Home > Mobile > Follow-along Campaigns

---

A Young Student and A Decisive Goal To Make 100$/Day ! Challenge Accepted (42)

---

Hi guys,

I'm Tien and have been brought to this industry by Charles Ngo. My first goal is very clear: I want to make my first 100$/day with PopAds in next 3 months !

Though It may sound crazy to newbie like me but I truly believe that with the help of a lot of elite marketers in STM plus my hard work, it is possible.

I've just digested quite many threads during the last 3 three days after joining STM and it's time to take rapid actions.

I am inspired by the direct linking method for newbie in a thread of Mr. Payne: https://stmforum.com/forum/showthrea...ing-to-Profits. So I decide to go with this.

This method is about trying to test different 1-click, 2-click offers in many verticals which are having high performance in one GEO and one carrier. Then optimize the ones that are converting well.

Here is my collection

Traffic source: PopAds (of course)

GEOs: India

Carrier: Idea

Vertical: As a part of the method, I will just pick which offers that are converting well. Currently I have: Utilities, Gaming, Adult.

Offers: I have already picked 11 offers from Advidi, Mobidea and ClickDealer

Campaign Structure: I will test a batch of 5 offers at one time to see how they work before adding more offers to test.

-------------------------------------------
6/7 - Camp 1
Offers: 2 Anti-virus, 3 gaming
Budget: 3$/day
Max bid: 0.1 (Legacy bid)
Category: Exclude Adult and File sharing
Browser: Exclude FireFox. As I learn from a post that FireFox Browser do not convert well in PopAds so I'd better exclude that at the beginning and can start adding it back if the campaign is performing better.

Here is a screen shot of the Traffic Estimator in PopAds:

And here is my performance if Voluum.

Some of my thoughts and questions:
- As vortex points out in a post: https://stmforum.com/forum/showthrea...oluum-Version). Should I abandon the offers after spending 5 times equivalent to the payout or optimize some targeting and continue to test them ?

- I just set a maximum spend of 3$/day in PopAds. But after an hour activating a campaign, I realized that it is spent nearly 4$. I do not really understand this. What if I let it run for a day and my budget may run out.

- What is the best category to add in PopAds if I want to test Ultilities and Gaming offers ? My target may be just too general. Do I have to add adult categories ? Or adult categories type is just for Adult offers .

- I hesitated to test adult offers at the same time with ultilities and gaming since i think it is chaotic and less focused. I already have some adult offers in my collection but I've decided to save them for the next batch.

- About testing a batch of 5 offers every time, if I want to test another batch, I will add them in a seperate campaign with the same targeting ? Am I doing right ?

I'm more than happy to receive feedbacks from you guys.

Regards,
Tien

---

I will leave all the major questions to the more experienced affiliates but I will address one issue really quick.

Bidding - Your bid of 0.1 means you are willing to pay $.10 cents per popup which would be $100 per 1000 impressions. With a budget of $3 I don't think that is the bid you was aiming for. When entering a bid on popads you need to enter what you are willing to pay per popunder. If you're wanting to pay $1 CPM it would be 0.001.

---

Hi Tien,

The main problem you have here is your bid. $0.1 bid in Popads is the equivalent of $100/cpm which is crazy. Check some of the bidding guides in STM and adjust accordingly ( also taking into account your offer payout). Then let the campaign run enoguh so you have statistically significant results. You cannot take decisions like category targeting or cutting offers when your results are not statistically significant.

If you don't want to have your budget spent as quick as possible, and have it more distributed during the day, use the throttling option on Popads to reduce the spending speed.

Cheers!

---

Check out PPCmode.com. The guy who runs it has a free PopAds guide (and a paid version too) that's dedicated to just this traffic source. Good intel.

---

Hi saigonscott,

Thank you I have already learned a lot from PPCmode. And that's the reason why I am dedicated to run PopAds.

Are you from Vietnam ? I notice your name has saigon in it

---

"I will leave all the major questions to the more experienced affiliates but I will address one issue really quick.

Bidding - Your bid of 0.1 means you are willing to pay $.10 cents per popup which would be $100 per 1000 impressions. With a budget of $3 I don't think that is the bid you was aiming for. When entering a bid on popads you need to enter what you are willing to pay per popunder. If you're wanting to pay $1 CPM it would be 0.001. "

Hi doctors, gameandwatch

As I learn from some campaigns in PopAds before, bidding this high does not always mean that I have to pay 0.1$ per impression. Like some campaigns in PopAds when I do some test by bidding 0.5 (500$ CPM) then checked the stat in volumm. What I found is that the cost for that traffic rarely reach 0.5$.

In other words, I think that bidding this high will help me get the best placements in PopAds compared to other competitors .

---

So are you looking for advice or you want to persuade us that you picked the right approach? (no pun intended)

You are right in one thing, any network that offers some kind of "smart" bidding algo, will charge you less than your maximum bid ... you will simply pay just a bit more than your closest competitor.

This still doesn't mean you have to run for crazy high bids. Think about it this way : you don't have a performing offer, or a LP, yet ... it's not the time to be the highest bidder. In a situation like this, it's better to start low, or let's say somewhere in the middle, definitely don't go for the top bid just yet.

So, scratch what you've done so far and restart with a way lower bid.

Check the traffic you have received so far, and make sure it matches the carrier and other targeting that it has to.

As for testing batches of offers - are you using LPs for this? Cause if you are not, you can test multiple offer types at the same time, just give it enough clicks to reach statistical significance. On a side note: definitely learn how to make/rip/use LPs, it's very hard to profit without them.

I also noted your decision to cut firefox based on some article that mentions poor performance of firefox traffic from popads... this is not a good decision again. You should rely on your own testing, not something that you read somewhere - it might be solid info, but you never know until you test it a bit.

Anyways, not gonna flood you with more tips now, you really need to restart again with a lower bid and get more traffic

---

Hi matuloo,

I am looking for advice of course. The explanation of my bid and my target is what I am thinking it is right, not to prove that I'm right . And you have pointed me some ideas to continue with this campaign.

I will rebuild my campaign from scratch again and update the results soon.

Thank you.

---

I live in Saigon, but I'm from the US.

---

Cool, didn't want to sound like an ass, just felt like pulling your leg for a bit

Let me know when you restart and we can take a look at your numbers again... there will be many more things to go over, but it's easier to explain stuff when we have some data to work with.

Mat.

---

I started my campaign from scratch.

My CPM bid is: 2$
My category target: all categories

And the result:

When checking which category they come from, what i get is one site from Adult and one site with a * mark. I do not know what * stands for ?


My thoughts and questions for advice:

- I do not know exactly whether I give them enough clicks or not. But there is an offer that convert and the ROI is ~50%. I read some thread in the forum that says if the offer do not convert once after 5 times spend of the payout, I should leave them.

Just think that I should abandon these non-converting offers and continue to test with this one plus other offers. Is it a right move or should I give them more budget to reach satistical significance ? Given that my budget to buy traffic for this month is 200$.

- Do I need to seperate between mainstream and adult audience into 2 campaigns to test the offers ? Should I use adult offers for adult category and sweeptakes, AV offers for mainstream. Two ideas I come up with when testing them in batches

AV and adult offers + adult category
sweeptakes, AV and app install + mainstream

- Can I test multiple campaigns with different offers at the same time and with the exact targeting and bidding in my traffic source ? Am I competing against myself ? I think that I should get the results faster so that it can speed up my optimization before the campaign become volatile.

- Does it make a difference with 40$ spend on testing 10 offers in one camp and 40$ spend on testing 10 offers in 2 camps (1 camp/20$/5 offers) ? What is an optimal way ?

Regards,
Tien

---

Just a new update while waiting for the advice. Today I am testing the batch of 6 offers from ClickDealer.

The 5 new offers are added together with 1 offer that have 2 conversions in my previous test.

All of them are erotic content and for Idea carrier. I think that it would be better to send these erotic offers to adult category targeting than mix adult and mainstream toghether. So in my targeting setting, this round is only for adult targeting.

My CPM bid: 2$ (Legacy bid)
My daily budget: 10$
Total budget: 20$

I'm gonna update the result very soon.

