Home > > Newbie Follow-Alongs

---

Experienced, media buying Freelancer switches to CPA marketing (28)

---

I thought it is about time to do my own follow along now
I’m here on STM (Best decision!!) for 1,5 months now and learned so much it’s unbelievable.
So first and foremost I need to thank you guys for helping us newbies so much!

I’m a German freelancer, running Facebook campaigns for clients, but I’m done being stressed out all the time and being there for my clients 24/7.
I had the plan to switch something different for a few months, but didn't exactly know what and as I finally made up my mind in November last year, I knew I wanted to do affiliate/ CPA marketing.

I do not have any experience in this field, but of course I do have experience running FB ads, but still I decided to make some success with Pop first.

I did the 1$ and the 40-day guides (they were super helpful! (thanks amy)), switched from Funnelflux to Bemob and keep on running campaigns day-in, day-out to learn.

My main affiliate networks are:

• Clickdealer
  
• Big Bang Ads
  
• Mobipium
  
• Haka

My traffic sources so far:

• PopAds
  
• Propeller Ads

My Vertical:

• Sweepstakes only so far

(Mainly tried iPhones & Samsung phones, but testing vouchers and giftcards as well)

Main Geos:

• UAE
  
• Saudi Arabia
  
• Thailand
  
• Malaysia

Toolstack:

• Tracker: Bemob
  
• Landers: Landerlab.io
  
• Adplexity
  
• Cloudflare

I try to avoid making most of the newbie mistakes, as far as I read on STM, by sticking to Sweepstakes, offers with PO < 1$ & Tier3/4 Geos.
I had and still have some trouble getting the tracking working properly, but with the support teams I’m getting results into my tracker step by step.

Like Amy recommends, I’m using a campaign journal to keep track of everything and this is so helpful, can’t even image going without it.

I think it also provides the best overview for a follow along to show you what’s going on:
https://docs.google.com/spreadsheets/d/1yku-8i_exz0gN0c5wi-F6a_lCBeGhNbGxnDUBSI2Bjs/edit?usp=sharing

Currently I’m averaging -94% (but because of the tracking issues and switching trackers in general I lost some data)

With 2 campaigns I managed to get 4 conversions, one testing 7 offers with 3 landers and an average PO of 0.33$. I didn't spend 90$ though, but only 30$. All offers were at -90%, so i paused at 30. Maybe that was way too early:





I found, that one of the offers was doing better with 2 Conversions, so I created a new campaign with that offer and 5 landers to test with. Also got 4 Conversions and two of them with one lander, now considering it being my best lander:



Still -77% after 12$ spend. I know the formula offer x landers x PO x 10 and this equals to 22,50$, but with -90% I thought I could stop the not-performing landers anyway.

Now I am very curious, if I am making any major mistakes here or if my setup is ok and what I should have done here and there 😉

My plan is to stick with Sweeps, but not iPhones and Samsungs, but maybe something not everyone is advertising...

Thank you for everyone, helping me along the way, already!

---

Wish you much succes @florakija or how you would say in Germany Ich wunsch dir viel Gluck!

I am curious how you will do. I am sure with your experience you will make a good progression soon!

Here are a lot of very helpful moderators. So you are in good hands!

Maybe you could check my FA out. This could answer maybe a lot of your questions. As I am a newbie also!

Here the link of my FA:

https://stmforum.com/forum/showthrea...table-Campaign!

Well... Auf wiedersehen!

---

Very cool stuff!

Are you planning on trying facebook too, due to your experience with it? Or going to stick with pops/push for awhile?

---

Thanks man. I’ve already read that! Great stuff there vielen Dank


Sent from my iPhone using Tapatalk

---

Thanks. Actually I tried BH FB stuff and of course stuff bannend, but it was 100% my fault. Not ready yet to step into the hassle of buying accounts and stuff.

I know it’s competitive, but I really like pop ads somehow.l and it’s great for learning the basics of CPA marketing


Sent from my iPhone using Tapatalk

---

Nice... very cool man

---

Very nice start for a follow-along @florakija!

I'm guessing you didn't use a statistic calculator?

Also: Why did the lander with the 2 conversions get twice the amount of traffic the other landers got?

As well: Try to drill down to placements/zones and pause out the ones that are in loss by 2x payout. When you've cut down to your winning lander+offer combo, you can give those zones another test (because they may perform better with better lander+offer). But for now, pausing those can curb the bleeding.

Let's start with this and see what the ROI of the best offer+lander ends up being. And once you have a winning lander, you can use that to retest the rest of the offers (the ones that didn't get enough traffic before) and also new offers.



