tanks completely for him.

For sweeps there are only 3 big landing page styles and one of them mostly works.

SOI sweeps usually work good on pops

Could be, NL is one of the better geos.

Although it´s a rather small country you can get good volume there and payouts are also attractive.

But what about another idea: The offer just wasn´t that good

---

Sacrilege!

Strangely enough, I've had better success on CC subs on pops and mobile. Most do it better on Desktop and push. This may be why I suck at them.

But, I also know a few who do really well targeting Desktop and Pops with adult traffic.

---

That makes sense to me. Pops is more interruption marketing, Push more targeted because they've already decided to click on the notification ad to get to your page. I would be less likely to trust some pop up and put in my credit card info if I browsing some torrent site when it popped. I've already decided to click to see the ad with push.

I could see that being better from the scientific testing perspective. Less changes between the two different tests (SOI to CC) so from that point of view, the test results could be more accurate.

I'll write something up when I get some positive green going . I think like most things, it's just more tedious in the beginning.

I had a similar idea . The multivariate split testing calculation gets pretty complicated. That was going to be my personal project trying to do the math calculations, but I might end up just trying to find someone on fiverr to do it for me. But again, that part of that stuff is on hold for me for now while I dig in more in the campaign side of things.

But there's so much that could be automated so this becomes more manageable .

I need to check with Binom support if they accept full campaign edits via API. I know for sure you can get the full info, and I know you can change the basic stuff like campaign name and campaign cost. If not, there's another option that I know of with python, it's just messier. UPDATE: they got back to me and they do options just not explicitly shown in description.


I'd love to man . I don't find myself in Germany very often, but I do have a friend that I met down here in Hawaii who I've done some quick day sails with. She lives in Sylt .

If you ever come to Maui, I'll show you some cool secret spots There's this hidden 600-foot waterfall that's pretty awesome. I have a waterproof notebook to scribble some code down on

And the coding stuff is a matter of when, not if. You can do it, no doubt in my mind. Anything I can do to help, let me know . Again, not an expert, but I've come across enough problems that I have figured some things out .


Honestly, just because in the past it's been higher quality for me. Other than that, really don't have a reason. I usually test Mobile with Chrome, but that could probably be just a bad assumption. Maybe worth a test without Chrome?

SmartCPM. Also something I have wondered about. I asked PropAds support about it once and I think they misunderstood my question. More likely that I worded it badly.

Interesting! Okay.

I've never had luck on AdCash, SelfAdvertiser, Hilltop, or ClickAdu, but that seems more like a Ben-problem than a network problem. Most trial campaigns were also on ZA, so maybe the competition just got the better of me.


Roger .


That's how I ended up there. I cut my NL campaigns down temporarily to a few better performing propellerads campaigns while I diagnose some stuff. Running just the WL, it gets like $2 profit, but I'd like to get it to a scalable point. If the offer will even work.

Haha, then the search continues.

One question:

Do you look at bots at all? I know you have recommended in the past not paying as much attention to how many bots are in a zone, as other experts on here seem to adamantly say remove them. More so, focus on the the metrics other than the amount of bots a zone has. Usually it seems that quality gets worse as percentage of bots goes up, but those zones can still convert.

Further on that, is there a good way to measure bots on desktop traffic? I think I read that Caurmen's method works but I haven't bothered to set it up. Is it worth it to add that extra blank offer in all of the campaigns?

I'm noticing in the harder GEOs that I'm getting Desktop to work better than Mobile.

Possible reasons I can think are other people have their lists dialed in better so I would need a super competitive to break into it, which I don't typically do on initial tests. People overlook Desktop I feel most of the time.

Sorry I put you through all of my rambling in my previous posts. Same to the Jaybot.

Thanks for the direction, man .

Adult traffic, you say? Sweeps CC with adult traffic?

I always have a post from your followalong in the back of my mind of someone who was pulling in XX,XXX doing/month doing CC, pretty much subbing them out for the SOI offers. There's a case study on STM somewhere where they go into it a little more. It seemed like they primarily used Pop in their methods. I have some campaigns named after them on my tracker, which are my attempts at the method.