In the meantime, I am trying to take some landing pages on Adplexity to get familiar with working with them. Because I do not know exactly which kinds of offers will be optimised and used landers (I have not finished my batch testing) so I decide to just see what is being run there

I see that there are some sweeptakes campaigns is being run in India and the offer is from Advidi. I shoot a message to my AM in Advidi is see if I can get this offer and run a test.

Just have an idea that I can use these sweeptakes lander and find the suitable offers for other countries instead of India alone.

I have a question that do I need to only focus on what I am testing with those SOI, DOI offers in India/Idea carrier or I can wait for the stat to come in and use that time on working with some landers and find new idea for the next campaign like what I've just presented.

Regards,
Tien

---

Hello again Tien,

as you can see, you received more traffic now, even with the low bid, so there is more data to play with.

Since your budget is low, I would recommend to focus on a limited amount of offers/campaigns for now. Looks like you've hit a working offer for India/Idea, so definitely keep that one in rotation.

Hard to say whether you should remove all the other offers already, you only sent about $2 worth of traffic to each of them - and part of that traffic could have been bots or otherwise poor traffic.

You didn't use LPs, so you can't look for placements that send traffic so poor that it won't even get through the LP - which would be one of the things I'd recommend to you now.

There are some things you can do now :

- start testing LPs
- run some kind of bot test, even if the simple one that Caurmen talks about in this guide : https://stmforum.com/forum/showthrea...n-Any-Campaign
- make sure you are getting the traffic you need - so Idea carrier in your case, what does Voluum tell you?

You can also continue with the setup you have now, so keep on testing more offers with a directlink and see if you can get a few more that would convert. Then, once you have 2 or 3, you can try to fine-tune the campaign and introduce LPs.

Right now, you need to make sure that you are buying real traffic and cut any placements that are sending you too much BOTs and find a few offers that are making conversions.

Now to your questions :

When mass testing offers, you need to stick to some rule of thumb, 5 times spend is a pretty good one.

The problem in your case is, that you are not using some proven LP but just a direct link. You also don't know anything about your sources and bot %... By running longer, you might be just wasting money. But in reality, you didn't run enough traffic yet, to cut the offers with certainty. I'd recommend to look at the data first and see whether you can find something in it.

There are no definitive answers to this, except for the obvious : follow the rules of the advertisers. Some don't want adult traffic, some are ok with it. In reality, adult traffic can work with a lot of verticals, and mainstream traffic can respond very well to erotic offers.

The only limitation here is your budget. You can test plenty of "things" at the same time and you can create dozens of campaigns at the same source, with the same targeting - in this case, you will obviously compete against yourself in a way, since the targeting would be the same. However, there is already a ton of other guys running with the same or similar targeting as yours, so a couple more campaigns won't really make a difference on these pop sources.

If you want to properly split test those 10 offers, you should leave them in 1 campaign, to make sure that the traffic rotation is as even as possible.

---

Hi Matej,

I have just set up the bot detection test for my campaign. Yet I have a little bit confusion about how to do it properly in the tutorial by Caurmen

I follow the code in the tutorial and creat my bot testing lander. Then I create the new campaign with the lander mentioned and the offer is my real campaign URL.

The lander:


The offer:



The new campaign


After that, go straight to PopAds and create a simliar campaign with the one I am testing ()

One thing to notice that even though I bid 2 CPM - 10$/day, the estimated traffic now drop significantly to 1000 impression. What the drop !!!



I jump right into my real campaign and it happen similarly. Now I can only receive 4k impression/day with 2 CPM and 10$/day instead of 10k like yesterday.

I test with the bid 100 CPM and it results only 7k impression. There must be something wrong here but I do not know how to explain and fix it.

Now I leave the this campaign as it is (2 CPM, 10$/day) since changing the bid may affect to my stats a lot (Am I right ?). I will let it this campaign to reach the spending 20$ then I will update the stat and consult for the advice.

About the bot test campaign, if everything I set up above is correct, I will let it run for 10$ (2CPM) and follow the guide to see if I can blacklist some placements.

I think I would do like you say before introducing some landers. At the moment, I cannot decide which landers to test with which offers so it may be better to find some converting offers with direct linking first then I will take on Adplexity to find some landers work with those offers and continue another round.

Now I am only have some offers left for the batch testing, I may need to sign up to some for networks to get the offers.

Do you have any recommendations for the network ?

I am now a part of Advidi, ClickDealer, Big Bang Ads, PeerFly and Mobidea.

I submitted the form for Avazu, Mobvista, AppFlood but I have not heard anything from them. I tried to contact their people through the Network directory thread created by Mr Green but have not got any respones back. I may find some more mobile friendly network to join in.

Tien

---

Popads is quite a diva. If you use too many targeting features (target to tight) you will hardly get any traffic and the CPM explodes. You should try a looser targeting.

---

Hi sebastian,

Do you have any suggestions for a looser targeting for my campaign since I am just direct linking these offers to see what convert best then optimize them with more specific targeting.

I am thinking that may be PopAds do not have much traffic for Idea carrier in India so it can be the problem leading to my low volume in traffic.

---

A quick update for my bot test campaign. It seems like I set up everything correct since I am able to collect clicks from the placements and decide whether the CTR is high enough or not.

These two placements above is suspected to have bots since the CTR is only 60% when they have a satistically significant number of clicks. I think I may cut them off quite soon if the CTR continue to decrease.

The other placements have not reach significant data, they mostly range between 10-20 clicks with 70, 80% CTR so it's hard to tell if they have much bot traffic. Since the offer page of the test is my real campaign URL so I may let it run for a while to expect some conversions.

In the meantime, I have 4 conversions after 10$ test for my batch test of 6 new offers and one offer that convert 3 times in the first round.


The highlighted one is the old offer. It has total 4 conversions and I think I gotta find some landers and test this offer alone with those landers.

---

I would test buying traffic for another carrier as well, or even wifi traffic and sending the redundant traffic to afflow.

If there is more traffic available for your carrier, and you just don't get much because of too tight targeting, you should see an increase immediatelly.

If this doesn't work, you can test some other traffic sources to find more traffic for your carrier, or use different offers that take traffic from different carriers.

---

My biggest issue I myself realise is that there is not so much impression for my campaign so that it can reach satistically significant numbers soon.

I just modify my bid from 0.002 to 0.008 and the estimated traffic increase from 3k to 6k/day.

I did try different bids to see if I can gain more traffic, but even if the bid of 0.05, 0.1, 0.5 the maximun impression is just 8000k/day.

The lack of impression makes it hard for me to move on to decide which offers are winners and also cut some placements in my bot test. I really need some help at this point.

I also look for some landers on Adplexity to test some converting offers in the batch test but I have trouble finding the offer's final landing page. Just start a thread and wait for help (https://stmforum.com/forum/showthrea...s-landing-page)

---

Leaving a post here to subscribe to the thread - detailed comments and suggestions to come!


Amy

---

The limitation in traffic volume is typical when you're targeting carrier traffic vs. bot traffic. However, if you're targeting carrier traffic, then there's no need to run a bot test. Bots is really only a concern when you're running wifi traffic. My theory is that this is because it's not easy to simulate bot traffic for carrier without using an actual SIM card from the specific carrier.

And you're right, the bid WILL affect traffic levels, as well as the daily budget. To get an idea of the amount of traffic you'll get, set your bid to something extremely high (like $1/view) but don't actually run at that level (i.e. don't save your bid), and go to the estimator to see how much traffic popads estimates you'll get - it will give you a good idea of the maximum traffic volumes available on the network. If there's not much volume available in the first place, then you'd know to forget about testing that carrier, at least on popads. If there's enough traffic, you'll need to set the actual bid to the average bid, test landers and offers until you find something promising, then test different bids to find one that will give you the highest ROI. That's the gist of the process.

If you're only getting 1k impressions/day at the average bid, then that's a sign to not test that carrier - volumes are way too low.

Also - are you finding that popads is spending the daily budget you set (i.e. the $10)? If so, then that would be one of the limiting factors, in which case you should set the budget to a higher amount.