Amy

---

No I didn’t use a statistic calculator yet. Need to learn how to use it

I ran the other on another campaign as well, that’s why it got more traffic.

About cutting zones, I thought I read somewhere to first get a decent offer + lander combination, before I start blacklisting placements and stuff.

Even though I did it anyway already, but too early I guess

Im still having a lot of problems with the tracking and postback urls.... even with the help of the support, but need to figure it out on my own anyway.

Set up some campaigns today and let you guys know how it’s going soon

Thank you already!


Sent from my iPhone using Tapatalk

---

UPDATE - 08.03.2021

So the last days I did more client work for FB ads to keep the income steady and I checked, that I have spent 1k for half of January and whole of February including tracker, STM & Adplexity resulting in -98% ROI. Hope that's not too bad for a newbie.
It's definetly not good either

I plan to spend 500$ a month without the tools to keep on learning and to finally get my first green campaign.
I also read through a lot of propeller ads blog and their webinars. I found them very helpful.

Campaign-wise it's almost always the same:
I test different offers with different landing pages to find out a winner. When I get 2+ Conversions I turned the campaigns off, checked the stats in the tracker and set it up anew with optimized placements, like Android only or region XY, if I found that it converted best.

BUT for the optimized campaign, basically nothing happens anymore.

Haka direct linking:


Haka winners:


* * * * *

Clickdealer offer test:


Clickdealer winners:


Maybe I'm testing with too less budget or optimizing too early, without having a winning offer yet...?
Maybe I'm judging too early on stuff.

This was still my best offer, converted 4 times, but...


after letting run only this one offer, I got 7 conversions, but still -81% ROI.
Not sure what would have been the right approach to go from there.
Would have been a good idea to test this very offer with even more landers?




* * * * *

Also I do have some very newbie questions:
- What are important KPI's to look out for in the tracker? (CTR, etc...)
- And what are good stats for each KPI (eg: good CTR on pop ads to look out for)
- I read everywhere to stick with one vertical, one traffic source and even one geo in the beginning, but also to test offers like crazy.
This seems very hard, because I dont find enough offers for that strategy. Maybe I got this wrong in some way
- How do I white/blacklist publishers or adzones on PopAds (or any given traffic source for that matter), whats the best approach here?
- Do I have to mess with the targeting optins in PopAds yet? ("general/imagehosting" and so on)

* * * * *

What I will do this week to make progress:
- Focus on building great landers, including JS alerts, implement more tokens to personalize the landers, and such
- Learn more HTML & JS, because landers are so important
- Maybe even start some push campaigns to test out


* * * * *

So that's it
Thanks to everyone in advance who will have some critical comments on what I did, but also on what I am going to do this week and if I am going in the right direction or completely wrong!

---

I wanted to reply specifically to this question in your last post:

==

- I read everywhere to stick with one vertical, one traffic source and even one geo in the beginning, but also to test offers like crazy.
This seems very hard, because I dont find enough offers for that strategy. Maybe I got this wrong in some way

==

If found that sticking to one vertical is helpful, and you can stack offers in the same vertical on one landing page to bump up your revenue. You can't do that if you are working different verticals.

Sticking to one traffic source but multiple geos has worked for me. It's pretty easy to tick some boxes to add multiple geos, and then you'll see which geos convert the best. Once you have that data then you can cull the lower performing geos. Then when you add another traffic source, you have a good idea which geos to start with.

I hope this helps!

---

Thanks a lot. Every hint clears out my way a little


Sent from my iPhone using Tapatalk

---

The last days I have paused a little with trying to make pop traffic work for me, so I thought I'll test one push campaign out of interest, because i never did one before (I know. Shiny object syndrome is not good ) BUT I got 7 Conversions right away!

You can imagine I was really happy having -36% ROI for the first time.

Screenshot:


The only problem was, that the tracking didn't work on Bemob (found out that there's a separate template for push campaigns and the standard propeller template doesn't work).

Screenshot:


Now I have set up the same campaign again, but added 1 more lander and some more creatives to test even more stuff.
Didn't change the targeting though.

Targeting being:
- No broker traffic
- High & Medium user activity only
- Country: MY
- Bid: 0,02 (CPC); recommended is: 0,039
- Daily Budget: 20$; Overall Budget: 40$
- Mobile & Android only

These are the stats after burning through half my budget (20$):

Screenshot:


Doesn't look as good as before, but at least the tracking works now.
Well, everything despite the cost.... Still can't figure that out.

I think I'm gonna stay with push for a while. I kinda like it.

So let's go to the questions part
- Should I change something already in this campaign? (optimizing)
- Would you consider the numbers (CTR, etc.) good or what are good stats for pop/push?
- Is the formula for push budget also: 10 x PO x Landers?
- Did I do major mistakes on this campaign (never ran push before)?

