A/B Testing for the Masses…
Now for something a little more unusual…
I got an e-mail yesterday (with article attached) from a friend of mine whose blog I help out with on the introduction of A/B testing in political campaigns – most notably, the last US presidential campaign. I read it with interest, because it means that like everything else that’s been happening for some time (quite some time actually) in online marketing, it’s finally ‘caught-on’ and no doubt a hundred thousand agencies offering it at excessive prices will spring up soon, like with SEO and social media.
I’ve spent a considerable amount of time doing the whole online marketing thing in one form or another – and to be honest, despite all the hype and ‘specialist’ chatter, nothing in online marketing is really anything new – people have been doing the same kind of things in shops, retail operations and companies for years, just in a more traditional sense.
In the last company I worked for, link building was a staple of their daily online marketing activities. It seemed like nothing more than an aggressive throwback to business association networking evenings, with the difference that in this case, you skip nearly all the pleasantries. Likewise, optimising for search engines – if you think about it, yes it takes some work, yes it’s new ground for a lot of people and it’s constantly shifting. But really, it’s a little bit like if you want to rent out a shop unit in a shopping centre – then you obviously want to be where the crowd are, as opposed to in the basement. What’s new about that? The only difference is, if the shopping centre was owned by Google, then you’d come in one day to find your shop was in the basement regardless to make way for a more popular shop.
Anyway, as is the usually problem with my attention span dwindling, I’ll get back to my main point – the article on A/B testing. If you think about it, shops have been at this for years – mixing up the aisles all of a sudden, seemingly overnight, and then checking out what people think. Some shops have been arranged differently, and so on until they find something that works – having the bakery near the front for example so the smell wafts out. And that’s basically what this is – you try something with one group of people and see what they think, and if it turns out they like it better than what you have at the moment, then you change it.
The problem with the internet is it’s made complicated by how far down you want to divide up the groups to test – age group, country they’re coming from, source (i.e. search engine referral, etc). But in one way or another, it’s been on the go for years as well for sites with any considerable traffic to them. It’s also tricky to know when it’s useful for – on this blog where I don’t really sell anything, what would I test for? I could try putting the adsense or the affiliate banners in the middle of the page and see if I could pull in a few more dollars – but with traffic on the site so small, it’d hardly make any difference considering the effort involved.
But on the other hand, I’m running an article directory and I need it filled with content (I don’t actually, there’s a world of spammers out there doing the job for me) – then having a ‘Post an article’ button right in the centre of the page where it can’t be missed might be relevant. So then it might be something to test for. This website from the article professes to make the actual testing – something which I used to spend considerable time doing – much faster and much simpler, and to be honest it looks great having toyed around with it for half an hour. Provided, as I say, it doesn’t create a whole breed of thousands of me-too services overnight (it probably will).
But at the end of the day, for now, I have neither the time nor the money to seriously play around with this on my own sites – although the website mentioned in the article, Optimizely is very cool and something I’d love to consider if the above mentioned issues weren’t as pressing. For now though, has anyone tried this out? Thoughts?