Does your offer only accept traffic from a specific carrier? If not, then it would be a good idea to include wifi traffic in your targeting as well. If you do, then a bot test would be a good idea. And I don't imagine you'd have problems with volume when it comes to wifi traffic in India.

Lastly - the drop in traffic volume can be a result of increased competition. If there are other advertisers increasing their bids to grab more of the traffic, then there'd be less for you if you're bidding the same as what you were bidding before. Another possible reason could be that the traffic network may have lost some publishers (i.e. placements) that were previously sending some of the traffic you were targeting. No matter the reason, you'll need to make sure that there's enough traffic volume available, before spending the time and money on testing and optimizing the camp.


That would have been a good approach - i.e. to spend the least amount of time and money upfront until you hit on something promising - if not for the fact that some offers will not convert without a lander.

What type of offers are you running? If you're running 1/2-click carrier billing offers, then direct-linking (without using landers) would definitely be worth a shot. But if you're running sweeps or antivirus for example, then chances are you won't be able to assess performance without using landers - you'll need to warm up the visitors and use some persuasion in order to get them to convert when you send them to the offer.


Try this tool:

https://stmforum.com/forum/showthrea...shing-1-Button

The affiliate networks available through this tool are newbie-friendly. But I think you already have a pretty good collection of aff networks - you just need to test more offers.


Nice testing effort and results!

However, if you're having trouble reaching even 60 impressions for most placements, then that's an indication that there's not much traffic available for your targeting. Again, you'll need to confirm this before spending further time on testing. ALWAYS confirm there's enough volume first, BEFORE spending the time and money on testing. If you don't have enough traffic volume to start with, you'll be left with even less after optimizing/cutting, and then you'd be making peanuts after spending all that time and money on testing and optimization.


Nice!

Reminder: Make sure you know how to compare offers and cut down to the winner properly. Here's the tool and method:

https://stmforum.com/forum/showthrea...211#post289211


This has confirmed my suspicion that there may not be enough traffic available for your targeting. 0.008/view is really pretty high for IN traffic. It would be difficult to find an offer+lander that will be profitable at such a high bid.

One thing you CAN do, is check out other pop networks to see how much traffic volume they have for your targeting. You could either check their estimator (if they have one) or ask their support directly. If you know for a fact that other pop networks DO have enough volume, then it wouldn't be a bad idea to do your testing on PopAds (since they tend to have the best quality, just lacking in volume), find the best offer+lander, and THEN scale to the other networks. The main point is you'd need to confirm there's sufficient volume first - it doesn't just have to be on popads, but it will need to be available somewhere.


Will take a look now. You're making good progress! Please keep it up!





Amy

---

Hi Amy, Matej and other members contributed to this FA

Thanks for your respone back to my Follow Along. Before getting advice from you (Amy), I did not have many ideas on how to continue the campaign in 2-3 days because of the traffic is relatively small for my chosen carrier in PopAds. It might have been for my lack of research and determination

Though I could test PopAds first then scale to other networks to get traffic, I have made up my mind to move on with a new campaign. I am not quite sure if it is a bad idea or good idea.

Here is the detail of the campaign I am working on so far.

Vertical: Sweeptakes
GEOs: SG
Traffic source: PopAds
Bidding: 4CPM

First off, I ripped four landers in Adplexity and test with a SG offer (payout 1.7$). I did not notice the payout initially since I refer this offer from the top converting sweep offer from my AM.

After the first test, it converts 2 times with 10$ spend for two landers (let’s say lander 2 and lander 4). The other 2 did not perform well so I cut them off

Since the CTR of lander 4 is much higher so I decided to send more traffic to lander 4 (80%). And this is the result I get after the first round.

-61%ROI for lander 4 !

I think that I’ve found a good offer + lander combo. I ripped two more landers and tested them with lander 4 for this offer in another round. Also, I cut some placements that has more than 30-50 visits but 0 click as well as placements reaching more than 100-200 visits but CTR <5%.

This is the result


Lander 4 is still the winner

What strange was that the two new landers has a really high CTR. I am a bit suspicious about this since I did not clear all the JS files after ripping them from spy tool. My discussion for this matter is here: (https://stmforum.com/forum/showthrea...l=1#post319769)

In the meantime, I found two new offers (1 from ML and 1 from SG) to test with this the current winning lander (lander 4). However, they seemed not convert well after the spend of 5 times the payout, so I cut them. Then I solely focused on the intial offer + lander 4

What I tried to do was asking my AM for some stat related to that offer and she showed me the list of the best ISP, OS and time. I decided to split test the ISP first.

I seperated a campaign targeting only the top ISP. I changed the bid from 4CPM to 8CPM since the traffic estimated was relatively small. This is my initial result


I am now sending more traffic to test the ISP. At first I just tried to target my best performing ISP but I only had 3k impressions per day so I went with the targeting of 4 major ISP. I will update the result tomorrow

One more thing I do is I use the blacklist of placements I cut in the first campaign and put them in every new campaign I set.

Also, I want to ask if Singapore is a good GEO to start since a member in our MasterMind group say that the traffic is kind of expensive. Picking SG as the GEO and the offer which pays 1.7$ with a low testing budget may not be a good idea, I doubt myself...


Tien

---

A real quick update for my seperated campaign to test with ISP

4 conversions in 4 isp. It's hard to tell which one is the winner.

Something I can get from this stat is that the traffic was not sent to each isp equally. 2 isp reach more than 1k visits while the rest only has a few hundreds. I am thinking that I should set up 4 campaigns targeting each isp.

The maximum traffic I could get for one isp according to the estimator is around 6k-7k impression (I set my bid to 1$/view). The other 3 isp only has <4k impression each one. Is it enough for split testing ?



Tien

---

1/ I decide to run more traffic to the campaign targeting ISP. The bidding is the same which is 8CPM, the only difference is dayparting. Today (Sunday) I run the campaign around 11am in SG timezone. This is the first time I run at this timezone in my entire a campaign so far. All the former testing rounds are run at 6pm to midnight.

2/ Another split test I conducted yesterday was OS version. I set up a campaign targeting iOS devices, since with a bidding of 4CPM or 8CPM does not give me much traffic, I decided to be a no.1 bidder with this targeting by bidding 0.018/impression. This bidding is slightly higher than the highest effective bid which is 0.017. My theory is that if someone bids this high for the given targeting, there should be some good placements in there which can convert well. Do not know if I am right or wrong

Only 1 conversion after 10$ ad spend.

One thing I can notice when combing the results of initial testings was that I have 7 conversion for iOS totally and all of them come from iOS 10.3. However, targeting iOS 10.3 alone is not possible since PopAds notice that there is no traffic available. I wonder if I can use this pattern as an angle

3/ One interesting thing I found in the campaign testing ISP was that 3/4 conversions comes from 1 placement. I thought that I gotta do something with this placement. I whitelisted it and set it in another campaign. Though I tried to bid high 0.01/impression but I have not got any traffic from it though it is estimated that 2k impressions are available.

This moment I realized that I am still in position 2nd in the bidding queue. Just think that the no.1 bidder get all the traffic. Then I increase my bid to make sure that I am the no.1 bidder for this placement. Let's see how things go. I will update as soon as I get the traffic comes in

Tien

---

First of all - kudos on being willing to keep testing landers!

Several points I'd like to make here:

1)Higher CTR does NOT necessarily mean higher CR/ROI. I've seen MANY landers with lower CTR beat out landers with higher CTRs. Some landers do a good job at pre-selling only certain visitors - the ones that would "fall for" the psychological hooks - so that although not a lot of visitors will click through, the ones that DO click through will have a much higher chance of converting. Also, some landers use stuff like auto-redirects (e.g. after x seconds, or immediately after some event such as when the wheel of fortune stops spinning) which will sky-rocket the CTR, but may not be very effective in pre-selling the visitor. In your case thought, the preferential treatment to the high-CTR lander is justified because the rest of the landers had really shitty CTRs. But generally speaking, you should cut based on statistical significance, and not on CTR (again, unless the CTR is so horribly low that it would be impossible for the lander to end up profitable).