I got three CC subs yesterday, 2 on Desktop pop and 1 on Push Mobile. I've gotten CC leads on pop in the past, just not recently. And I still haven't made them profitable. Used pretty much the identical lander I was using for the pin submits, but added the pay for shipping message before the CTA to the offer.

Still negative ~-50% ROI. Part of it is test budget spend, but I also just need to get some new offers.

I don't think I'm anywhere near nailing down the concept of Good BLs. So far, I have managed the most of my success with WLs and a handful of well-converting zones. And then there are the good straggler zones with a couple here and there.

I just checked my tracker and combined all of my data for all time for PropAds. I have about 6 zones that have 1000+ conversions amongst all of my campaigns, then about 60 more that have more than 50 conversions. A good chunk of those were lower payout offers, less than or around $1.

On the flip side, I have maybe 30 zones that have 0 leads and more than 10,000 impressions, ~70 with 0 leads and 5,000 impressions.

Anyway, I'm on the hunt for new offers.

---

As I always say, what works for one doesn´t necessarily for the other

Anyway, let me clarify my statement a bit.

I also had profitable CC submits running on pops but these were rather uncommon Geos for it like SA, BG, some Asian geos.

The big boys I would rather run on Push.

Spot-on

From that point of view you are right but you should also consider that you don´t run the test on same requirements.

Different conversion flows are already a different variable in the tests anyway.

Oh, I would do it much easier like "I test 10 offers and when offer converts 7 times and the other offers only have 0-1 conversions stop all other offers" or something like that.

Really man, I never target by browser when I don´t have to because of offer restrictions or so.

Check here for example



The "profitable" browsers are accountable for 86.7% of all traffic and they are accountable for 98.7% of all conversions.

Chrome Mobile is by far the biggest in terms of volume and conversions.

The non converting browsers are accountable for $-195 compared to $1915 total profit, so when I would exclude them my ROI would theoretically be about 10% higher.

BUT...

I want to go for maximum volume and to achieve this I just keep the other browsers just running and I don´t care if they just don´t perform that good or if they are slightly in loss or if they tank completely.

As long as the biggest elements are profitable they will always compensate the losses from the smaller elements in a good campaign.

And the thing is, trafficsources want to sell all their traffic, the good traffic as well as the bad traffic.

That means when you target too tight you risk to receive less volume for the good traffic as well because you don´t buy the bad traffic.

In your example let´s say that you now receive 50k impressions per day for Chrome because you target only Chrome.

When you would widen your targeting and accept all browsers = buying also the not that good traffic you often get favored then with more volume to the good traffic = Chrome as well.

So when you are willing to sacrifice some money for non converting traffic/browsers then in return you get more volume to the good traffic so that then you could receive 80k impressions per day for Chrome.

And this is it, you should always try to work towards maximum profit and sometimes it needs to eat some losses to get more traffic to the good elements to ultimately increase the total revenue.

Let´s say with your 50k Chrome impressions per day for targeting only Chrome you make $100 profit.

Then you run exactly the same campaign but with all browsers.

Chrome is still accountable for 85% of all traffic, the bad browsers for the remaining 15% will result in $-20 losses but at the same time the additional traffic for Chrome will result in $60 more profit.

Then the total profit from Chrome would be $160 - $20 losses from the other browsers = $140 profit = $40 more compared to only targeting Chrome.

One more thing about my screenshot



As you can see, Chrome Mobile is by far the biggest element followed by Chrome.

It´s a mobile campaign so usually there shouldn´t be Chrome for desktop anyway.

This just happens and probably I couldn´t avoid it even when I would target only Chrome Mobile.

Same for Android Webkit, often you just can´t exclude it.

And when we take this into consideration as well then I would receive about 95% the same traffic when I would target only Chrome vs targeting all browsers.

Of course that´s all a bit simplified and very theoretical but I hope you get what I mean.

When it sounds too confusing just let me know and I try to explain a bit better.

When you bid $2 fixed CPM then you pay $2 per 1k impressions.

When you bid $2 SmartCPM however it can also happen that you only pay $0.20 per 1k impressions.

This is very good for testing and to keep costs low but for scaling it´s often better to run WL on fixed CPM to get more volume.

It all has advantages and disadvantages.

