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

---

[Funntastic] My AM trip. Le'ts start getting my hands dirty! (16)

---

Hello to everyone,
and welcome to my follow along!.

I'm looking for some good tips or advices to better understand the optimization process and all the stuff related!

(I'm not confortable with emotional stories or long rounds of words... so i'll be extremely concise and technical to get stright at the point).

Skills/Tools:

• Voluum : 90%OK (it works but I have to go more deeply) - maybe I'll change tracker due his recently price raising.
• Adplexity : 100%OK (mobile+desktop)
• Amazon S3+CF : 100%OK
• Custom domain for tracking + offers + DNS : 100%OK
• Affiliate networks : 10+
• HTML + CSS : 75%OK (I need to improve my speed in ripping and optimize LPs + create a library of templates/scripts)
• 40days Vortex guide : OK (no green campaigns till now).

Traffic Source :

• PopAds : 70% (need to better understand how bidding works)
• I'll try also new traffic source (Propellerads, Zeropark) but now I'm focusing on PopAds.

Campaings :

• several direct link testing (no good results)
• 1 FB campaign in tier1 geo (suicide)
• 4/5 mobile sweeps campaign in tier2/3 geos (no good)
• 1 desktop AV campaign in tier 2 geo (I want to share and take a look at this)

My AM show me a good offer in AV vertical (desktop) in Europe (tier2) that is converting better than the other sweeps I tested till now.
Other affiliate networks doesn't have offers that i could test with same LPs in the geo with I'm more confident.... but anyway I decided to start with one single offer (this may have been a serious starting mistake...) and one geo.

Offer payout = 1 $

Step 1 (BOT Testing):

• budget 15$ at average bidding - 14.000 views
• Cutted all the placement that were >35% bot
• 1 conversion

Step 2 (LP Testing):

• budget 40$ at average bidding+25%
• 5 LPs tested
  
• total : 28.175 visits - 899 click - 6 conversions - CTR 3.19 - CR 0.67%
• Best LP : 11.177 visits - 350 click - 6 conversions - CTR 3.13 - CR 1.71% (the gray one - I didn't go totally black)

(2 LPs were cutted suddenly for very very low CTR < 0.5%)

Before step 3 I cutted low converting edges in PopAds (I could re-test eventually later but I tried to cut down strong - preserving anyway enough volume):
- Chrome browser - no conversion
- AD Block - no conversion
- site id with no conversion and 5payout spent (it's only one placement)
- category id with no conversion and 5payout spent (it's only one category)
- High screen resolution - no converting (nerds with 4K screen and so don't install antivirus on pop...)

Step 3 (Script testing + first round of cutting in Popads):
Now for the first time I've got a landing page+offer that converts (low...) but maybe with some scripts? Let's test!!!

Budget 15$ - average bid+25%

• LP1 : audio script - beep at open : 3001 visits - 87 clicks - 3 conversions - CTR 2.90% - CR 3.45% - cost $7.45 - revenue $3 - ROI -59.74%
• LP2 : back button script - link to same offer : 1705 visits - 72 clicks - 0 conversions - CTR 4.22% - CR 0.00% - cost $4.37 - revenue $0 - ROI -100%
• LP3 : js alert on load - insane low ctr : CUTTED
• LP4 : js alert on exit - insane low ctr : CUTTED

While writing (from step 1) I've spent 70$ and get 9$ revenue in this campaign.
Now I have no great ideas for the step 4...
I really appreciate any suggestions you want to give me!

Peace and love!

---

Hello guys!
From the step 4 I thought having a nice offer and 2 decent LPs... so I want to start to split testing traffic source variables (popads).
From to step 4 to 5 I also added SSL to all the redirect chain domains, ad switched from Voluum to Binom (I triple checked this setup and works fine).
I used my best 2 LPs and added 2 new ones (same design and script, but very aggressive text).

Step 5 (PopAds Bid split testing)

1 - High effective bid +25% : Views=4545, Clicks=224, CTR=4.11%, Leads=2, Cost=$13.11, CPC=$0.00240 ROI=-84.74%
2 - Average bid : Views=6469, Clicks=146, CTR=2.26%, Leads=0, Cost=$9.41, CPC=$0.00146
3 - Average bid -25% : Views=2456, Clicks=76, CTR=3.09%, Leads=0, Cost=$2.60, CPC=$0.00106 [low traffic aviable at this bid]

In all the previous steps I was bidding near that point (in popads) where the amount of traffic start to breakdown (near average bid + 25% according to Vortex advice).