2)Next time, consider ripping more landers in the beginning instead of testing them in batches. Each time you test a new batch of landers, you start a new split-test, and will need to include the winner from the last test (i.e. the "test control"). So you'd effectively be testing the same lander twice. If at the beginning you'd rip 10 of the most popular landers and test them at the same time, chances are you'd end up with at least 1 decent one. Plus you wouldn't need to re-test as many landers. Another benefit is that when you have a lot of candidates in a split-test, chances are that some will do significantly better/worse (i.e. the range of performance would be wider) so that statistical significance could be reached faster. This is not always the case, but the probability of being able to cut landers faster is higher.

3)The "cutting something when it doesn't convert after 5x payout" rule of thumb I suggested, was only for direct-linking to carrier-billing offers. That rule works fairly well because carrier traffic converts well in general (unlike for wifi where bots are a concern), plus when you're direct-linking you have one less thing to optimize (i.e. landers) such that there's less room for optimization. In that case, if the offer doesn't convert after 5x payout, then it would be very hard to take it to green because 1)there are no landers to optimize and 2)there are no bot placements to cut to drastically increase the ROI.

However, when landers are involved, then it would be a different story.

When you're trying to find a promising offer+lander combination, there are 2 factors that need to be locked down - the offer and the lander. I like to do this in 2 steps: step 1 - lock down a good lander; step 2 - use that good lander to mass-test offers.

Important thing to note here: During step 1, even if none of the landers seem to stand a chance of becoming profitable, it's OK - because you could hit on a profitable offer in step 2.

So during step 1, the emphasis should be on finding the best lander you can, and NOT on profits/ROI. (Although, the offer you're using should still convert at least semi-well so you don't lose your shirt while testing landers!)


Good thinking and good testing effort! But as you've found out, when you try to get too granular in your targeting, you could seriously limit the traffic volume you get. Next time only go to this type of trouble with you're targeting something that has a lot of traffic volume.

Also - you can't make any observations based on just 1-2 conversions per ISP or whatever. You need to collect more data.


Good approach. However, make sure you have a blacklist for WIFI and one for carrier - placements that convert very badly for WIFI may convert very well for carrier.


Of course ideally you'd want to run offers with decent payouts compared to traffic costs. But you can't really judge an offer just based on payout alone.

There are low-payout offers that convert extremely well. I've made lots of money with offers that paid as little as $0.10-0.20. The CR matters a lot - and this is something you won't know until you actually test the offer. (Plus, what you're bidding will affect the CR as well.)

There's a reason why every offer is priced the way they are. If higher-payouts are always better than lower-payouts, then we wouldn't need to test so many offers - we'd simply pick the ones with the highest payouts to run.


Reminder: You're NOT trying to pick a winner here!

When testing offers and landers, you want to pick winners because you can only show a single lander and offer to each visitor.

However, when talking about which traffic segments to target, you don't need to compare them to each other - you would want to target ALL traffic segments that will give you profits - the more the merrier!

Example: If all 4 ISPs are making you profits, why would you need to find a winner and cut the rest? You get my point.

Another minor correction: It's not that "traffic was not sent to each isp equally". It's the other way around - there are different levels of traffic coming from different isps. Some isps just have more customers than others.

You could set up separate camps to target each ISP, and the main point there would be so that you could tweak bids for each ISP. But again, only do this when bigger traffic volumes are involved. You can experiment for yourself for education's sake, but in the end, you'd probably want to to avoid having to manage a ton of itsy bitsy campaigns that are each only making peanuts. However, if you'd be fine with that business model, then feel free. More testing and experience is always good as well.


So you've been restricting your traffic by dayparting from the start? Well then that would explain why you haven't been getting a lot of traffic! Since there's already not a ton of SG traffic on PopAds, why not just open it up to all hours from the start?


Again - good observation that was worth a test for sure!

Perhaps try another pop source that does allow targeting by OS versions and see if you have more luck there?


This is a common experience - whitelisting in pop camps often doesn't work as well as blacklisting. Often when you start whitelisting good placements, the traffic dries up. I don't know the reason for that either.


Tien - you're testing anything and everything you could think of. This is the fastest way to learn, and this type of attitude will get you far. Keep up the great work - not every test you do will prove to be beneficial in the end, but the more you test, the more tricks you'll find. Try to find a big geo to do your testing with, so you can collect data faster.




Amy

---

1/ Lately today I am playing with how to optimize the bid since I am used to bidding based on my instinct plus the traffic estimator, I have no rules about how to bid effectively.

Thanks to the recommendation from Amy with this thread http://stmforum.com/forum/showthread...Considerations, I find a method introducing by caurmen http://stmforum.com/forum/showthread...-The-Landscape.

I think that I've found a bidding point where I can get volume of traffic with a resonable cost and I will apply this bid with my campaign tomorrow.

2/

Oh no, I've not received any visit with this targeting and wondering why ? Is it because PopAds does not allow a whitelist with only one placement?

3/ I found a tool introduced by Amy about how to cut some placements (http://stmforum.com/forum/showthread...Banners-Part-2) but it seems like I've not got enough data to see the results. Almost all of the placements are concluded to "Keep Running" while I cut some of them after 30-50 visit without a click or <5% CTR overall in the first round.

4/ I try to rip another lander from spy tool to test with the only winning lander I am running at the moment. For all of the split test campaign set up at the moment, there is only 1 lander + 1 offer which I found in the initial test. Here is a src shot of my the best performing lander aka the only one I am running so far


There are actually two options pop in my mind

a) I should start to do further testing my targeting (OS, ISP, bidding) for my current lander + offer
b) I should try to test that lander with some new ones to find the better. I come up with this option since I see the CTR of this current lander is acceptable but the CR is not quite good, combining with the thread by caurmen about satistically significant (https://stmforum.com/forum/showthrea...nificance-quot) when the maximum possible CR (0.1%) < the minium viable CR (0.28%) with 4 CPM, Payout 1.7 and 20% minimum ROI


Tien

---

LOL - we posted during the same minute! What are the chances?

That tool is overkill for pop, because pop camps are so volatile and short-lived and the tool requires quite a bit of data before it will give you a "verdict" on a placement.

Best way would be to use rules-of-thumb - such as cut a placement when it's in loss by over 2x the payout.


As I've just pointed out in my recent post, testing landers and locking down a good one is just the first step. The second step is to use that winning lander to mass-test offers.

Right now you're pretty far from green, and trying to get there by cutting stuff will be costly, plus will leave you with little to no traffic in the end. You need a way to increase your ROI massively - and testing offers may just do that for you.

Testing lander variations would be another option as well. The point is that your current funnel (i.e. lander+offer) isn't good enough. You need to test until you have a good enough funnel, so that you're close enough to green, and then cut the rest of your way there.




Amy

---

It takes me much time to digest what you've pointed out Amy. But it is absolute worth it

For the last few days, I stopped running the campaign in SG and start to find ways to improve my ROI first before cutting things. Therefore, I come up with the two ways for me to go ahead

1/ lock the winning lander + find some more landers + test with some more offers in bigger GEO

I found two offers in PH which has a ridiculous difference in payout 0.05 vs 0.8 . I tried me best to get approved by other networks (including the STM tool you suggested me but it does not work). Only get accepted from AffiliaXe, but I do not find some suitable sweep offers for me to play with (most of them come from US UK). So I decided to go with these two offers.

I ripped other 2 landers to put in the test. So now 2 offers + 3 landers for PH

One thing I did not not feel comfortable when these 2 offers have different price:

offer 1: for iPhone 6s, only 1 carrier, payout 0.8$
offer 2 : for iPhone 7, all PH traffic, payout 0.05$

So I have each of my landers modified for two versions: version for ip6s and version for ip7

I set my initial bid to 0.0006, a little bid higher than the average bid.

This is my intial result after nearly 10$ spend

55 conversions for the for offer 1 ! The revenue is so small but the number of conversions is noticable after the first test. Just think that the payout is so small that it can convert quite easily. I know it can be hard to really get profitable with this payout but I think I can play with the data to find some hidden gems. Things like good placements where I can whitelist them for another test with new similar offers.