BL campagns I would always run on SmartCPM if possible, WL campaigns on fixed CPM however are very good for scaling.

Not necessarily a Ben problem.

It´s just normal that not everything works equally good for everyone.

I don´t really use Self-Advertiser but I have good experience with Adcash, Hilltop and Clickadu.

On Adcash for example I had a campaign running in NZ quite some time ago that made about $300 or $400 profit per day.

No, I don´t care about bots at all.

The only thing I care about for placements is if they are profitable or not.

When they have 95% bot traffic but the remaining 5% convert so good that the placement is still profitable I keep it running.

When a placement has 0% bots but isn´t converting at all I stop it.

Yes, it´s true that a higher bot % can result in lower quality.

But instead of cutting the whole placement I rather clean my traffic from bad user agents, bad ISPs and stuff.

That way I send the "bot" traffic to a smartlink or so and send the remaining "clean" traffic to the real offer.

When you do it like that you already send better quality to the offer even when you still keep bot placements running.

In most cases desktop beats mobile when it comes to high payout offers and complicated flows.

One reason is that mobile users are mobile, they can be at work, visiting friends, in a restaurant or whatever.

Desktop users however are in most cases at home.

That´s why Desktop mostly works better for more complicated flows and landers like advertorials.

The users just have more time then to go through it.

Users probably also feel better to grab out the credit card and pay for stuff when they are in the comfort of their home and not in a overcrowded subway or so.

---

Sorry for the delay, but I think you misunderstood me @twinaxe

I meant hiding the offer, or the traffic source I'm running it at from the spytools or so.

---

Alright, got it.

Of course you can also use a cloaker to hide from spytools.

Basically you have two options to do so.

First option is to set the cloaker up to redirect the cloaked traffic to a different destination.

That way you can keep your domains completely out of spy tools but it could trigger curiosity from the trafficsources as well when they see the additional redirect.

The second option is to set the cloaker up to have a page and then show different content for reviewers/spytools/whatever and real users on the same URL.

Basically you add the landing page https://yourdomain.com/index.php to your traker and then for reviewers and spytools you show the content of https://yourdomain.com/cleanpage.php on the index.php.

For real users you show the content of https://yourdomain.com/dirtypage.php then on index.php.

The advantage is that there is no additional redirect then.

The disadvantage however is that your domain still gets caught by the spytools but instead of showing your real lander they will show the cleanpage then.

---

You could put a smartlink on the backbutton so that you do not have to bother updating it.
There was an interesting thread on tracking revenues for Main CTA vs Back Button CTA a few weeks ago:
https://stmforum.com/forum/showthrea...914#post397914

---

I am not a big fan of Jupyter notebooks. They are ok to some science, but to develop code, nothing beat writing in a pro IDE like VSCode (free) or the JetBrains tools (PyCharm for Python - paid): Code correction, auto completion, server debug console running inside the IDE. It took me time to get use to it, but I am so much faster now.

A few articles from people that do not like them to balance @bennimen enthusiasm :-D
https://datapastry.com/blog/why-i-do...ouldnt-either/
https://towardsdatascience.com/5-rea...k-4dc201e27086

Totally agree with that. Go Go Go @twinaxe

---

Hoo mama! Nothing too significantly different with my campaign stats but twinaxe's response really made me excited! I'm going to jumble around the order of the post a bit.

Oh man. There are those moments in life that hit you and stick with you forever.

I think the last post by twinaxe is one of those moments for me in my affiliate life. And it kind of flipped on a lightbulb for me.

I clicked thank post probably 100 times on my phone, hoping I could glitch out the forum and that my thanks would show up a bunch of times.

I've noticed a handful of times that when running a very narrow WL camp (1-5 zones) side by side with a BL camp, with all other targeting exactly the same, that the WL zones converted in the BL camp instead of the WL camp.

In my mind, since the WL zone was just targeting those WL zones, instead of all of those other zones that the BL camp targets, that the WL should perform much better.


My original thought process example:

WL camp:
$10 budget
5 zones
$2/zone of funds allocated (I know that's not how the bidding works and that there are many other deciding factors of how zones get traffic, amount of available traffic, cost per zone, etc, but for the sake of simplicity and the example)

BL camp:
$10 budget
100 possible zones
$0.10/zone of funds allocated (again, example sake)

But there have been a bunch of times that I've noticed that the whitelisted zones seemed to perform better in BL camps over WL camps, when looking at it on a zone by zone basis.

