- Melbourne, Australia
- FAQ
- About AML
- The List Report
- Contact Us
A/B Testing. For a hundred years it has been a tried and true tool utilised by direct marketers to understand the effects different marketing approaches have on results. A/B testing, sometimes called split testing, is just what it’s name implies.
Basically you are splitting your sales prospects into group “A” and group “B” and exposing each group to different marketing communications to see which approach generates better results (note:
for this newsletter we are focusing on marketing communications testing, not target market testing).
A/B testing is the most basic form of marketing testing, first introduced by mass direct mailers to help them improve results. Marketing testing gets a lot more advanced and complicated utilising multivariate testing, but we’ll save that for another newsletter. For now, let’s take a look at the basic principals of A/B testing with the simple example below:
A/B Testing Example - Direct Mail:
You have a list of 1,000 prospects which you split randomly into group “A” of 500 and group “B” of 500. You then send group “A” mailing #1 and group “B” mailing #2. You find that mailing #2 generates twice the response rate. Now you know that this should be your “control” or standard mailing. Pretty simple example, but you get the picture. What you don’t see from such a simple example is that this process has been the key to some of the most successful direct marketing programs of all time.
The other thing to know is that this very principle can easily be applied to your online marketing programs. In fact A/B testing is one of the best ways to maximise the return of your online marketing investments.
A/B Testing is Perfect for Online Marketing:
We are currently building a website that will allow our client to test the effects of random homepage layouts on their e-commerce conversion rates. In addition to homepage testing, you can (and should) test PPC Search Marketing ad copy, website landing pages, email marketing approached, banner ads - all of these can easily be tested and tracked.
Let’s take a quick look at some other A/B testing thoughts:
1. Be Random - You must use the laws of ransom samples to get accurate testing results. If you pick and choose who gets which marketing communications, A/B testing is no longer valid. The key is to measure the effects of the differences in marketing, so your audiences must be randomly chosen.
2. Don’t Stop - Great Marketers always test against the controls. If you find a new marketing approach that doubles response, keep testing to see if you can beat this new “control”.
3. Vary the Variables - The marketing variables are the things that will affect results. You can test offers, colours, messages, styles, anything that you want but remember not to test too many at once or you will require sophisticated multivariate analysis to know what caused the effects. You can improve the results of your marketing with simple A/B testing, but vary it to see which variables make the biggest difference.
If you focus on your two key variables and test these for now, then from there evolve your way through some of the other variables to see what other changes can improve response. Remember Direct Marketing is a marathon, not a sprint!
Posted on 02nd April 2006