So in this step I went from -3%ROI to -85%ROI. Where am I wrong?!?
1. Offer is still running (talk with AM - offer is still converting for other affiliates).
2. Affiliate network tracks clicks that I'm sending to
3. Redirect chain works good with SSL
4. The offer doesn't have a cap
5. PopAds campaigns are clones of the first (I change only bid amount)
6. The domain is not flagged by google and runs perfectly via Amazon S3+CF
7. Page time load is always the same (< 100ms)

Next step :
1. Hire a translator on fiverr to test same offer in other geos (DONE)
2. Bang my head on the wall several times (STILL DOING).

Peace and love!

---

Hey!!!

ROI in this campaign still sucks.... so I decided to keep it run at lower volume, split testing other new LP and old one with little improvement.

As I told in the previous post I hired a translator to try catch some fishes in other geos.. where I have also many more offers I could test.

Step 6 (Try to export campaing in other geo)

As usual (at first) when I'm in a new geo I run a bot detection round. ($15 with my best landing page and offer + parses JS).
To speed up the optimisation process I absume that previous test in other geos could be right even for the new one.
So I started with only Firefox user. (I can test chrome in another round)

Results were interesting..
Really low bot presence and 2 conversions (in this geo same offers pays +50% -> $1.50) - so let's start at -80%ROI

After cutting few bot placement and one category that I usually don't wont to have, just now I'm running 4 LP + 5 offers test and I'll try to optimise in the next step.

I keep testing! But if someone could give me some tips... I'll really appreciate ))

Peace and love!

---

I don't do pops, but if you post some of your ads, I'd be happy to provide general feedback on them.

---

At this link you can find some stuff i used for this campaign.

Windows alert fake popup works better than others

https://www.dropbox.com/sh/sol6m9fc1...MaPIHdyoa?dl=0

Copy isn't full black... It says "you could have a virus" ... "System could be in danger.." and so on.

---

Okay, I don't really do black hat, so can't help out too much. If you are running more mainstream offers, the one of the keys really will be making sure that your ad and our lp match and reinforce each other.

---

I'm running Pop, so I don't have an AD... In the latest LPs I tried even to match as much as possible my landing with advertiser page.

---

Step 7 (Give up this campaign)

Yesterday I decided to stop this campaign to analyze and think about what has been done. I had one good day, so I was going deep into that daily statistics.
While checking about this and that... I get an e-mail from the AM that tolds me that the payout, for the offers I was running was 20% decreased.
This was the last straw...

If the payout has been decreased I think that other affiliates are going high volume and high conversions too.. So I checked into AdPlexity what they are running, looking for the campaigns matching the url of my offer. Big surprise when I found only my landing pages (only two of the +10 I used....) and only another one with low volume...

Now i have some questions...

1. Are the other affiliates cloaking? (I think so... for AV offers I belive this could be a standard practice).
2. Why AdPlexity catched only two of my LPs? (It could be really useful understand how it works... So I could let AdPlexity catch the landing that I want and hide better ones...
3 Offer payout reduction mean thats other affiliates are doing well?
4. Name the path into my server like : "https://mydomain.com/av/uk/lp1/index.html" is a good idea? I don't think... Adplexity shows exactly that paths... So anyone can change the url with different geo or different lp number to see easily all my landings (i know that this isn't the point why I've got non profitable campaign... but it should be a thing to take care from starting).
5. My best day was on Sunday... does big affiliates go to the sea on sunday?

Next step:

Stop campaigns for this weekend and watch a video-course about CPA affilate marketing. Hoping to get some useful tips. Then reorganize my ideas and start again with POP traffic + desktop offers - the most of newbies goes POP+mobile, I want to try it different.

Thank you for the answer you'll give me.

---

1. Yes, many other affiliates are cloaking, esp. more aggressive stuff like AV. Cloakers are too expensive and too much trouble if you're just getting started (imo). I would suggest rather testing something a little less blackhat, even say sweeps (leadgen or carrier, but more traffic to work with for leadgen) for example, among other verts. This will save you some headache and give you more time to focus on your camps.

2. Adplexity just has bots scraping ads they see around the internet on known publishers that they watch. If you're not doing much volume, it's not surprising it wouldn't have seen your ads. Or you could buy ads on placements that adplexity doesn't even monitor, and your camp would never show up in there. Though at some point, keeping your landers out of spytools could be worth thinking about, don't stress about it too much right now. If you learn from the landers your rip and the split tests you do there, you can come up with small things even, like cleaning up the copy, testing new headlines, etc, that will still blend in with the "rest" of the landers on adplex and not pose as big of a threat to your lander standing out.

