Why ‘A/B Testing’ is a Complete Waste of Time For Most Affiliates

When is A/B testing NOT A/B testing?

Below is a scenario that feels like optimisation. Unfortunately, it is the same as babbling perfect common sense with your head up your arse.

Let’s say you launch a campaign and track these variables:

  • Banner Ad 1 vs. Banner Ad 2
  • Landing Page A vs. Landing Page B
  • Offer X vs. Offer Y

After three days of testing, you establish that Banner Ad 1, Landing Page B and Offer Y are the best performers.

How can you be sure?

You can only optimise one variable at a time.

Testing multiples is not optimisation. It is guesswork. Trade-offs.

To establish a best performing variable, test it in isolation.

For example:

  • Banner Ad 1
  • Landing Page A
  • Offer X vs. Offer Y

This process is time consuming, yes, but infinitely more reliable with the small datasets that affiliates swear by.

In what order should you test variables?

  1. The offer
  2. Your angle
  3. Landing page theme
  4. Landing page content
  5. Banner/text ads

I’ve dissected this process at length in Premium Posts Volume X.

If your testing process currently resembles a madman throwing turd at a hurricane, I suggest you start there.

Recommended This Week

  • Volume X is sponsored by Adsimilis, one of the top networks for CPA affiliates. Adsimilis has hundreds of top offers (specialised in dating), with industry-leading payouts, and international coverage. Get onboard and start making some money! Email me if you need a referral.

About the author

Finch
Finch

A 29 year old high school dropout (slash academic failure) who sold his soul to make money from the Internet. This blog follows the successes, fuck-ups and ball gags of my career in affiliate marketing.

6 Comments

Leave a comment
  • Testing in isolation and then eliminating the non-performers one by one is just as bad statistically as testing multiple offers at the same time. If you want to do it right, you have to test all 2^n combinations (or however many variations you have). Problem is, it’s usually pretty expensive to do that.

  • Good point. I agree, you’d want to test every combination in a unique campaign if you had the time/money. But in practical terms, very few affiliates actually wait for statistical significance before dropping an ad/LP/offer. We have to make rough snap judgments. Given the limited data, I’ve always had more success isolating the most important variables (offer and appeal) and testing these before moving on to the rest. I don’t think it makes sense to split your budget every which way before you’ve nailed the variables that will define whether you’re in the red or black. When I see affiliates testing 5, 6, 7 different variables on $200 ad spend, I think that leads to much more guesswork than a systematic approach that focuses on finding 1) the best offer and 2) the best appeal. I wouldn’t want an awful banner to mess with my chances of isolating those factors, which it’s much more likely to do on lower spend from my experience.

  • I prefer testing the angle and ad text first. The reason for this is because the angle is what determines if profit is even possible. After that, I then test the ads to see what gets the highest CTR so that from there on out, I am getting much more traffic, yielding more data at a lower price. If I start out getting the most clicks possible, I can then test the other variables more quickly. Waiting until the end to test the ads means you’re going to spend twice as much testing every variable that’s in line before it.

  • Yeah I should point out that while I test the ads last, that’s in reference to things like colours/wording/imagery. The ad text goes towards the angle, which is obviously very important. Your familiarity with the market is a factor too. If I’m moving in to a volatile foreign market, I want to know that there’s an offer I can make money with first, which is why I test it before working my way through a bunch of angles.

    Generally speaking, if you can find the hottest offer + the hottest angle, you’re going to make money.

  • Time and Money consuming. How would you expect the average n00b aff, expecting to put $100 and turn them into $100,000 to do that. Profit is grown not magically produced out of thin air. The techniques are all solid Finchy, but they take second seat until the mindset of this revolting, children owned industry changes. “Look at me mah – I made 1m and spent 950k on toys. Oops… did I just leech all the money from my business and hindered growth? Maybe can’t even pay bills next year, cuz FB cloaking died?”…. ooops. Until the #RyanEagle style of running/doing business vanishes from this industry your posts will probably fall on deaf ears :(. To all the true peeps out there looking to create a real long term, high revenue (lower ROI) business – kudos. Keep it up. Stay away from those who sell lifestyle as a credential for their success. Hmmm, time for another guest post Finchy ? 🙂

Leave a Reply

Your email address will not be published.

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <s> <strike> <strong>

Copyright © 2009-.