Thank you already, guys and girls

---

hahha, you should know that its the same here. you should be 24\7 with your campaigns, at least I do that.
I like your post, I also started with similar offers and you made a good choice - its very easy to start with push and pop traffic these days, i think.
I also test propellerads and popads now, but with mobile subscriptions, antiviruses and sweepstakes sometimes.
You'll get good results soon!

---

Hey calipso,

You got a point with being 24/7 at campaigns. I can see that already, but it’s still different in my opinion.
But I just started out, so I don’t really know yet.

I hope I get good results soon
Your post motivates me, thank you!


Sent from my iPhone using Tapatalk

---

Time for another update on my journey

After some more trial and error I finally got a green campaign with 15% ROI (1,50€ lol) But hey, it's something.
I actually focused on push and got way better results right away then on pop, so I sticked with it for now.

With this campaign I started to test one of the clickdealer top offers in Germany, because I thought maybe I try a higher tier country and I'm a native German, so I got the language advantage here.

I started with 2 landers I have used before (in tier3 geo though), PO being 3$, so my initial testbudget was 60$.
Is this enough for push?

I ran it for 2 days and got -40% ROI. 1 Lander got a double as high CR, so I deactivated the other one.
Also deactivated 2 of the 5 creatives, because worse CTR:

Was this a right decision?

Then I duplicated the campaign with the following settings:
I found that in only one zone I got 4 conversions. I aggressively cut ALL other zones. I also only kept 7 regions out of 16 in Germany.
Only Android 10 & only Chrome 86-89 was being kept.
Is this approach too aggressive?
Should I at first test more landers, before even thinking about cutting placements, zones and everything else?


The duplicated campaign started slowly and after half a day I got 1 conversion on Propeller Ads and Bemob:

_______________


BUT I got 3 conversions on Clickdealer:


which actually means, that I'm profitable right now!

I asked the Propeller ads support and they have said, that it can take up to 24 hours for conversions to be tracked....
Is this true or is my tracking broken?

I also added 3 new creatives to the running campaign, because the only creative left has a CTR of 0,76%.
Is this the right approach or should I never add creatives to a running campaing, now in moderation again of course.

I also got some other questions:
- Propeller ads charges a fee for depositing budget. Can this be avoided?
- How can I optimize bids. I read that in Propeller Ads you can somehow see how many bids you win and then decide either to up or down your bids....
- What CTR's for landers, offers and creatives on push to look out for? What are considered bad CTR's?

Thank you everyone

---

Lots of progress has been made! Congratulations on the first green! That's definitely a milestone!

Sorry for the lack of response - I was away for a couple weeks and still playing catch-up.

I'm quite concerned about your methodology here - can you please elaborate further what you mean? 2+ conversions in most cases aren't enough to base any optimization decisions on.


None of the "winners" made a single conversion - am I missing something or did you attach the wrong screenshot?


The "winner" is at -77% ROI - that's a long way to green. Were you testing landers at the same time as well? Would be good to see the ROI of the best offer when run with the best lander. Would also be good to collect a few conversions and then look at placements stats.

If you were already using your "winner" lander after split-testing a few good landers, and still getting that kind of ROI, I wouldn't suggest to try optimizing that all the way to green - would take too much time and money, and chances are you won't have much traffic left in the end.


Testing more landers MAY be good, but if you've already testing a few before, and if they were popular landers that were optimized for speed, then I wouldn't test more landers at this time - yet.

If you can find more offers to test for that lander, I'd suggest to do that.

Or, find another type of offers to test for the same geo - e.g. sweeps offers with different prizes than what you've been testing - but of course then you'd need to test new landers for them.

Or - try another geo altogether.


-Important KPIs: Basically first step is testing a good number of landers and offers to find a combination that has the best ROI. Then you need to drill down into various stats variables and ask yourself the question "do I have enough room to cut my way from the current ROI to green?"

E.g. If your ROI is quite low, and you can't find one or more big traffic segments that are bad performers and cut them, then don't try to optimize - test more offers and landers first.

E.g. If your ROI is quite low and you're not getting a lot of traffic volume in the first place, there's no room for cutting - don't try to optimize.

CTR is worth keeping an eye on - e.g. if a placement/lander has significantly higher/lower CTR than the rest, it may be good to find out why (if it's a lander) or just cut it (if it's a placement). Overall speaking though it would be better to look at ROI, because landers/placements that have lower CTR can result in higher CR or ROI, so it may not be a good idea to judge something based on CTR alone.

-Good CTR on pop ads - can't tell you - varies widely between geos and verticals and specific landers.