3. No, I wouldn't necessarily think it means others are doing well. Maybe the advertiser found out their numbers aren't panning out appropriately with the LTV of a customer compared to the cost of the lead/subscription . This could be a number of things. A better gauge of an offer doing well, would be perhaps seeing it running a lot in the wild, spying, with volume and a decent amount of competition running the same offer.

4. Name your paths in a way that is organized and makes sense to you, but yea, I personally just kinda change up the numbers /naming scheme of my individual lander folders so its not easy to iterate through. Maybe like 001, 005, 012, or harder, some random letters in front of each separate number, etc. Also, make sure your server is set up to serve a blank or error page when someone goes to an index-less folder, so no one can just browse your LP directories.

5. There's a lot of volatility running camps, and some things may work better on the weekends, some during the week. It's really going to vary across verticals, placements/sources, etc. Weekends are often better for me, but I've had plenty of times where that's not the case too.


You're doing great here learning the ropes. My biggest suggestion would be to not focus too much on any one camp. In order to really find good ones, you'll need to test a lot of offers & campaigns. One good way is to set up 3-5 different camps at a time whenever you launch. You could launch different verticals in the same GEO (*if the volume is there), or you can test numerous GEOs at once in the same vertical on one source.

The latter is particularly easier when testing some GEOs that speak the same language (like German can run in AT, DE, CH, even parts of BE, or Spanish). You can reuse your creatives across each camp. Of course, more generally, this will get easier with time as you run more and more camps and collect a stash of cleaned landers for a vertical in many languages.

As for testing desktop or mobile, go for it. It's the same process setting up the camps, so throw it in there and keep launching! Build funnels that are an appropriate size for the given traffic volume for your targeting, and budget accordingly (that is so you can run each split test to stat. sig in a reasonable amount of time.)


If the volume's not there (for desk/mob/whatever) where the camp could be foreseeably worth your time profit-wise, then move on and find something else (whether testing something else on that same source, or testing the same camp elsewhere).

---

Thank you Chris!
I greatly appreciated your advices
After the weekend spent reflecting and rearranging ideas, I'm finally ready to re-start with sweeps!
I've choosen one new geo, translated LPs, and ready to start with all the phone, vouchers and 'win something' offers that I can find!
I've also created a new account on Propellerads... (today popads doesn't allow me to create new campaings....) so I'll test even a new traffic source.
Stay tuned!

---

Wow - have NO idea how your follow-along didn't make it on my radar until now! Profuse apologies for the late arrival!

Many thanks to Chris who jumped in to answer every single question!

Funntastic (great username btw!) - you're already doing a LOT of the right things. It's just a matter of testing more stuff in a systematic fashion.

I only have several comments/suggestions at this point:

1)Try to stick to a testing methodology. This post gives a bird's eye view:

https://stmforum.com/forum/showthrea...l=1#post348214

2)Try to launch in several geos at a time now that you know what you're doing. Yes you'd be spending more money on testing, but chances of finding something profitable will be higher as well. That way you wouldn't be focused on one campaign holding your breath for it to succeed.

3)Once a campaign is profitable or breaking even, try scaling it to other traffic sources as quickly as you can. Offers tend to be short-lived (I'm talking about the kind that converts on pop traffic). So the bigger and quicker you can scale something before it dies, the better.


As for keeping your landers out of spy tools - do some research on your own (try googling).

Looking forward to your next update! Now that I'm subscribed to your thread, I should respond a LOT more quickly.



Amy

---

Thank you Amy
I didn't notice your reply... It seems that I can't get alert from my thread (I have to check better my setup).
So I launched some new sweeps campaigns in the same geo, I'll update this post ASAP with the new test results.
Thank you for that link that show the process optimization, I will test for sure.

---

Awesome! Looking forward to seeing some stats soon!

If you go to your profile and click on the subscriptions tab you should see all updates of all posts you're subscribed to. And when you post in a thread you will be subscribed to it. Every time someone updates thr thread with a new post you'll see it in the subscriptions tab.



Amy

Sent from my SM-G930W8 using STM Forums mobile app

---

Hi to all guys!
I'm still testing but I'm not having too much time to post my results!

My tracker shows that in this last 30 days I've tested 76 different offers, 103 landing pages, 45 campaigns, 8 traffic source and 13 Affiliate Network!!
I've done all my best to be consistent and focused!!

And finally I can see some green! I found my first 'near ROI' converting campaign changing a little my approach.

As suggested me by veteran affiliates I choosed one geo (tier 3), setted up a custom landing page that gives me the possibility to test lot of differents sweeps offers without so much adjustment from a type of offer and another.

In this way I can Test 'win voucher', 'win iPhone', 'win cosmetics' and all this kind offers in a massive way with low low effort. The main focus at this point is only to find an offer that can make some conversions. Once this happen I can 'pull out of the garage' my 'sunday LPs' and start to test his potentiality for real.