And that makes total sense to me with twinaxe's explanation.

I can totally understand how networks with some kind of quality algorithm would send more volume of higher quality impressions mixed in if you're spending more money with them.

It's like if you owned a business catering to different companies and you got a contract with a really big company, Google, that would bring you more revenue, you would totally focus more effort (higher quality clicks) on Google and making them happy (affiliate that spends more) than the small fry (me as an affiliate right now). Sorry little companies.

It may not be final and there still should be some amount of uncertainty in the bidding process because there's no way that their algorithm, however sophisticated, can guarantee 100% that this one click is going to convert over another. There must still be chances that a low quality click will convert when a high quality one won't. But I'm sure they can manipulate the quality to some extent. At least whichever traffic networks have an algorithm to detect and predict some kind of quality.
@twinaxe, have you noticed this high-quality-low-quality-volume click relationship with most of the pop networks? I think there are still a few networks out there that don't give an option to target higher quality traffic over lower.

I feel really happy after reading that post . Haha. Now I just need to find the right offer so I can put it to the test .

That also helps me wrap my head around scaling too. Up to this point, I've had a really hard time trying to scale any of my offers. I've always tried to fine-tune and find a way to increase the traffic but keep it targeted, just to minimize my losses since scaling inherently causes ROI to go down anyway, at least, based purely on the numbers.

When scaling, the volume should ideally make up for the loss in ROI, which is kind of the point in scaling.

From a number standpoint, if you were scaling by raising your bid to get more traffic, you're margins would get smaller, maybe to the point where you'd end up in loss if you chose too high of a bid. So always before I tried to scale, I'd drill down in to my stats and see if I could exclude something more, some kind of badly performing metric; os, device, browser. That way, I could minimize my losses. Doing this might help or might make my chances worse of getting higher quality traffic.

I also shouldn't be trying to scale an offer that can't handle the loss in ROI that comes with scaling anyway.

I find myself struggling with this. Just because an offer converts with your lander+offer combo, doesn't means it will convert well. That's something I definitely need to get better at realizing and determining. When testing an offer and it converts, I put a lot of effort into trying to make it convert, usually in the form of over-optimizing and throwing more money at it. And you can only milk it so much.


That would totally be doable .



You're the best at explanations .



That's a good note for me to internalize.



Thanks for saying that Twinaxe . I'm pretty it sure it was a Ben-problem though .

I think I'm going to do what I can to bunker down and focus on PropellerAds as my main one for now. At least 80%.


Good to know!


Do you send the bots to a smartlink specifically because they can convert or just as another option to send them anywhere but your offer?

When I tested some dating offers a few months ago, I had one that I got cut from for quality.

And that was because I was sending Lithuania traffic, LT, to a Luxembourg offer, LU.

I know, I'm a dummy. I'm not very familiar with all of the European countries.

I've never had problems with any sweeps offers though, in any of the offer's that I've run. Does that happen often? If it does, probably happens more at scale?



Good suggestion . I have done that in the past but that's definitely worth the test again. Thanks for that man .

And Interesting! I did see that on monetizer that you can attach a postback. I never played with it, I just put off until later. Meant to get to it at some point but it wasn't at the top of my list.

I've also haven't played with custom conversions yet either, although I'm pretty sure Binom allows them. At least there's a parameter called conversion status 1 and 2.

I have had a problem at one point redirecting all of my 404 traffic to a smartlink and getting domains banned that way.

I've had way too many domains banned. It has been a while now though since I've tweaked some things. One specifically that you mentioned somewhere about "eval" in the js script. I don't do a lot of that anymore.

The way I currently have my landers set up, simplified, if you try to go to one of my landers without going through a proper link (through a spy tool link or back button once the visitor has gone to the offer, even refreshing the page), it'll send you to a 404 error page. I think spy tools will probably still show the screenshot, but I don't think they'll get to the actual page, just the 404. I tried redirecting those directly to a smartlink offer.