-For pop, there's less of a need to stick with one vertical and one geo, because it's so easy to just rip landers and test offers. But yes it would be good to stick with a good-quality source for testing - PropellerAds is good for this.

There ARE advantages to sticking with one vertical and one geo, because then you'll have proven landers and know what the good and bad placements already, which will save you money when testing any new offers. Keep this in mind, but don't let it become a limitation.

Also: You can always sign up to more affiliate networks!

-Are you talking about adzone-cutting criteria? I like to cut when a zone/placement is 2x payout in loss, because I find that it'd be more difficult for a placement to "catch up" to turn green when it's behind by 2 conversions. For the bigger placements (the ones with more volume)

-As for targeting categories on PopAds - don't bother. PopAds already has little traffic volume. Also, pop traffic is broad traffic in nature - categories don't really matter in most cases anyway.


Will reply to your other posts below...



Amy

---

I wouldn't call that shiny object syndrome - push traffic is a good thing to test! It's not that big of a step from pop.


The cost parameter for PropellerAds is {cost} - just add that in Bemob's traffic source settings for PropellerAds:



For "Query Param" just put "cost" (without quotes).


You do that! And be sure to read all of @twinaxe's posts on push!


Hard to say without seeing details on creatives, landers, offers, and placement stats.

Budget would be different now because you'll need to take ad creatives into consideration (which you didn't need to for pop) - there's an extra variable to test now.

Will ask @twinaxe to take a look at your last post - he's the push expert!



Amy

---

Reporting for duty

Hey Flo, at first congrats to your green campaign.

Doesn´t matter if the numbers are low or not, every green feels alot better than a red campaign

I don´t know many other languages that well but so often I shake my head when I see the German translations other affiliates are running so maybe a good idea to do so.

There are also many good offers available for DE, I would even go so far to say that DE is one of the best geos worldwide for SOI sweeps.

What LPs are these?

Spinwheel and giftbox?

If yes also test a questionnaire LP.

Only one offer?

Then it´s enough.

Don´t only check the CR (but also don´t only check the CTR), chedk which lander generated highest profit.

A double as high CR wouldn´t get you anywhere when the CTR would be only 1/4.

You mean 2 out of 6

Well, with only 2 conversions altogether it´s hard to make decisions but based on what you have it´s ok to deactivate them.

But same as with LPs, CTR is not everything.
It´s only important how much profit you make.

I also recommend to run first tests without a big image.

That way you can test faster, safe money and it´s absolutely enough to do so to check for potential.

When you see success you still can splittest with big image vs without big image.

So you run a WL campaign with only that 1 zone or what?

How many conversions has the campaign altogether?

You don´t do yourself a favor when you target that tight, especially when you don´t have enough stats for such decisions.

To be honest but I can´t remember only one campaign where I targeted by regions (expect the offer only allows specific regions).

I also never target by Android or browser version.

Mostly you receive the most traffic from specific browsers or versions and these are the elements that have to be profitable.

I don´t care if browserversions or OS versions that receive only 3% of the traffic are profitable or not, it´s only important that the big elements are profitable.

Don´t make the mistake and optimize the campaigns to death.

Yup, too aggressive.

First task is to get a converting funnel (creative -> LP -> offer) running, then you can think about other things.

About zones, just exclude zones that burn through the budget without conversions.
That way you can control better that your traffic is diversified across more different zones so that a single zone doesn´t have that much impact on the overall campaign performance.

Test these excluded zons then again when you have a good working funnel.

Again, is only the one zone with 4 conversions in the campaign or do you run more zones there?

Compare click IDs to see exactly which campaign triggered the conversions on which day.

Can happen that you receive late clicks or late conversions so that you see them in the CPA network for "today" but in the tracker for "yesterday" or so.

Propeller Ads is not Bemob so even when it can take up to 24h for them to track a conversion it still should appear in Bemob.

I personally prefer to test new creatives in a new campaign.

It depends on what option you chose.

When you use credit card or Paypal or so then it´s normal to pay fees.

With wire transfer you often can avoid it although then it can still happen that you have to pay fees to the bank for currency conversion when you send money from a Euro account to a Dollar accunt for example.

All such estimators are only this: estimators.

Use it to get a rough impression but then the best way is to just run traffic and make decisions based on the results.

Too low volume: bid higher
Enough traffic and conversions but negative ROI: bid lower

Of course this is just very rough but it´s just to show that it´s easier to decide on bids when you have stats to work with.

Also check the different zones, they can also perform pretty different and can work on different bids.

No general or cut in stone numbers there.

As I said above, don´t just focus on one metric like CTR.

100% CTR are worth nothing when it has 0% CR.