Meanwhile, offer 2 barely get some visits since it only targets one carrier.

For landers, here is the quick result



So excited to see one new introduced lander is winning over the current winner. I think I may spend some more tomorrow and collect more data. That is for the first campaign.

2/ I found 3 offers + 9 landers in IT and decide to go with them.

However, I notice that the advertisers are extremely strict when they set many rules such as: no 3rd party brand, no promises, no instant winner, etc, which makes it difficult for me to match my ripped landers with them. I decide to set A Flow in the campaign to match the lander with the offer

offer 1: lander 1,2,3,4,5,6,7,8,9 (the advertiser is quite easy going haha)
offer 2: lander 1,2,3,4,9
offer 3: lander 1,2,3,4

I do not know if I set up everthing in the funnel is ok or not so I put the flow here for being reviewed



My initial bid for Italy traffic in PopAds is 0.002 (once again, a little bid higher than the average bid). I will update the intial result of this test soon.


Tien


Update: I sent some traffic and realize that the flow has a problem that lander 9 has much more traffic than others and offer 2 only have a small portion of traffic. I made up my mind and closed the flow. Now I only test 3 offers with lander 1,2,3,4 to make sure that the distribution is equal.

A quick result after my test, not quite many visits since I only spend ~9$. I am conducting a test based on Caurmen's thread to find the approriate bid before sending more traffic. Gonna update very soon

---

Update

Today I try to find other similar offers for IT in my current networks to put in a campaign and continue running the bidding test but I do not find any.

I have signed up for these following networks: Mpire, F5Media, Max Bounty, YeahMobi to get more offers to test. Really hope that I will review my application very soon.

I think I need to spend this waiting time to read some Follow Along in the forum and get some more ideas...

---

Glad to hear! I tend to be long-winded and keep repeating the same concepts, so thanks for bearing with me as well.


Were you logged in before accessing the tool?


PH camp stats looking promising! Would be good to add more landers to the test. Since PH speaks English, you'll be able to find a ton of landers on Adplexity.


Feel you for sure, but it can be difficult to judge the potential of an offer based on price alone, without knowing how their CRs will compare.

I've seen lower prices work better than higher prices and vice versa. It depends on the individual offers - differences in conversion flows, how attractive the offer pages are, how saturated the offers are, and (*whispering*) how much shaving/scrubbing is going on behind the scenes...

The only way to find out would be to actually test them.


Well there you go!

I REALLY like your strategy of using low-payout offers to cut placements for cheap, and then testing higher-payout offers on the good placements.

One thing to keep in mind though: The targeting needs to be similar, and preferably the vertical (although pop traffic tends to be broad-appeal anyways, certain placements will still convert better for certain offer types depending on the topic of the placement site).

What I mean is, don't use a low-payout wifi offer to cut placements, and then run a higher-payout carrier-billing offer on the remaining placements - or vice versa. A placement that converts like crap for wifi can perform very well for carrier (mostly due to lack of bot traffic).

However, I don't share in your pessimism that it'd be hard to get profitable with this low-payout offer. You're doing -30% currently, and have only tested 3 landers. And you still haven't tested bids. Which carrier are you running? If it's the SMART carrier, there's good volume on PropellerAds, so try your luck there.


A good reminder for next time: Confirm there's enough traffic volume FIRST, and THEN find offers and landers to test on the traffic.


Good job! Once you have a winner in this round of split-testing, add 5-7 more landers and you may have even better luck!

9 landers - very nice! With that many landers, chances are at least one of them would be a decent converter.


I always like to ask my AMs for the "real' rules.

The rules stated in the offer description - half the time your AM will tell you that the advertiser is actually not nearly as strict as they sound. There would still be traffic source rules to adhere to though, unless you're cloaking.


The strategy itself is good - but you're making things complicated for yourself.

(BTW LP7 is nowhere to be seen - perhaps you had that set up and just didn't include that path in the screen capture? Also, LP9 should be for offers 1&2, not 1&3 - perhaps they were just marked wrong?)

Next time, just put a single offer in each path, and all the landers permitted for use with that offer - should be easier to set up.


Looking forward!


This is because path 3 only has that one lander (LP9). Next time if you like, just tweak the weight of path 3 to a lower percentage so the path would get less traffic.


Always good to experiment!

When testing unproven landers and offers, it's a good idea to bid average or above, just so you can be confident in the traffic quality. The logic here is there is that if the landers and offers still don't convert half-decent with such quality traffic, then it may be time to move on.

Or - you could adopt a different strategy: By starting everything with a low bid to save on costs. Doing so could lower your test budget considerably, but the flip-side is of course that you'd be prone to erroneously giving up on offers and landers that COULD HAVE converted well if you had been bidding higher.


I feel your pain for sure - but know that it could have just as well gone the other way. Thus is the nature of testing.

Also - keep in mind that IT is more of the more expensive geos, AND the fact that you're testing 9 landers and 3 offers!

Plus a part of the budget was spent on very high bids - that you normally wouldn't be running at for any extended period of time.

What's the payout range of the offers? I do believe you should give those offers and landers another chance! Especially after spending all that time cleaning up those landers.

Keep in mind that you only need 1 lander + 1 offer to be profitable for enough of the traffic (placements etc.) to have a green camp.


Ah OK! Hm...

So you've spent around 10x payout testing each offer.

I still hold the same opinion: With so many landers and considering a part of the budget was spent on very high bids, I think it would be worth running the camp further. I would throw $20 more at it. If one of the offers converts, I would JUST run that offer to see if I could start cutting landers.

If I find that that approach ends up burning a hole in my pocket, I would swap out the offers for a batch of new ones and retest the landers.

Basically in this initial stage, you need one semi-decent offer to help cut landers down to a winner. Having many landers in the test is a double-edged sword. I love it because they almost guarantee I could find a half-decent lander in the mix, but the downside is that I'd need to throw more budget at the testing.


Are you targeting carrier or wifi? If carrier, then no need to run a bot test. If wifi, then yes.

In terms of controlling costs: Stick to tier 3/4 geos and low-payout offers.


In this case you'd need to fork out the cash to do the testing. There's a Chinese saying: A needle can't be sharp on both ends. If you want to break into a competitive geo, then you'd need to give up the initial investment.

Trust me, I understand the frugal mentality when it comes to running camps. But this is something you'll need to overcome. I used to think the same way until one day I get so sick of worrying about spend, that I decided to change my thinking: Instead of aiming to minimize my test budget, I decided on a weekly test budget that I would aim to exhaust by the end of every week. Once my aim was to "spend x/week" instead of "make x/week while spending as little as possible", it lifted that pressure to start making profits, and everything turned around for me relatively quickly after that.

This is one of my fondest memories from my newbie days - thought I'd share it with you.


The PH camp has potential - but locate the traffic volume first. There would be no point in optimizing further if there isn't much traffic in the first place.

As for IT - if you're really wanting to save on budget, try another geo. But like I said, I would at least give it one more chance - since you're already spent time fixing up 9 landers.


I'm not wanting to discourage you, but $200/month would take you forever to succeed at this game.

I'm not suggesting for you to quit - don't go that route. You're learning a lot in a short time. It would be a shame for someone who has so much potential to quit.

What I AM suggesting, is for you to do freelance or get a part-time job, in order to generate more disposable income for testing purposes. There would be a lot to juggle for a while yes, but persistence will get you there.

And like I've mentioned above - stick to tier 3/4 geos and low-payout offers.

Actually - would you be interested in trying your hand at mobile display? It doesn't take much money to test carrier-billing offers on go2mobi for example, by targeting IP ranges.

Here's a rough outline you can start with if you're interested: https://stmforum.com/forum/showthrea...l=1#post299115

Don't know if I've mentioned that option before - if I have then please excuse my forgetfulness. Thought I'd mention it again just in case I haven't.

You're making good progress although the results may not look like much. This is typical when you're just starting out - the learning curve is steep and you won't see great results for a while. Hang in there!




Amy

---

Are you targeting carrier traffic or wifi? If you're targeting the SMART carrier, then like I said there's more volume on Propeller. So even if the volume on popad isn't great, you'll still have that to look forward to.

