Many non-marketers would likely wrinkle their brow in bemusement when asked if they knew what A/B testing was. Firstly, we can assure you that it does not involve any kind of medical procedure. Secondly, we can assure you that A/B testing can in fact be one of the most easily executed ways to help you improve your marketing engagement – and more importantly, your results – with a surprisingly small amount of effort.
A/B testing can also be referred to as split testing and usually refers to an instance where you test two (or sometimes more) versions of an activity – typically a web page or email – to determine which is more effective. This exercise can actually be quite easy to run for your organisation depending on how complex you want to and what the asset is you want to test it for. By running an A/B test though you can sample a small portion of your audience or test different ideas and then analyse the stats side by side to help you optimise your marketing execution.
Here’s a couple of examples of how you can run an A/B marketing initiative in your organisation:
Example 1 – Email Marketing
- A/B test different subject lines. Pick two different subject lines that you feel could be relevant to your email content and send a version of each to a small portion (perhaps 1%) of your database. Review which gets you better engagement rates such as click throughs and open rates and then take that best performing line and use for the other 98% of your database. It can really have great results!
- A/B test different content layouts. Perhaps you could choose one email option with a more visual or low text design and a second with a more text heavy format. Take the same approach as above and trial different options.
Example 2 – Web Pages
- Whilst it can take more time, to help you optimise your web traffic you could create 2 different designs or content options for a specific page or section of your site. Working on the same date and time ranges, trial each option for a specific period of time and then assess the audience engagement stats you get for each. Depending on the results, opt for the more impactful page and implement – you can run the same exercise for as many or as few pages as you want, but for more scientific results it’s best to try this with one or two pages only at a time, perhaps a homepage for example.
These examples are relatively simple and straight forward to run as part of your own strategies. Next time you’re planning an email campaign, write A/B testing as part of your checklist and just do is a matter of routine. You’d be surprised the impact a simple thing like subject line amendment can have – I myself remember trying out a “flashy” and attention grabbing subject line as part of an email campaign and feeling confident of its results. After running an A/B test on my creative subject line idea versus another more clear and yet simple option, the results were nearly 4 times better in favour of the more straight forward option.
A/B testing can actually be executed in a range of different activities, but conducting it on your marketing campaigns it can help take away some of the guessing work and reduce the need for assumptions. By getting a more analytical understanding of your audience and customers through A/B marketing you’ll be surprised at the extra returns this small change can make.
So you’ve understood the principles behind A/B testing, so perhaps it’s time to look into the reality of how you can adopt this approach long term. There are numerous tools and software programs available to you organisations to help you to automate your A/B testing. When it comes to email testing, typically your email engine will have the tools built into it that allows you to do this. When it comes to website A/B testing it’s never quite so simple – trying a manual approach as described above is hugely time consuming and also hard to accurately track. Here’s a selection of 3 A/B test tools available to help test websites and web pages:
- Google Analytics Web Optimiser A completely free tool from the biggest search engine around. Can deliver “experiments” to help you optimise your web design and content options relatively easily
- Optimizely.A powerful and detailed tool that enables you to deliver widely customised setups and test numerous elements and options. It’s used my many of the largest companies in the world and is certainly not the cheapest but is a highly powerful tool.
- Unbounce. A site designed for those mainly not using custom packages such as WordPress or Wix and allows you to make custom pages with the ability to personalise and tweak designs effectively.