The smartlink pages/offers were getting flagged as adware, so my domains were by association. Advertizer sent me a report after suspending a campaign I had with them showing the redirect chain and that it was my monetizer domain that it had a problem with. There were a couple other factors around that time that were flagging everything, but I don't think that specific campaign had any of those other factors.

Hahah. Have I not always said I'm no expert coder?

I am aware that true coders don't seem to like jupyter and how I use it. And pretty much if you get that debug console working how it's supposed to work, I'm sure you can work much faster and more efficiently.

It was just quick and easy to install and when you separate things in cells, if you have an error in your code, you can break it up and debug it relatively easily. In its own way

Everything I've heard, pycharm is good. If I'm not mistaken, you can even import jupyter notebooks in PyCharm.

Anyone jumping into python is probably better off in the long run getting VSCode or PyCharm or some IDE. Haha, jeremie .


Go twinaxe .

---

Seems like I prefer using it to direct to a different landing page on my current URL.
But I guess the real question is "do I need really need it"?
If I already use the landing page people already use, but how about hiding the offer?
Then if its not possible hiding the traffic source that I use then maybe I could just hide the landing page so that people at-least don't exactly know what it is I'm running there? (although based on the creatives it would be quite easy to guess but still)
Do you use cloaking for these kinds of purposes @twinaxe?

---

I installed VS Code and Pycharm and wasted so much time since yesterday just trying to get that shit running properly.

Usually I decided to go with VS Code because it seems to be faster and easier to work with.

I installed it and craeted a hello world project but I can´t get the integrated terminal working.

It always says Linter pylint is not installed.

I had to check what the hell this is, then i installed it but I still can´t get it running.

So running the scripts in VSC won´t work because the terminal won´t work and debugging also doesn´t work.

I am running OpenSUSE on my work PC so maybe it has to do with it, wouldn´t be surprised about it.

Pycharm community edition seems to work, maybe I will check it out more then although I still would prefer to get VSC running.

But it´s really annoying, finally I want to test a bit and then I can´t because that stuff is just not working properly.



As always it all comes down to testing.

When you have a big enough WL and run competitive bids you can get enough traffic to make it worth it.

When you only have a small WL with placements that don´t that much volume anyway it can help to rather run a BL campaign and cut bad placements then more aggressive.

On pops it´s not only that a high bid unlocks access to better placements and increases volume.
A higher bid can also lead to a better bid position so that your ad is shown before the users have seen 10 other ads from that placement already.

This can also help for better quality traffic.

To be honest, I never ran real splittests or so to check it for each network.

But to receive higher volume you either have to

a) bid higher
or
b) target wider.

In both cases you would pay more for the traffic than before thus getting better quality.

At least that´s how itn is in theory, in reality we have to test to find the sweetspot

Yep

It´s true that a good campaign can also turn red when you set the bid too high, it´s all about finding the best balance.

But I usually never drill down to browsers or such stuff.

The OS is something I set in the beginning of the campaign and don´t touch it anymore later.

For browsers and devices and such stuff you mostly have only two options.

Either the biggest elements are profitable and can compensate the losses from the smaller ones, then you don´t need to optimize it because you want max volume.

The other option is that you have to cut the biggest elements to get your campaign profitable.
Then you will often be left with low volume so that it´s not really worth it anymore.

In such case it´s better to move on and find a better funnel.

Of course there are always exceptions and probably everyone is handling it s bit different.

But this is how I do it

In the end most impotant is that you don´t optimize your campaign to death.

When you have to cut too many things to get your campaign into profit then it´s not worth it because then it comes down to the ROI/profit thing again.

A high ROI won´t get you anywhere when you can´t scale.

Thanks, I am doing my best.

With PropellerAds you can´t go wrong.

They have good volume, different ad formats and add new features all the time.

But here are also several members who don´t make it on Propeller and who have more success on other platforms.

In the end it´s up to you where you send the traffic to but a smartlink is often the best solution.

Not all of the filtered traffic is bots, it´s also normal users that are just outside your targeting.
It´s normal to receive a small part of the traffic that is untargeted, for example wrong geo, wrong OS or whatever.

When you send it to a smartlink you have the chance to still monetize these real users who wouldn´t convert on the original offer anyway.