Now my best campaign is near ROI, without cut any placement (only targetting right devices and carrier) and soon I'll start split testing Title, CTA, colors, etc... I'm starting right now to scale it on other geos and after on new traffic source!

Today I also achieved a +30% payout bump on this offer (just asking).

Never give up!!!

---

Testing extensively in a systematic way is a sure path to success!


Congratulations! Very well-deserved after all that testing!


That's what I call a systematic test methodology - very inspiring!

So you're using one lander to mass-test offers to find a winner first, then testing more landers on that offer. I like to do the opposite - test landers first to lock down a good one, then use that to mass-test offers.

I choose to test landers first because there are typically only a few popular ones running (there are a million different versions, but only a few that look and feel completely different), whereas there can be lots of offers available.

But whatever works for you!

(But you said you tested 103 landing pages? I'm a bit confused there.)


Good plan - however, just a reminder:

Mind the 80/20 rule. Testing extensively in a cheap manner, picking promising offer+lander combos, then scaling hard and fast is the way to go.

In other words, be FAST and EFFICIENT.

So - instead of testing lots of lander variations to make a bit more of the traffic profitable, it may be more efficient to just cut placements until your campaign is profitable.

You could be spending that time on testing offers and/or scaling profitable campaigns to more traffic sources instead, to end up with a LOT more profits.

8 traffic sources is AWESOME for a start! But think how much more you can scale if you were to test, say, 50 more sources, to find 10-20 more that have quality traffic.

Testing more lander variations is great - I'm not trying to discourage you from doing that. Just a reminder to always heed the 80/20 rule - because when optimizing a campaign, it's so easy to fall into the trap of over-optimizing past the point of diminishing returns.


Very nice! Have you checked to see whether the same offer is available on other networks? Would be worth testing - conversion rates can be night-and-day.


I know you must be very busy optimizing and scaling - but it would be amazing to see some stats if possible! I hope you're as proud of yourself as we all are of you! And thanks for sharing your experiences and knowledge along the way!



Amy

---

Thanks for all the explanations! Understood.

Sorry to hear that you only got lucky with one placement! But that was another good lesson learned:

Next time, make sure that there is a good number of profitable placements and not just 1-2 placements BEFORE scaling massively. If only a couple of placements are in profit, you may still want to do a quick test on another traffic source (that you know has good quality traffic), but no need to scale hard and fast to a lot of TSs. (I know you didn't this time - just for your information.)

Also: Don't make the mistake of spending a lot of money cutting placements, when there are not enough profitable placements that are generating enough profits!

After running traffic to your best offer+lander for let's say 1-3 days, look at your placement stats, and look at all the placements that have made at least 1 conversion. If most of them are in profit, it's a good sign - I would probably keep running the campaign longer. And the more placements that are profitable with 2+ conversions, the better. You can also estimate the average profit per day, by taking the profitable placements, and adding their profits together, then dividing by the number of days you ran traffic. If this number is high enough for the campaign to be worth your time to manage it, then keep cutting placements. If not, then it means no matter how many placements you cut, the campaign won't give you enough profits. (Of course, this is assuming you're not throttling your traffic.)

Many new affiliates make that mistake - they don't verify that there's enough profits first, and keep wasting money keeping the campaign running to cut more placements. When there are enough profits from profitable placements, you can keep cutting the unprofitable ones until you're profitable. But when there are NOT enough profits in the first place, no amount of cutting will end up in profits.

This don't just apply to placements, it applies to other traffic segments as well, i.e. OSs, carriers, browsers, times of day (dayparting), etc. As long as you see one or more traffic segments with enough combined profits , you can continue to cut unprofitable segments.

Your mastermind buddy knows his stuff. Here are a couple other possibilities:

-Your competition may have found a way to find out what the url of that placement is, and approached the site owner to buy traffic from them directly.

-Your buddy mentioned the possibility of the TS not allowing your ad to show in certain placements due to it being overly aggressive. Some TSs allow publishers (placement owners) to report/stop ads that they don't want to be shown on their website. This way the publishers are "helping" the TS to monitor and moderate the landers that are being run. This may be part of the reason why sometimes, you start off getting more traffic, then the amount decreases after the 1st day or 2. You can try to create another campaign that is exactly the same as the first one - sometimes that can get some of that traffic back. It can be annoying to have to keep doing this though. (Perhaps something to automate if you have the skill, or can hire someone that does.)


That's another very good discovery: Don't judge an offer by ONLY its payout. Very nice!



Amy

---

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