As for how much traffic is "enough" - that would depend on the traffic costs and CR of the offer on that traffic, plus your standards on the minimal amount of profits you're looking to make in order to feel that it's worth your effort in managing the camp. It would be difficult to pull out an arbitrage number that will work for all situations, but I'd say to forget about anything with below 5k impressions a day unless the CR is high and/or the traffic is cheap.

In your specific case though you already have some stats. For your best lander, for the 4000 impressions, you made $1.80 revenue for around $2.60 in spend. Let's assume you can take the camp to 30% ROI (this is basically the lowest I'm willing to go - after all there are business costs, plus preferably I would like to get paid ). So...

Profit/Cost = 0.30

(Revenue - Cost)/Cost = 0.30

Revenue/Cost - 1 = 0.30

Revenue = 1.30 * Cost = 1.30 * $2.60 = $3.40 (approx.)

Which means you'll make $0.70 in profits for every 4000 impressions. Now even for a completely newbie, I wouldn't imagine anyone would keep a camp running for less than $5/day. This means you'll need $5 / $0.70 * 4000 = 28500 impressions...

So this would be your gauge right there.

BTW - have you adjusted your costs in Voluum with the ACTUAL spend you see in popads? The costs that are passed back automatically using the [BID] token are almost never accurate, and can be off by as much as 100%! DO check the actual costs in PopAds every time - until you can estimate the percentage difference between the actual costs vs. costs displayed in Voluum, for your specific targeting options. (For example if you're always seeing around 30% difference, then you can mentally adjust the costs you see in Voluum. Either do that, or manually update the costs in Voluum.)

Reminder: Please verify that you're targeting all the intended options before checking the estimator.

And no - there's no need to constantly change your bid to stay above the average. The suggestion to bid average or above is just so you can be confident in the traffic quality while testing offers and landers. It's not a guarantee that you'll get the best ROI (by staying at or above average bid).

Once you're running on a promising offer+lander combo you can test bids. The level of competition will surely continue to change all the time, so it would be a good idea to retest bids once in a while. However, many pop camps won't remain profitable for very long, so we'll cross that bridge when and if we get there.

If you're finding that you're getting a lot less traffic volume all of a sudden - then maybe consider increasing your bid to get the testing done faster.


Most traffic sources will naturally start sending traffic at a slower rate when it's approaching your daily budget (and in this case, also when the amount you've assigned to the campaign is approaching depletion). So even without you setting up throttling, the camp would effectively be on throttle.

So if you're wanting to spend more money, you'll need to increase your daily budget plus assign more money to the campaign. Or if you're wanting to keep spending $10/day then just accept the throttling effect.

BTW - speaking of money and budgeting - I just came across this thread which you may find useful: https://stmforum.com/forum/showthrea...-(Need-Advice)


Yes - both the bid and time of day will have an effect on traffic volume.

One possible reason for the significant change in the bidding landscape, is a big advertiser (or multiple smaller ones) is running an offer that converts very well that has a cap - so that they're bidding very high only during certain hours.

This is when dayparting will help, but you'll need to run continuously for a few days at least, in order to collect enough stats to see a pattern.



Amy

---

Update:

I have a problem with the offer today when I run traffic to the offer, seems like it redirects my traffic to another offer. I see the report on my network that says my PH offer does not get any clicks, instead it is somewhat a Global Sweep (Smartlink).

One thing to notice that my most successful lander does use third-party brand (I assume many landers out there for sweep use Facebook or Google), but the advertiser does not allow so. I don't know if I've got shut down or not...

Unfortunately, my AM is on the vacation so I have to contact another one that my AM prefer me to. I will keep updating the news.

Tien

---

In Voluum stats, instead of choosing "ISP/Mobile", choose "Mobile Carrier" - that way you'll see stats for each carrier (3g/4g), and "Others" will represent the wifi traffic mostly.


Good move!

And yes - DO target wifi and carrier in separate camps. Your carrier traffic is breaking even already. Carrier traffic is usually considerably more expensive than wifi, so to avoid overbidding for wifi traffic and underbidding for carrier traffic, targeting separately would be the way to go. When you're targeting carrier traffic separately, you can bid high to get more volume without having to worry about overbidding for a ton of wifi traffic.


If this stats tool says a lander is ready to cut (i.e. above 90%), I would suggest to just cut it. Although - in recent months I've moved the bar up to 95%+ because I feel that the extra accuracy is justified given how crucial landers are to the success of the camp.

There's no need to worry about sample size if that's your concern. The tool already takes that into consideration.

There's always a possibility for the calculator to tell you to cut a lander, and then if you keep it running it may perform better and become the best lander - but because that doesn't happen very often, I wouldn't keep the lander running if the calculator tells me to cut it. Being overly accurate can get expensive - especially considering how many landers and camps you'll be running over your lifetime.


I actually have no idea!

I've never done a comparison between 1/24 hours vs. 1/7 days. That's something you can find out easily if you can afford to spend some money on testing.


Yes the idea is for primespot traffic to be of better quality and convert better - however, keep in mind that this traffic will also be more expensive.

Traffic prices are like the stock market where prices will constantly adjust themselves according to market supply and demand. The big difference is that unlike stock prices, which can be heavily influenced by perception of value, traffic prices are driven by profit margins - such that nobody would pay higher prices unless they can still turn a profit.

This results in a general trend where the better-converting traffic will cost more, and the worse-converting traffic will cost less.

Therefore, there's no reason to JUST target primespot traffic. Non-primespot MAY convert less well, but it would likely be cheaper as well. What you may want to do is target them separately and bid higher for primespot - but only do this if you're getting significant amounts of traffic for both types. If you're not careful, you could end up with a million camps all targeting little tiny segments - just not worth the bother.

Pop traffic in general is annoying, primespot or not. But people DO still click on them and go on to convert. It's good to speculate, but in the end, we just need to let the stats talk.


That would depend on the nature of the offer.

Sweeps don't tend to convert well on adult. Perhaps if you test adult-style landers you could make sweeps convert better - but that's something you'd have to test.

Stuff like antivirus, 1/2-click entertainment-type offers, and of course the typically-adult stuff like adult dating and male enhancement will all convert well on adult traffic.


Why don't you try and see? If you don't want to mess up your stats, set up a separate camp, set a budget of $2.50, and let 'er rip. Take screenshots of all the different stats, then manually adjust the cost, and then look at all the stats again to compare and look for differences.

I almost never adjust costs manually - too busy and lazy. I would just adjust/estimate in my head as I'm reviewing stats. It's not the most accurate way for sure, and I wouldn't recommend that you do the same.


Are you using caurmen's step by step method? If so, then the CTR should represent the % bot traffic, in which case you'll want to cut placements that have 80%+ bot traffic, or even 70%+ bot traffic.

It's true that you can't control how many impressions each placement will get. What I do is I would set a low campaign budget, check placement stats when budget runs out, and pause placements that have reached the required number of impressions (I think caurmen recommends 60 impressions?) Then I would give the camp another low amount to spend and repeat. Don't keep repeating until EVERY placement reaches 60 impressions. The smaller placements won't be worth your time. Just aim to weed out the bigger bot placements.


Haha thanks - but it could also have gone the other way. I've had a couple people tell me e.g. "the tool told me to cut the lander but I ran it some more and now it just became the winner". Nothing is 100% accurate and we shouldn't aim for such high accuracy, because extra accuracy requires extra test budget. As long as we're right MOST of the time, we're good.


Up to you.

If you want to be accurate, you could cut landers for EACH campaign separately just based on each particular campaign's stats. The results may very well be different due to the different targeting. However, this brings us back to the accuracy vs. efficiency dilemma - the extra accuracy will cost you. So this would be a personal decision you'd need to make.


If you've already tested quite a few landers, then just focus on testing all the offers you can find for that geo.

Testing any single element too much will get you into the area of diminishing returns. You can always test lander variations, yet, but I wouldn't do that until I've tested lots of offers.

Regarding getting enough clicks to the landers - as was mentioned a few paragraphs before, the stats calculator already knows to take sample size into account, so no need to worry about that.


