Home > Questions and Answers > General Questions

What's the right testing strategy? Feedback on my gameplan (6)

04-12-2014 05:45 PM #1 xckt56 (AMC Alumnus)
What's the right testing strategy? Feedback on my gameplan

I'm focused on making one gaming offer work on Facebook. Going to spend this weekend working on the campaign, creating ads, and testing.

I read up a lot on the forums about testing for angles initially, particularly these posts from Mr. Green, Finch, and Ngo:

- http://stmforum.com/forum/showthread...gnoring-angles
- http://stmforum.com/forum/showthread...icient-Testing
- http://www.charlesngo.com/angles/

So with that in mind, here's my gameplan (please give feedback):

1. Test for the best offer landing page variation (or signup page, since mine has multiple). I'm going to use one of the offer's banner ads provided and direct-link it to all the different landing page variations provided by the merchant. This will discover the best offer signup page that I'll use from now on.

2. Test for the best creative angle. I'm going to come up with a few angles and create 5 banner ads for each, and then test for the highest avg CTR of each angle. So if I have 5 angles, I'll have around 25 images/ads to test, like Ngo says in that post:

Launching a campaign with 1 image of each wouldn’t work because there’s not enough statistical significance. Maybe the men next to fancy cars is the best angle, but you chose a horrible image to represent the angle.
3. Create landing pages based on that angle/ad and test for the winning landing page.

The I'll stop direct-linking to the offer and try to find the best landing page with the highest conversion.

4. Keep testing different variables (banners, landing pages, ad targeting/time) based on that angle until I have positive ROI.

Once I have the right angle combo (banner and landing page) figured out, now I can be more granular in my testing (LP headlines, buttons, images, etc). Maybe even test for different targeting, times of the day, etc.

5. Scale that shit. Don't be lazy, and keep testing and optimizing.

Is that about right? Would appreciate any feedback or input...

Now I see why a lot of experienced guys say to focus on one offer, one traffic source... takes a lot of focus/time and testing ($) to discover a winning campaign and make one offer work.

04-12-2014 05:49 PM #2 davidwikes81 (Member)

This sounds good. You trying to simulate a campaign here?

Go and try rolling a campaign first. Fail, fail badly then ask here for suggestions.

04-12-2014 06:00 PM #3 xckt56 (AMC Alumnus)

Quote Originally Posted by davidwikes81 View Post
This sounds good. You trying to simulate a campaign here?

Go and try rolling a campaign first. Fail, fail badly then ask here for suggestions.
If that sounds good, great - all I needed to hear.

Before joining STM, I failed badly with doing like 10 POF ads and not really focusing or having a solid testing gameplan. I hear you though, ultimately going through the motions, making lots of mistakes, is the best way to learn any skill. Just wanted to make sure I'm spending my $ testing the right away this time around.

04-14-2014 05:06 PM #4 caurmen (Administrator)

That does indeed sound like a solid plan - go forth and test it out!

The only modification I'd recommend is to test more than one offer, across more than one network. Once you've found one that works, though, narrow down and crush it!

04-15-2014 01:29 AM #5 xckt56 (AMC Alumnus)

Quote Originally Posted by caurmen View Post
That does indeed sound like a solid plan - go forth and test it out!

The only modification I'd recommend is to test more than one offer, across more than one network. Once you've found one that works, though, narrow down and crush it!
Thanks a lot caurmen, yup got lots of ideas for ads and getting to work now, will keep myself disciplined with my follow along.

Question though... why do you test the same offer across diff networks? apart from payout, what else could really be different that you're testing for?

04-15-2014 02:03 AM #6 zeno (Administrator)

Different networks use different tracking systems, may broker the offer or be direct (different redirect hops involved), use subtly different landing pages, etc.

In an ideal world the aff network would have all offers direct and have similar redirect/tracking performance to others but this is rarely the case. There are also other shadier things that go on like lead shaving by both the aff network and the advertiser, and this can depend on the performance history of the network for the advertiser.

In the end, there are many interlocking parts and it's easier to not think too deeply about it and just collect empirical evidence. Offer A gives $0.37 EPC at network A and $0.21 EPC at network B. The landers are identical and redirect hops similar. Who cares why, follow the profit.

Home > Questions and Answers > General Questions