Dating advertisers are often stricter when it comes to quality and for such situations it´s better to filter the traffic a bit.

Good question if you need it or not.

It´s true that you don´t really have to hide your lander when everyone is running it anyway.

But when you want to hide the offer you would have to have an additional check with redirect when the user leaves your lander so that you can filter out unwanted traffic.

When you would use a cloaker for it then the outgoing traffic would be hidden from spytools and such stuff as well so it doesn´t really make sense to cloak the outgoing traffic.

And when you cloak the incoming traffic then the cloaked traffic wouldn´t see your real flow anyway.

---

Why do you need a full-blown IDE for Python?

It's such an easy, readable language that a good text editor is usually all you need. And since you're on linux already, python is right there in your command line. A good book for learning Python, which used to be free everywhere is here

Depending on how complicated the code you want to make is (using tons of plugins), an IDE like VS or PyCharm will work, but for most things, unnecessary.

---

You don't need one. The same way a chef does not need a 200+ USD Japanese Santoku knife. He could well use a 25 USD knife
I think it depends on how much time you spend coding, and if it is worth the initial effort to get things running at the beginning.

Thanks for the book!

---

Hmm. I don't think it would specifically not work because of the OpenSUSE. Getting IDEs working and set up in the first place is one of the big reasons I never ended up using them.

I feel your frustration man.

There's a solution out there.

I'll see if I can install OpenSuse on a virtual machine and get it working.

I get where both jeremie and jaybot come from.

In the meantime, I think jay is right that python should be pre-installed on your computer already anyway, in case you wanted to get a head start without your IDE.



Now to just keep building my lists .


What about languages? Do you exclude languages too or do you include non-targeted languages to go after higher volume, and therefore higher volume of good zones?

I have usually targeted the language of my lander/offer. And I feel like it makes sense to continue to target that, because otherwise the visitor wouldn't understand the page.

Although, I did mistakenly put a Thai spinner lander on an Arabic offer the other day and didn't realize for a while. It got a conversion.

I wonder who converted. I don't go clicking around on Chinese popups to see if I would win a prize. I probably wouldn't even know if a prize was being offered, let alone know to enter my email.



I can never read that too many times.



Great advice! I have seen a few times that traffic specifically targeted for a certain GEO ends up bringing in other country traffic and I thought that the advertiser wouldn't be happy with that if that click ended up converting. Or at least that would be a reason to flag the quality. Throwing a rule on there to redirect traffic that doesn't meet the offer's guidelines would be a great way to keep the traffic clean.


Thanks Twinaxe

Haha, I could probably use the IDE. I tend to make things more complicated than they need to be. Nice share with the book man. Thanks .

Good comparison man . I probably should have started using an IDE a long time ago. I do have some specific reasons I think why I rely so heavily on jupyter but they're unrelated to affiliate marketing for the most part. Writing code with selenium would be tougher in an IDE.


Thanks guys .



Testing out a few new offers for BE that I found. Got a conversion pretty quickly on a WL, so it at least passed the first test. I'll start a BL camp soon to start testing the few different offers against each other.

My JO offer is doing okay on Push, so the next step is throwing more volume at it to see how it scales.

I opened up targeting on the JO Pop campaign, including all browsers instead of just chrome and I raised the daily budget a little to get some more traffic it toward to see how it would work. Will report back. It seems to be doing better on Push than Pop though overall.

Just on a fluke while testing my WL, I got a CA Galaxy Flip CC conversion. Instead of using a static image, I used a gif for its main image. It was hosted on giphy to reduce the stress on the server although I have no idea if that's allowed in their terms of service.

I think the gif itself was 8mb but it seemed to load quickly enough with how it was set up. I didn't check it with gtmetrix or anything yet. It may have slowed the lander down a little but when testing it against the static image, the gif lander got a higher CTR, and more conversions when testing an SOI offer. Wasn't stat sig though.

I noticed this when I was doing my quick stint in dating. There was a video lander that worked the best for me out of all the landers I tested, which wasn't many, but I was wondering why that wasn't more of a thing in sweeps. Is it only because of the load it puts on lander and its speed? In the business and ecom world, it seems like video backgrounds are a big thing. You would think, with how they draw attention, that we would use more video sweeps landers, especially if server load and speed were magically not an issue.