Again - it would be up to you how accurate or efficient you want results to be. More accurate = more expensive, less accurate = less expensive.

Is it true that different landers will perform differently for different geos? For sure!

Ideally, for maximum accuracy, you would retest all landers for each geo - but that would also be expensive.

So how much would you be willing to compromise on accuracy to save money? Again, that's a personal decision, and one where experience will help. But then every vertical is different as well. For antivirus for example, I've seen the same landers "win out" consistently for different geos. For sweeps, it would depend on lander angle.

At one time I was asking survey questions on the lander, that were on trendy topics specific to the geo. In that case of course you wouldn't want to use the same lander in other geos.

However, landers like the spinning wheel, I've found to work well across many geos.

So - if you're wanting to be accurate and efficient at the same time, I would suggest to 1)stick to non-geo-specific landers, and 2)test all popular landers on adplexity for the first 5+ geos (preferably from different continents) and see if you can spot any lander that converts consistently well for all geos, and then use that for all subsequent geos.

Whenever you find a good offer with this one-size-fits-all lander, it wouldn't be a bad idea to allocate at least a part of the traffic towards testing other landers for that geo.


Oh WOW! Happy day indeed! So pleased about your progress - and congratulations! More and more green to come no doubt!


You can test bids - you're already down to your best lander.


Why let the throttling warning annoy you? If you're wanting to spend more, just allocate more balance to the camp and set a higher daily budget. If you're concerned about spending too much, then be content with the lower spend.

Limiting your budget is a good idea when you're testing stuff and need to check for statistical significance, or need to cut placements. The limited budget will make sure you don't overrun (i.e. keep running stuff long after you should have cut it).

If you're already down to your best offer and lander, and have already cut the worst and biggest placements, OR your camp is in green or close to breaking even, then just allocate a larger campaign balance and set a higher budget - basically set them to something you won't exhaust in a day. The traffic estimator will tell you how much traffic you can expect to receive, so that would be a guide to how much you should expect to spend each day.


Perhaps the offer ended or cap was reached?

Yeah you're correct in alluding to the fact that the best-converting landers are usually misleading/aggressive. There ARE offers that will accept aggressive marketing - just ask your AM for those if you're worried about getting caught.

This is why so many people are cloaking - so many restrictions from the advertiser and the traffic networks. I'm just stating a fact, not encouraging anyone to cloak (but then - not discouraging anyone from doing so either).

Always get your AM on your side if you're doing something shady - most of them will help you out as much as possible without pissing off the advertiser to the point of non-payment. They know which advertisers are more/less strict so will help you to navigate the treacherous waters. If you keep shady stuff from your AM and then get caught, then they have no reason to pay you.




That's all I have for now! All the rest of your observations and decisions sound great! Keep up the momentum!



Amy

---

Targeting wifi and carrier separately is great! And yes if you target them both with the same bid then of course you would get a lot more wifi traffic than carrier. This is because there is less carrier traffic available in general.

But why are you bidding so high? In both cases you're bidding much higher than the max bid.

It would be good to clone the camp and test different bids as well. Different bids will trigger traffic from different placements. You can identify profitable placements at every bid. If a camp has enough profitable placements for it to be worth your efforts in keeping it running, then cut the unprofitable ones. If a camp doesn't have enough profitable placements, pause the camp.

The rest of the settings look fine. Eager to see how they'll do, but please keep an eye on the spend and ROI, seeing as how high you're bidding!



Amy

---

Ah OK got you! That would be fine then.


How many conversions a placement has is irrelevant if it has received a significant amount of traffic but is still in loss.

Could you please cross-check those placements in Voluum to see how much they are in loss by? I can bet you that they're all in loss by many times the payout of $0.05.

I normally wouldn't even check placements for statistical significance using the spreadsheet when it comes to pop camps. Pop camps don't stay profitable for very long, so speed is of essence. I would just cut anything that's in loss by 2x payout.

The spreadsheet requires a lot more data in order to reach a verdict, compared to the rule-of-thumb of cutting at 2x payout in loss. In other words, when the spreadsheet tells you to kill something, you should kill it - because there's next to zero chance of the placement ever becoming profitable.

Of course, the spreadsheet is not a human - it will only calculate based on the current set of conditions. So, if you change the bid, or test new offers or landers - then a "kill" placement can turn into a profitable placement.

Don't put your faith in a placement just because it has made some conversions. Put it this way: If every time you give me $10, I give you $8 back, would you continue to give me $10 again and again? Well, it's the same sort of thing here.


It's definitely good to give these metrics some thought, like what you're doing!

However, setting rules of thumb like cutting at 70+ impressions without a click, or at 100+ clicks without a conversion, has one problem: Neither of those are a direct representation of how profitable a placements can be. They tell you nothing about the cost of the traffic, the conversion rate, or the payout.

For CTR: Instead of establishing a one-size-fits-all rule-of-thumb (e.g. 70+ impressions without a click), a MUCH better way would be to compare the CTR of each placement against that of other placements in the same campaign.

For example, If for a specific campaign, most of the placements have CTRs that fall into the range of 50-100%, and you see some placements with <5% CTR (or even <10% CTR, depending on how aggressively you're wanting to cut placements), then you can tell that something is off - and cutting based on CTR may be justified.

As for X number of clicks without a conversion - that would heavily depend on traffic costs, payout, and CR. I would highly suggest that you NOT use this as a metric, but instead cut when the placement is in loss by x payouts. Remember that your goal is to cut UNPROFITABLE placements, not LOW-CR placements. If the traffic cost is low enough and/or payout is high enough, even low-CR placements can be profitable. So I'd suggest not to use that as a criterion for cutting.


Congratulations! Usually for pop camps, when you whitelist placements you'll see a decrease in traffic volume or even dry up completely, so blacklisting usually works better. So I'm glad you're able to hit on an exception to find some profits!


You can of course do both.

Your idea of using the best-performing placements to test new offers for cheap is GREAT! However, I would encourage you to add more good placements to this whitelist camp first - otherwise the testing may go slower than you have patience for.

As for raising the bid - you can do that gradually (for example once a day at the same hour each day) until you start seeing a drop in overall profits, then revert to your last bid (the one that got you the highest profits).


Correct and correct!

However, just base your decisions on your own stats, and take actions that can maximize your profits. You may or may not be able to afford to pay as much as the top guy - you don't know what they're running and how they're running it, or what their profit margins are.


No - it's not the case that every placement is charging the same cost.

Look at the whole thing from the traffic source's perspective: Their goal is to make as much money as possible, so they'd want to sell most of the traffic to the highest bidders where possible, and the rest to the lower bidders (both to make profits and also to keep customers happy).

Now imagine this: There are hundreds of other advertisers bidding on traffic from the same geo (let's ignore finer targeting such as OS and carrier for now). They're all bidding differently. They're cutting placements that don't convert well for them. Some will cut bad placements to leave good placements running, then raise their bid to get more traffic from these good placements. Some will do exactly what you're doing - set up whitelist camps and bid higher to get more traffic.

The result? The better-converting placements have more advertisers bidding higher for, and the worse-converting placements have fewer advertisers bidding lower for.

So what does it mean for YOU? If you bid lower, you'll get more traffic from the unwanted placements and less traffic from the sought-after placements. If you bid higher, you'll get more traffic from the sought-after placements and less traffic from the unwanted placements.

I'm probably simplifying the bidding system by a bit here. But the overall idea to keep in mind is that when you bid differently, you'll get differently percentages of traffic from different placements, and also get traffic from different sets of placements.

A simple optimization strategy would be to clone your original camp several times, and bid differently for each from low to high. Then run them for a bit to see which placements each of them is getting traffic from, and how many of the placements are profitable in each camp. If a camp has quite a few profitable placements that are jointly doing good profits, then continue to cut unprofitable placements. If a camp doesn't have enough profitable placements to be worth your time in running it, pause it.

There are other ways to optimize, but this approach would be a good start.

Let me know if you need further clarification!



Amy

---

Hi Amy

Thanks your insights, it really help me for my next steps in this campaign.

As you state above, today I do 3 following things.