Btw, what activity levels did you choose and is it only web push or also in-page?

And please check my sig, there´s a little tutorial how to properly insert images in posts.

---

Hey Amy,

thanks for the great tips!
I indeed inserted the wrong screenshot regarding the Haka winners...

You mentioned, that you wouldn't consider an offer a winner yet, after 2 conversions.

It's just that I read it here on STM, that after 2 or 3 conversions an offer can be considered to be tested against landers and worth spending money in general, because it seems to have potential.

But maybe I got that completely wrong, because let's say an offer-lander combination has 7 conversions splitted on 2 landers, ROI is -80% it can not be considered a winner, right?
And even with 3 or 4 conversions, but ROI -70% and beyond, an offer can also not be worth testing anymore?

Did I get you right, that if I do not have an offer-lander combination, which is in the greens, I should not cut placements and optimize.
I should rather look for an offer+lander, which at least breaks even?

Thank you Amy

---

Haha, thanks for helping!
Didn't think it would take such long time to become green lol. I think I underestimated this industry and yeah, green does indeed feel better than red

Cool! So I can focus on German offers a little and use my advantage here!
Those landers are for the most part GRAUSAM indeed

I have used a questionnaire and a spinning wheel lander. I could have tested a gift box lander then as well.
It was only one offer, so I estimated the budget correctly.

I saw, that you have somewhere said, if after 50% of the budget you get no conversion you can stop that test earlier.
What if I do get some conversions, but still like -70% ROI?
Should I keep it running until the budget has run out or start optimizing already, or at least adding new Landers/ Creatives?

And yes, I ran it only in one zone
(Will never do that again, I promise )

I guess I should have not cut anything yet but maybe test more landers, if I was not profitable?

This is the inital campaign in propeller ads:


Bemob:


The duplicated campaign, which was profitable for short time, but then I destroyed it, I guess.

[img]undefined[/img]




I think I should not give up on this campaign yet, because it was profitable at least for one day or so, but I'm not sure what I should do now, starting from this -50% ROI.Should I create one more lander variations, or test 3-5 more landers even, but run with the same targeting as the original campaign?
Maybe I could also add some more creatives to test.

Would you advise me to test more landers or even other offers now, to be at least green before I start cutting placements or optimizing in any way?

And what about SmartCPC on Propeller ads?
I am lazy and so I could avoid cutting placements and optimizing myself and let the machines do it :P

So many questions XD Thanks for helping out!

---

Helloooo, back with another post.
Its been a while, because stupid work always gets in the way
I had some initial success with push campaigns, but they have died out pretty quickly So some help is definitely appreciated!

I ran two campaigns, one carrier only, one wifi only, because I read, that the bids are very different, so I have separated it that way.
I have used 2 spinning wheel landers, because they were the best so far in previous tests.

I wanted to test 3 sweepstake offers in Germany, one Galaxy S10 and two iPhones; I created the paths like this:


The first two days I had really good results with the CARRIER campaign in my opinion:


The WIFI campaign did not nearly as good, but I want to show the stats as well:



From here on I will only talk about the CARRIER campaign.

On the third day I did some optimizations. I deactivated 2 zones:


Deactivated 2 creatives:



Deactivated 1 of the paths in Bemob:


I then also increased the bid from 0.06 to 0.08, because Propeller Ads only spend like 5$ the whole day.
But from there on everything went to shit and became unprofitable.


My question here is of course, if I made any major mistakes and how could I crash this campaign so hard?
Also strange is, that I had that exact lander with only one offer (the galaxy S10) running before and it didnt do as well, as with two offers.


The next day I created a campaign with the best spinning wheel lander and two offers (carrier only this time), one being an Amazon voucher and one being a supermarket voucher (Penny, Rewe, Aldi).
After two days these were the stats:


The supermarket voucher did way better here and these were the creatives:



Its been two weeks since I have stopped this campaign and haven't had much time to set up a lander with only the voucher to test it. (Should I even do that?)
Should I now test running the voucher or has there been too much time in between the test and today?

Thanks a lot for your help

---

Wow - sorry I thought I had replied to this thread - must have opened it in a browser tab and didn't get to it - thanks for updating the thread again so I'd see it!

OK I see the confusion regarding the 2 conversions.

In the beginning, when you don't yet have a proven lander, you need to split-test landers to find the best one of the batch.

In order to split-test landers, you'll need to collect a certain number of conversions - and you'll need to use an offer to get these conversions.

So which offer do you use? You have a catch22 situation here where you don't have a proven lander to test offers, or a proven offer to test landers.

The methodology I suggest in the 40-day tutorial is this:

-Throw a bunch of landers and some offers into a campaign.

