The Art of A/B testing (on a landing page)

It clearly wasn’t working.

7 days and only 4 conversions.

Something had to change – so what did I do?

We’ve all been told that A/B testing is the clear path to success, riches and worldly power – but:

-Have you really tried it out?

-Do you know the rules?

-What should you test?

A current client of mine runs a gym where he helps people whose health has deteriorated.

We’ve been running FB Ads to drive leads to get new monthly clients.

This specific campaign was quite broad, we were looking for local people, and using a 30-day challenge as the enticement.

We had different ads set up, and the landing page was ready for them, with a simple form for them to fill out in order to get a consultation call and sign up.

However, after the first week, we weren’t getting the conversions we needed.

Not a great start.

But rather than get upset, or annoyed, or feeling defeated – I saw it for what it was, the  time to test.

The landing page was set up with the information that my client had given me.

Features, benefits, descriptions…

It was solid, but I thought it might be too long, So I tested a shorter, more to-the-point version.

I ditched most of the text, cut down the images from 8 to 4, and had to make the tone more relaxed.

The result was the difference between profitable and not:

Screen-Shot-instapage

Now bear in mind – 6.5% is NOT a wonderful conversion rate.

This post isn’t to show off, it’s to show how simple changes can make big differences.

But then you keep going, another 40% increase here, a 25% increase there, and suddenly you’ve created a machine that spits out £10 for every £1 you put in.

That’s a machine everyone wants to own.

And that’s what I’m going to be building with this campaign.


The rules of split testing:

  1. Have a definite purpose:As with all your marketing, mindlessness is a surefire way to lose money.
    So have a clear picture of what you’re trying to test.
  2. Test one thing at a time:This should be obvious, if you change 3 things on a page, you won’t know which one makes the difference.
  3. Test randomly, and at the same time:Again, its about having data that is actually valuable to you, so use a tool that tests against your control at the same time, (i.e not one for 1 week, and 1 for another)
  4. Wait until you’ve got enough data to draw conclusions.Its no good making a change, leaving it for a couple of days, then freaking out when your results are wildly different, if you get 200 visitors a week.
    – Firstly, you need to leave it for a week regardless of your traffic, otherwise you might miss natural fluctuations.
    – Secondly, you need to avoid (or at least be aware of the fluctuations caused by) national holiday etc

So down to specifics, What can you test?

So now we’ve got the point, and the method of testing down, what do we actually test?

  1. Headline
  2. Page length
  3. Main copy
  4. Background colour/ image
  5. Other images
  6. Call-to-Action
  7. Text colour
  8. Page layout
  9. CTA button colour

Tools to use:

experimentcalculator.com – How long will your experiment take?
commercesciences.com/tools/ab-test-significance-calculator – A/B test significance calculator.
instapage – The landing page builder I like to use.

So that’s it – get testing, and tell me about your tests/wins/losses below.

Jamie


Are Facebook Ads right for your business?

If you aren’t sure, lets schedule an informal 20-minute chat; and I will give you honest advice – no sales pitch, no bullshit, no pressure.

Enter your details here and I will be in touch to arrange a time to talk.

[contact-form-7 404 "Not Found"]

Leave a Comment

Your email address will not be published. Required fields are marked *