1/ I take the stat from my Android campaign and add some more placements in my whitelist campaign. Also, I block my placements that have significant loss though they generate conversion.

Sure I do not want to give someone $10 again in this situation, but except you since you've helped me so much for my learning LOL

I blacklist and whitelist placements based on this input

Confidence: 80%
Min ROI: -20%
Min CTR: 2%
Payout: 0.05
Min Spend to cut: 2

2/ I've cloned 3 campaign from the Android campaign and set them with 3 different bid

-Clone camp1: 0.0004
-Clone camp2: 0.0008
-Clone camp3: 0.0017

I will let them run for and while and report the stat.

I have a question that when we clone campaign and adjust different bids. Because I've drilled my campaigns to Android - wifi traffic, when you mean my original campaign, do I have to go back to my very first campaign when I've not split tested anything to clone or I can do it with my setting above (which is cloning the Android campaign while choosing all the connection from carrier to wifi) ?



Tien

---

Update morning 11/8

1/ While continuing to blacklist placements in clone camp3, I also try to add these good converting placements (according to the tool as "whitelist") to my whitelist campaign. I am trying to get more traffic for my whitelist campaign. But the bidding between my whitelist campaign is slightly lower than my clone camp3, so if I add some placements in clone camp3 to the whitelist campaign, is there any chance that I would get traffic from them ?

2/ I have added 2 new offers which accept only carrier traffic (Globe and Smart in PH) to my white list campaign. I set Voluum rule for my campaign.

Is >>> Mobile carrier >>> Smart , Globe >>> Lander 7 >>>offer 1, offer 2

So if the traffic is from these two carriers, I want to send these two new offers to them. And if the traffic is not from these two carriers, they will be sent to my original offer. Is this setting correct ?



Updating....


Tien

---

Fantastic! Let me PM you my bank account details....

Seriously - I'm very glad you're patient enough with my long-windedness. And posting a follow-along is the best way to get your value out of your STM membership - plus you're contributing to the community as well, as other members are following alongside you. So - I would like to say thank you as well on behalf of STM.


Question: Why would you set the Min ROI to -20%?


I would suggest to just take the camp that already has the most-profitable targeting first (so this would be your Android+wifi traffic camp) and clone that to test bids. Of course if based on the carrier traffic you've received so far, you feel that it has a chance, then you could set up separate camps to test bids for carrier traffic also. Carrier and wifi traffic should be targeted in separate camps anyways to avoid overbidding for wifi and/or underbidding for carrier.


This is what I mean whenever I advise people not to optimize so much, that they run into diminishing returns. The bigger gems will always be discovered first, so by whitelisting those you'd get the biggest benefit. Then as you whitelist smaller and smaller placements the benefits you'll get will become less and less - when the effort of analyzing stats and adding placements to the whitelist camp will remain the same.

Right now you're still in the learning stage, so over-analyzing and over-optimizing is good for learning purposes. But when you start mass-testing and mass-launching camps in the future, always keep in mind the 80/20 rule by asking yourself: Am I spending my valuable time trying to squeeze a few more pennies (or even single-digit dollars) out of this campaign, when I could be doing other stuff that has potential to grow my profits by leaps and bounds (such as testing new offers or scaling profitable offers, or scoping out new traffic sources, new verticals, new traffic types...)?


I can't explain it either - I never bothered to ask PopAds support about this - perhaps you could if you like? Maybe they could shed some light on this.

Actually - if you want a rough estimate of placement volumes for the biggest placements, you can do this: Whitelist the placements you're wanting to find out max available volumes for, bid higher than max bid, allocate a limited amount of balance to the campaign (IMPORTANT!), then keep refreshing your stats until you start to see traffic come in. Wait 1-2 minutes or more, then pause the traffic. Then you'll have an estimate of how much traffic you can get from that placement per minute, and can do the math the estimate the daily volume (it would be a very rough estimate, because volumes won't stay the same from hour to hour - but at least you'd have an estimate).

This would be harder to do if you're not a VIP on popads, i.e. your camps are still being manually approved instead of being auto-approved immediately. Set the camp up during their work hours and it should get approved pretty quickly - at least from what I remember from pre-VIP days.


Well you can play around with the estimator - specify your bid, put placement IDs in the "Included" area, and look at the Bid Distribution Map to see how much traffic you'd get in total from those placements.


Yes of course you'll need to apply other targeting, depending on what your offer requirements are and what you know to be converting well etc.

A placement is essentially a website where the owner has implemented code from popads that would make ads pop up/under on their site. The website's visitors could be from multiple countries, using multiple devices/OSs/carriers/ISPs/browser etc. etc. So by targeting different subsets of that placement's traffic, you'd get different traffic volumes.


Why would you set Min ROI to a negative number instead of something like 30%? Your ultimate goal is profits, correct?

If you're still testing landers and/or offers, then setting your target ROI at a negative number may be justified - because at that point you're not wanting to cut all unprofitable placements - you just want to cut the worst budget-drainers and still get enough traffic to test your landers/offers. But if you're already down to your best lander+offer, and there aren't other major traffic segments that are converting badly (OSs/browsers/etc.), then you should set the target ROI at at least 30% and just cut away (or, better, just cut placements that are in loss by more than 2x payout for example).


Good call! Don't waste your time with anything that doesn't have profits potential. Your time is valuable.


For simplicity's sake, I would just leave the placement running in all camps where it's profitable. Again, I don't like to over-optimize my camps - I like to focus on testing and scaling instead, because these are tasks that bring the greatest returns typically.

You could try to see which bid is doing the highest ROI for a specific placement, and blacklist this placement in the other bid-camps except the highest-ROI one - and see if the remaining camp gets more volume from that placement or not.


Wifi and carrier - both types of traffic will work. Wifi will take more cutting, but the volumes are a lot greater. Less cutting for carrier, but volumes are lower compared to wifi. There are offers that convert better for wifi and there are offers that convert better for carrier, and then there are offers that will only accept carrier traffic. In the end everything comes down to testing - can't say in general which type of traffic is better. But if you have a small budget, then it would be good to target carrier, because wifi will take more upfront investment just to identify and weed out bad placements.


For sure for sure for sure! So far so good!


Sometimes yes!

This is why some people would just clone a profitable camp and run both at the same time without changing anything - sometimes you WILL get additional traffic. It depends on the traffic source.


Looks correct! Anything that doesn't fit that rule should go to your original (default) path. You must have received traffic by now - check your stats to make sure the rule is working as intended, i.e. that the correct traffic is going to the intended offers.




Amy

---

Hi deetei,

I appreciate letting us follow along with your journey. It definitely helps me to learn, and soon I'll be starting my own follow along. Thanks so much!

---

Hi falcon,

Though I have not done anything big but starting a follow along helps me learn a lot and you should definitely do the same

Good luck with your journey


Tien

---

When comparing performance between offers, it would be appropriate to ONLY look at the timeframe starting from when you added the new offers, instead of looking at the entire lifetime of the campaign.

Specifically - we want to know whether offer 2 is doing better/worse ROI than offer 1. Based on the screenshot it looks like offer 2 is doing 25% higher ROI than offer 1 - is this REALLY the case? We need to judge based on the period when you were running both offers simultaneously.


Yes this will sometimes happen - it would seem that the more placements you're targeting, the more the traffic source likes the campaign and the more traffic it will assign to it. This is why very often, when people start a whitelist camp, the traffic dries up (compared to the traffic the SAME few good placements were getting in the original camp).

A better approach may be to just blacklist instead of whitelist. So instead of whitelisting the few placements, just blacklist as many of the bigger placements as possible and leave the few good ones running. That way you'd still be targeting a bunch of placements that are (hopefully and ideally) so small that they won't be giving you much traffic in total (many of the smaller ones would just give you e.g. 1-2 impressions/day). So it would look like there are a lot of placements in the camp still, but you'd only effectively be paying for the few good placements.




Amy

---

Hi,

I've paused the campaign temporarily since I am running in the cash flow issues

I will try to get more disposable income and come back as soon as possible.

Tien

---

Home > Mobile > Follow-along Campaigns