-Wait and see which offer makes 2 conversions first.

-Use that offer to split-test landers to find a winner, pausing all other offers.

-Use that lander to test offers to find a winner (can include all the offers that were paused before - because the offer split-testing had NOT reach statistical significance remember?)


Some things that need to be pointed out here:

-The offer that reaches 2 conversions FIRST, does NOT mean it's necessary the winner of the bunch of offers being tested (again, statistical significance needs to be reached first).

-The "winning" lander of the batch only means it's the best of the batch. It does NOT mean it's necessarily a very good lander in general. This is why you need to make sure to include at least all the popular landing page styles, landers that have received a lot of traffic (according to spy tools), AND that all the landers are functioning as intended and are optimized for speed. When you do all that, THEN we can ASSUME that the winning lander is a decent lander. With that assumption, you can then use this winning lander to test offers.


As for your example of 7 conversions across 2 landers with -80% ROI - you'll need to look at the ROI of the INDIVIDUAL offer+lander combo.

So how high does the ROI of an offer+lander combo need to be, for it to be worth optimizing?

For newbies, a VERY rough "rule of thumb" would be to optimize a campaign when the ROI is -50% or better, better any significant optimization. You can use that as a crutch in the beginning. But there ARE exceptions.

Example: If you're running in a geo on a traffic source that has a shit ton of traffic from a shit ton of placements, and/or your offer accepts traffic from many traffic segments (e.g. all carriers + wifi, ios+android, all browsers, desktop+mobile...the more the better), then your starting ROI can be more negative than -50% ROI and you'd STILL be able to cut all the way to green.

Another example: The ROI seems really negative at first, but when you dig into your stats you find that more than half the traffic is from a big placement that has spent over half your budget without giving you any conversions. You may be able to cut it to drastically increase your ROI.

In the end, it all comes down to 1)how much room for optimization there is, and 2)whether it would be enough to optimize the campaign from the current ROI to your target ROI.


And what gives you room for optimization? Main factors are traffic volume and [B]how many traffic segments the offer lets you target.

Hope this helps! This post is getting long - I'll start another post to reply to your other posts above.



Amy

EDIT: Just remembered you switched to running push - my reply in this post was in reference to pop. Similar reasoning should still apply though.

---

You can start optimizing whenever you feel is suitable, whether the budget has run out or not. Just remember a couple things:

-When doing a split-test (offers or landers), don't introduce new candidates before the split-test is done. (e.g. When split-testing landers, don't add new landers to the batch before a winner is found. You can start a new round of split-testing with the winner AND new landers later.)

-If you start cutting placements etc. before you have a promising lander+offer+ad combo, keep in mind that you may want to unpause them for a retest once you DO find a good lander+offer+ad combo, because a placement is only as good as the lander+offer+ad you're running. (But since you were running in one zone, this doesn't apply. BTW how did you decide on that particular zone?)


See this is why you need to put so much work into landers BEFORE the testing. What I said in the last post about including landers of every popular style, that have received lots of traffic, that are optimized for speed and work correctly - the AIM is for you to be confident that the split-test winner is a decent lander.

So that when you don't get good ROI, you can confidently (or at least semi-confidently) say: The lander isn't the problem. It's probably either the offer or the ads, or both.

(Of course the other factor is traffic quality - which is why you should always do testing on a source you know has high quality traffic. Not bidding too low would be another smart thing to do.)


It's weird, but it's common to get better-quality traffic at the start of a campaign - like for the first hours/day, and then the ROI would drop. At -50% ROI it may still be worth it to optimize the campaign (again, everything depends on whether there' enough room for optimization to get from -50% to your target ROI).

What you can do to optimize the campaign:

-If you haven't tested all the popular lander styles that have "received most traffic", or half-assed the speed optimization etc. - by all means test more landers.

-If you're confident in your lander, test more creatives.

-If you're getting good traffic volume, you could even cut placements before testing more landers/offers/creatives. Just keep in mind that you may want to unpause those placements to give them another chance later, when you DO have a good lander+offer+ad combo (as mentioned above).

-If you feel you've tested enough creatives, and you don't see enough room for optimization to get to the target ROI, test more offers.


Questions to clarify my confusion:

-The 2 spinning wheel landers - one was for iphones and the other for galaxy 10, correct?

-I see that each path has 1 lander going to 2 offers - shouldn't you have the iphone lander going to the 2 iphone offers, and the galaxy lander going to the 1 galaxy 10 offer?

Just wanted to make sure you have the right landers going to the right offers!


Why?


It also looked like 2 of the offers for the remaining lander were almost ready to be cut as well. (It's one percentage away from 90% statistical significant - depends on how accurately you wish to cut.)


