Home > Paid Traffic Sources > Facebook & Instagram

A/B test (6)


07-03-2017 06:07 AM #1 liocamdiosong1986 (Member)
A/B test

One offer is in three networks. Everywhere different landings. What is the best way to proceed?
Choose one network, drive a certain number of banners on the selected landings and then the best banners split on other networks or do I just have to do a split test on all networks for each banner?


07-03-2017 08:36 AM #2 frozengod (Member)

EDIT: I thought you are asking about advertising networks. 100% my mistake
I agree with what Mobidea said below.


07-03-2017 09:21 AM #3 Mobidea (Veteran Member)

Hey!

First - is the payout for the offer exactly the same in all three affiliate networks? And what kind of offer is this?
For pin submit offer you could try to run it direct-linking.
Otherwise - landing page is recommended, and it would be great if you don't take the one that an affiliate network suggests, but create your own, or rip it from Adplexity and modify. Why so? Check Matuloo's answer here: https://stmforum.com/forum/showthrea...863#post317863

If I were you - I'd try to test the offers first (this could help you define one affiliate network to work with in regards to this offer), and then split-test the creatives.

I hope it helps!


07-03-2017 11:47 AM #4 caurmen (Administrator)

Yup, that's almost exactly the same way I'd do it.

Although, to add an extra step: if you're in the situation where there are a lot of similar offers that could work with a small number of landing pages, take the approach Vortex outlined in AWE's pops panel, and try to narrow down the best available lander first. Then you can use that lander to test all of the offers, rather than picking an offer first then trying to narrow down landers.

That's a great approach if there are more potentially-working offers than potentially-working lander models out there.


07-03-2017 03:54 PM #5 liocamdiosong1986 (Member)

Thanks to all who responded.
But I will clarify my question.
There are, for example, 10 landings. How best to test - every creative to run for each landing? Or choose 2-3 landings, then the best creative take and drive away to the remaining 7 landings?


Under the creative I mean here is not the site - but the picture that the person is clicking on


07-04-2017 04:56 PM #6 caurmen (Administrator)

Ah, I see!

If you know this offer works, or you're already in profit and you're optimising, then run all creatives to all landing pages for the offer, and use Bayesian A/B testing to determine the winner. Just be sure to eliminate losers as soon as it's obvious they can't compete with the current winning page.


Home > Paid Traffic Sources > Facebook & Instagram