Home > General > Affiliate Marketing Forum

---

Propeller Ads CPA Goal 2.0: Conversion Type Splittest (17)

---

This is a pretty short case study but I hope it will be interesting for you anyway.

In THIS THREAD we discussed some time ago a bit about the different conversion types for Propeller Ads CPA Goal 2.0 campaigns.

You can chose there between:

- Install
- Single Opt-In (1 step)
- Double Opt-In (2 steps)
- Complicated Conversion



Now I finally ran a test to compare the different conversion types.

For this test I only used 1 step, 2 steps and complicated conversion.

I created 3 campaigns for each conversion type and targeted Martinique, Mobile, Android, WiFi and 3G and I set $0.20 as CPA Goal.

As an offer I used a click2sms offer with pin submit from Haka.

I didn´t use any pre landers and ran it as a direct link campaign.

Ok, I don´t want to bore you with uninteresting stuff so let´s check the outcome of the test right away:



You can see that all three campaigns received more or less same traffic volume but the „complicated“ conversion type converted much better than the „1 step“ and „2 steps“ is in between.

Based on the statistical significance calculator the sample size is also big enough to make a final decision about what conversion type was really best for this test.



But what do we really learn from the test, how can it help us?

Well, at first we can see that it really makes a difference what conversion type you choose.

The complete flow of the offer I used is like this:

- User visits offer page
- User clicks CTA button
- SMS application on phone opens
- User has to send SMS with pin code
- Conversion happens

It´s obvious that it´s indeed rather a „complicated“ conversion flow and no „1 step“ flow so the conversion type that matched best for the offer also performed best.

This is good to know because it shows that (at least for this test) the Propeller Ads algorithm worked just the way it should work.

But why has the conversion type such an impact on the performance?

I could imagine that Propeller sends higher quality traffic when you chose „complicated flow“ and lower (yet good) quality to „1 step“.

This would be logical because simple 1 step offers usually convert easier than offers with a more complicated flow.

Following this logic we have seen that choosing an easier conversion type than it actually is won´t work out that well.

A good topic for another test could be to check what happens when we use an offer with a real „1 step“ flow and test it with the 3 different conversion types

I hope you enjoyed this little case study and would be great to get your input about it as well.

---

This is fascinating.

Especially since propeller ‘support’ says Pin submit counts as 1-click and to select that. And click2sms isn’t much more complex than that.

This just goes to show that nobody really knows until you test it

Next step would be to split test a CC submit with complicated and 1-click

---

Oh, didn´t know that.

Well, maybe it also depends on what kind of pin submit or whatever

@jaybot Did you (or anyone else) test CC submits with CPA Goal 2.0?

And when you just have to enter your CC number and hit "Buy" wouldn´t it then be a "1 step" conversion type?

So many questions...

My next step would probably be a real 1 step offer like 1 click or email submit SOI and then testing it with the different conversion types.

---

@affiliatecase case study a couple months ago was exactly that with a CCsubmit in PT using complicated conversion. Then erikgeypes chimed in and said he chose 1-click iirc.

I’ve done it in FR and some LATAM geos a few times with complicated conversion. It works. 1-click doesn’t not for me and CCsubs.

---

Oh yes, you are right.

So the thread that was the foundation for my splittest was for CC submits already.

I really should work on my brain

Thanks, good to know.

Would be interesting to see if maybe choosing "complicated" in test stage is always better than choosing "1 step" and when the flow is easy enough and offer converts good then test the other conversion types as well.

Hope you get what I mean

---

From what I’ve been told, from 1-click, 2-click, to complicated simply sets a larger budget for testing.

That’s what the said in the CPAg webinar. I didn’t actually watch the whole thing (too boring) but that part I remember

---

I didn´t attend the webinar but this is pretty interesting.

Of course it would make sense to set a larger budget for a more complicated flow, that´s what we do all the time anyway.

But in my test results all 3 conversion types received more or less the same traffic volume and have big differences in performance, this looks rather like the traffic quality of the different conversion types was different.

Otherwise it should have more or less same performance when it´s only the budget.

I forgot to tell that I set unlimited daily budget and $10k total budget.

---

There is a link to it somewhere on this forum, probably propeller's official thread. You can skip around past the boring parts

One more thing you also didn't mention: how long did this test take?

You don't have to, I'm just curious because I run MQ too and it has a very tiny amount of traffic that getting to 60k visits would take months on most camps

---

I love these kind of posts, the ones that answer a question that I thought of but never got around to figuring out.

It also could be that the AI "waits" longer for a conversion to trickle in, kind of like a conversion window on FB and thus is able to make better decisions.

---

I kept the campaigns running and the campaigns still run on very low volume but about 90% of the traffic and conversions happened in the first month.

That´s great to hear.

Well, in the end there is only one way to find answers to such questions: Testing

Also possible.

But for me the main argument is still that all 3 campaigns/conversion types received almost same volume but the performance is different from "complicated" being best, "2 steps" second best and "1 step" worst.

This looks rather like a different traffic quality for me.

Would it be only a larger budget for testing or to wait longer for a conversion to be posted back the overall quality should still be more or less the same.

---

Looks like you broke it

They have removed the option for 1-click, 2-click, Complex from new CPAg camps.

---

After @twinaxe case study I also tried the complex conversion on propeller's cpaG. Worked much better than SOI.

Probably it worked too well.

Today I tried to use it again but for whatever reason it cannot be chosen no more. Looks like they killed that feature.

---

Wouldn´t be the first time thst something like that happens

Seems there are some changes at the platforn, now there are also 2 CPA Goal options for Push.

---

I... think that only shows up on super-affiliate accounts.

Doesn't show up for us plebs

---

I saw this like about 2, 3 weeks ago but when I wanted to test, it disappeared and never show again

---

The difference between GPA Goal and CPA Goal 2.0 seems to just be bidding CPC vs CPM?
Thanks for sharing this, I'd have assumed that simple conversions would be much easier for their system to optimize than complex ones, so its unusual but nice to see the inverse is true

---

Yes, seems so but I am not 100% sure.

On CPA Goal 2.0 it says

"While you are paying by CPM, the system is finding the best traffic and optimizing your campaign according to the desired CPA and targeting."

On Cpa Goal it says

"Your campaign goes through a paid test period based on your campaign targeting. During the test period your campaign will be CPC-based."

So it could also be that on CPA Goal you only pay CPC during the test period and later pay the CPA you set.

It´s not 100% clear in my opinion.

---

Home > General > Affiliate Marketing Forum