The change MAY be due to the bid change, but may also have other causes - volatility in performance is the name of this game, which makes it hard to strategize sometimes.

Also: It would help to look at placement/zone stats. Maybe by bidding higher you triggered traffic from some shitty placements.


Only actual testing will tell. I think it would be worth a test - too early to tell but your best creative has made 2 conversions at $3.68 CPA and your payout is 2.25 euros so ROI isn't bad. Of course there's too little stats still to draw any definite conclusions.

I will ask the legendary @twinaxe for input as well. He's the push expert around here.




Amy

---

Hey there,

I am a bit confused with your setup

You say it´s 1 Galaxy S10 and two iPhone offers, why is there 1 offer in 2 paths?

Then this



These are stats for the carrier campaign sorted by landing page.

I wondered a bit because it seemed that you run all 3 offers on both landers although it´s for different prizes so I checked the landing page to see how it looks.

First thing, don´t use a .de domain for your landing pages.

Get a whatever domain where you can get whois guard.

Then the landing page itself.

On the js alert you say the user can win a Samsung Galaxy S10, on the prizewheel image you also show a Samsung Galaxy.

But when the user "wins" then it´s an iPhone that rotates in.

Then on the pricepage you have two offers to choose from.



What I see first is that the buttons are not shown properly.

Next thing is, don´t make two offers available to choose from.

It works better when you make only offer available and and block the other because it´s not available anymore.

It´s also the Samsung Galaxy image from the spinwheel but a different iPhone image than before.



But why did you blocked the zones?

By what criteria did you decide?



Same here, why did you block that path?

First thing, when you increase the bid from 0.06 to 0.08 you pay right away 33% more for the already existing traffic.
Additionally you also receive more volume and also traffic from more zones so it´s normal that the ROI the goes down.

But that´s how it works.

You have to check the stats more in detail, don´t just go by how much the campaign spent total.

In big or high volume geos you will receive traffic from many different zones so that you can spend quite some money and still have only few clicks per zone.

When the campaign has potential we just have to eat some losses first and will make it back many times when the campaign is more optimized.

How much did you spend altogether for the carrier campaign?

You have 3 offers with 2 landers and payout is about $2.40 or so.

Germany is also a big geo so when you only spent like $30 or so then you don´t have enough stats for any big decisions.

You also received few conversions.

Do you use the same spinning wheel lander as before?

What prizes do you show on the spinningwheel and rotating in?

I just ask because it´s again two different prizes on one lander.

There´s only one way to know -> Testing

Or you could ask your AM if these offers are still converting

But few things, you really need to wait more before you make changes in your campaigns.

What you did looks pretty random, when you don´t have a clear structure how you run your campains or when to make decisions it will be very hard to see patterns in the campaigns.

You also need to spend some more when you want to run campaigns in geos like DE.

When you want pay less you need to in cheaper geos.

And it´s better to run in cheaper geos and letting the tests runtill they finish instead of running in expensive geos and stopping zones/offers/landers/campaigns before they reach meaningful stats.

That way you basically just throw money on random campaigns and in the end this will cost more than running less campigns but therefore properly.

You know the saying:

"Wer am falschen Ende spart zahlt doppelt"

---

Thank you so much for that information!!
Pure Gold! I guess I need to go back to the drawingboard again

You asked, how I decided on a specific zone to deactivate. I found that it costed a lot, but didn’t bring in conversions.
Maybe I was too fast here....

---

Thank you for the very detailed answer, seems like you’ve put a lot of time in it.
I really appreciate this!!!


You say it´s 1 Galaxy S10 and two iPhone offers, why is there 1 offer in 2 paths?
I wanted to test 3 offers and thought this was the best way to do it. Yeah I know, not very beautiful.


First thing, don´t use a .de domain for your landing pages.
But why? Not even for campaigns in Germany?


It works better when you make only offer available and block the other because it´s not available anymore.
Will do! Thanks


But why did you blocked the zones?
By what criteria did you decide?
I found that it costs more than the other zones and has no conversions....
Too early I guess.


Do you use the same spinning wheel lander as before?
What prizes do you show on the spinningwheel and rotating in?
Yes, I found that this Lander works best on all of my campaigns. Still I should test more landers.
Not sure what picture to put in the spinning wheel, if I want to advertise more than one offer, you told me to use only one offer, so that’s out of the way.


On the other hand, how do I bulk test offer, without putting more than one offer on a Lander?
Do I need to create a Lander for each and every offer or is there a shortcut of some way?


