A/B testing on your landing page

I was just talking about this with a client and thought it might make a good post. If you’re running PPC campaigns then presumably you are also optimising your landing pages. But if you are really serious about optimising your landing pages for low cost, and increasing conversion rates, then you might want to also look at multivariate testing. Although at a lower level, you can do it the same as you do with your PPC ads, running simple split or A/B tests. A/B testing measures the actual behavior of your visitors on a live page, meaning your results are a true refection of what is working better for your site.

To run an A/B test first set your control page and let it run for a while to get some base data on performance. This page should be as close to perfect as you can get it using usual design methods. Then set up your test page, changing out elements such as headlines and copy text, subtle alterations in the layout, using different sized and coloured images etc and compare the results this page gets against your control page. Don’t make too many changes in your test page – in fact it’s better to just make one change at a time, because A/B testing is only really useful for experimenting on one clear goal at a time. When your test page starts to out perform your control page, make the better performing your control page and then experiment on your new test page to see if you can make it even more effective.

A/B testing doesn’t require a huge amount of work, but it is something you do really need to stick with to get decent results from. Really think about the change you are making and why you are making it. Don’t just change the colour of a font just to see what happens, put a bit of thought into what you might think will actually make a difference in getting people to react to the page. All it takes is time, which we’re all short of now days, but if you’re looking to increase those conversions then it is really worthwhile giving it a go.