Anyway, I paused the CC offer while I was ahead and plan to get back to it at some point. Right now I want to get back to testing things that I can afford to test. Smaller Budget things. And also, right now I'm profitable on at least one CC conversion .

I have a CZ offer on push, $5+ payout, that gets 1-4 conversions a day on one zone. I have 5 campaigns set up for it right now: 2 WLs with different bids for one zone only, 1 global WL, 1 BL camp and 1 CPAgoal 2.0 for Push, all camps on PropellerAds. I made while I still had maximize ROI on the mind over maximize profit, so they have lower daily budgets. So far, the only zone with more than 2 conversions is the one I have the separate campaigns for. I'm going to bring it to Pops too. I had it on Desktop and had "eval" in the js code but I removed it so it doesn't bring down any more domains. Fingers crossed . It had potential to do well on Desktop before I stopped it.

I have a few CPAgoal 2.0 camps on Push that are profitable and they seem to be getting conversions on lower volume zones, when comparing them to my BL campaigns. They also have been more profitable than unprofitable for me. They're based on CPM instead of CPC though, so you pay per impression, which is less desirable.

I'm also playing around a little bit right now with raising daily budget or removing it altogether to see how it affects the campaigns and the amount of impressions/clicks they get on both push and pop. On the campaigns I removed the daily budget, I still have a total budget set. I can't afford a mistake like that. But I'm making a guess that a campaign with a total budget of $50 will act the same on a given day as the same campaign with a daily budget of $50 but a total of $100.

Also still testing NL. Found some other offers to test along side the old ones to see if it can be worth it. I don't think I put enough traffic into it before.

Will post screenshots on next post.

---

It´s no problem for me to get Python scripts running.

I use Sublime for coding them and then run it in terminal.

Right now I am trying to make a script to grab info for specific campaign groups from Binom.

Getting the info from the Binom API is no problem.

My plan is to get the campaign IDs from these groups, check the campaigns stats and then pause landers or offers based on simple rules.

But it´s annoying to write the code, save the file, type command in terminal.

That´s why I want to use an IDE where I can write the code and then run it with a click of a button.

It depends a bit on the geo and trafficsource.

Usually I use local language and english.

In some trafficsources you can´t target several languages at once, there you can only target all languages or just one language.

When I have countries with 90% local language then I also target only that language.

But there are many countries with several languages where it´s better then to target all instead of only one.

It´s not only about the most spoken languages or official languages in the country, it´s about the browser language so there can be 90% of the people speaking the local language but still can have many different languages in their browser.

Another reason why targeting only local language(s) is that it can help to increase quality as well.

---

Sorry to sneak in here Twinaxe but I wonder how to do this when you have 30 campaigns but only want to see the combined placement stats from for example 5 of these campaigns?

---

It depends on the tracker you use if it supports to group campaigns together.

In Binom I can do it so I would just check the 5 campaigns I am interested in and then check their stats combined.

This feature is not only helpful for zones.

When I run the same campaign on different trafficsources I also combine the stats and check what offer or creative is doing best.

Can help to gather stats faster compared to running and checking only source by source.

---

Oh cool, that's a nice feature. I use Voluum.
I know the're a way with Google Sheets to combine multiple sheets, but it's time-consuming. Nice one Binom (maybe someone from Voluum reads this

May I ask you a follow-up question? Let's say you have found the zones that you want to block in 5 campaigns, how do you block them?
The way I do it is copy them from Voluum to a .txt file, and then from there I copy them into the blacklist of each campaign seperately. That's a bit a boring and annoying task. For example in Propeller, I always have to right-click the
campaign name, open in new tab, then scroll down, insert zones, click save...
It would be nice to have a tool with a "global blacklist" feature, where you just insert the zones you want to block and click on the campaigns where you want to apply it. It doesn't seem that theoptimizer has this ability, nor does Automizer from Voluum have it.

---

Very detailed and interesting case study!

Do you keep running campaigns with PropellerAds traffic? We’ll be happy to buy your case studies about successful campaigns. All details are here: https://stmforum.com/forum/showthrea...l=1#post402612

---

Home > POP / PPV / Redirect > Follow-along Campaigns