>> I have learned that I need to wait waaaaay more, before I decide on anything. I think that’s the biggest learning here. <<


I’m still a little unsure how to test a lot of offers and landers to finally find a profitable campaign.
If I need to adjust each and every Lander for every one offer, than it’ll take a lot of time (But maybe it is just the way it is ;P ) or is there an easier way to do that?


Im also asking myself, if I just run all offers for one geo for example in one direct linking campaign and if one of those converts at least to a certain degree, I could run another campaign WITH landers and that offer. This would basically add another step to testing, but I that could save me lots of time.
Costs a little more money though.


Do you think that’s an idea to consider?

---

You have no imprint nor any DSGVO stuff.

There are lawyers in Germany who do nothing else than searching the internet for any stuff that´s worth to sue people for.

Yes, too early.

Yes, create a lander for each prize but you don´t need to create the landers from scratch each time.

Search and replace the name of the prize, switch 1-2 images, that´s all.

Do it few times and the task takes only few minutes.

No.

Of course you can do it that way if you want to but I don´t recommend it.

If you want to save time I would rather prepare landers for all prizes in the geo I want to run, then add all landers and prizes to the same tracker campaign in separate paths and splittest all in one campaign with more general creatives like "Herzlichen Glückwunsch. Klicke hier um zu gewinnen!"

---

In addition to what @twinaxe said:

You need a cutting criterion. Usually the suggestion would be to cut when a zone reaches 2x payout in loss.

(So if your payout is $1, then cut the zone when it's in loss by $2. )

But if you're getting a lot of traffic from a lot of zones and losing money fast, and wish to cut faster, you can cut at 1x payout, or in some cases even 0.5x payout. However, in those cases, you'd risk cutting out potentially profitable zones. What you can do is cut aggressively at first to curb the losses, then once the campaign is profitable or at least not losing money very quickly, you can unpause some of the zones to test them some more.

That's it in a nutshell. I'm writing a new lesson in the 40-day tutorial right now on how to cut zones - will drop a link in this thread when I'm done.


However you do it, make sure the prize on the lander you're using, matches the prize in the offer you're promoting.

Of course you can promote more than one offer in a campaign - in fact you're encouraged to when possible.

There's nothing wrong whatsoever with having 2 paths in the tracker! Again, you just need to make sure that each path is promoting a different prize. For example the default path can consist of iphone 11 landers and iphone 11 offers, and the second path can be iphone 12 landers and iphone 12 offers.



Unfortunately, most of the typical sweeps landers are designed to promote a single prize only.

There are landers that ask the user to select from 2 or 3 different prizes for example. So you can link them to 2 or 3 different offers promoting different prizes. You can find these landers on Adplexity. However, they usually don't convert as well as landers that promote one offer, so you should at LEAST test other landers styles as well.


Yes and no. Not waiting enough can be a problem, but waiting too long can be a problem as well.

It's difficult to know when the optimal time is. We have some rules of thumb that can help (e.g. cutting zones that are 2x payout in loss). If in doubt, read the 40-day tutorial again (focusing on the lessons on testing and optimization). It will get easier as you run more and more campaigns - you'll develop a SENSE on when to cut what, when to abandon a campaign, when to scale etc.


What @twinaxe said.

Also: You could hire a programmer so you wouldn't have to do this yourself.

It would be tricky (although possible) to make landers that can promote more than one prize. However, you can find/create landers that can shown multiple languages.

So for example you can produce 5 landers for iphone 11, that contain all the popular languages. Then you can pass a parameter in the lander url to specify the language you want the lander to show, e.g. https://landerdomain.com/LP1/index.html?language=en (for english) or https://landerdomain.com/LP1/index.html?language=es (for spanish). Then you can use the same set of landers for all iphone 11 offers for all geos that speak those languages.
@twinaxe is working on a tutorial on multi-language landers so stay tuned!


I had the same thoughts when I was a newbie - just direct-link to all offers for an initial test. Then only pick the promising ones to prepare landers for.

This method CAN work for certain offers - ones that have easy conversion flows that don't need preselling. Unfortunately, most sweeps offers don't belong in this category.

For sweeps offers though, landers are recommended. Without presell, conversion rates can be dreadful, and you may waste a lot of money on testing as a result.

You can do a comparison yourself:

1)Create a campaign to run, say, 5 iphone12 offers direct-linked.

2)Create a campaign to run the same offers using landers.

And see in each case, how much money you need to spend to get your first few conversions. Then decide from there.


Alternatively: If you don't like to mess with landing pages, you can look for offers that can convert when direct-linked. Offers with very simple conversion flows, or offers that already have pre-landers built into them (pre-landers are just what aff networks call landers), make good candidates for direct-linking.

Hope that helps!



Amy

---

Home > > Newbie Follow-Alongs