5 Dos and Don’ts of Effective A/B Testing

Ready to perform some A/B testing on your website?

The concept may be simple, but A/B testing’s role in inbound marketing is so significant that you should always know what you’re doing. Here’s a cheat sheet of “dos and don’ts” pointers:

Image Courtesy: Small Business Database Marketing

Do: Start with the Goals, then Test

This might seem like trivial common sense, but it’s in fact a frequent mistake that derails the entire effort. Without direction, your A/B testing is meaningless; and your goals dictate your direction.

Don’t: Set Goals Not Aligned to Current Needs

Always align testing goals with inbound marketing improvement. Even if you have set goals, but they’re not aligned to desired improvements in your inbound marketing, then the results of your A/B testing efforts won’t be useful until such time as it can be employed to improve one facet of your marketing or design.

Do: Setup as Many Letters as Needed

It isn’t called A/B testing because you can only test version A and B — you can in fact test as many versions as you need to gather the sort of data that yields useful insight. Remember: inbound marketing is all about collecting, cross-referencing, and analyzing relevant data, and if you need three, four,or five versions to test to get the required data, then go for it.

Don’t: Setup Too Many

Of course, that doesn’t mean you can go berserk and set up an alphabet’s worth of versions for testing. Set up as many as needed, not as many as is called for by whimsy.

Do: Measure the Right Metrics on the Proper Channels

Don’t be shy and explore the more advanced metrics too. It’s one thing to stick to relevant metrics and another to deliver into the added insight that advanced measurements can provide. If you are testing mobile for example, you may want to measure if users are coming only from large cities such as Toronto.

Don’t: Let Scope Creep Sidetrack Your Testing

Stick to your A/B testing process and always focus on what you’re doing. When you’re looking at the data you’ve collected you can get sidetracked by other areas of your Google Analytics that might be affected by the changes applied to your different versions. Take note of it and save it for another A/B test, otherwise you’ll see no end to your current one.

Do: Zero In on Conversion

It’s always about increasing the bottom-line, so if you’re testing some on-page SEO tweaks and you notice one version is getting more traffic than another, you should also always check if that version is likewise getting more conversions. Don’t be easily misled by positive changes in metrics as these may not directly translate to equally positive changes in conversion.

Don’t: Forget Your Funnel

Don’t be too focused on conversion, however, i.e. you should not sacrifice the integrity of your entire inbound marketing strategy to increase conversion. For instance, conversion rates can spike for a certain page, but if at the same time exit rates go up as well, then you know you’re creating an exit from your sales funnel by focusing too much on making that one page convert better.

Do: Test as Many Facets as You Can

Inbound marketing is a dynamic effort — it’s not a set-and-forget system. Always test the different aspects of your marketing and design that you think can do better; even the ones that you don’t.

Don’t: Test Them All at the Same Time

Let’s get one thing clear though: never test multiple aspects at the same time. You’ll get conflicting data. You won’t be sure which change affected the outcome, and that will result in a poorly executed A/B testing round.

A/B testing is extremely useful for inbound marketing if you know your dos and don’ts.

Do you wonder what constitutes a great marketing campaign?
Unlock the secrets of smart marketing and download our latest eBook
The Smart Marketer’s Guide to Creating
Effective Inbound Marketing Campaigns

  • Storewars News

    read! Very informative. Did you know that Diageo stocks $2.3 bn ahead of open
    offer for United Spirits. Full story here: http://bit.ly/1r7Ly96.