Perform Split Testing with phpA/B
Split testing, also called A/B testing, is a testing method where you compare a base sample with another test sample, changing one specific variable inside it.
For example, you could have a base landing page, and then you would compare it against a new landing page where you changed the main headline.
As a blogger, there are many things you could improve with split testing. Examples include:
- The number of subscribers you get by testing the headline and position of the subscription form.
- The AdSense earnings by testing different AdSense units and placements.
- The bounce rate of the blog, by testing different placements of your content above the fold.
- The number of retweets you get per post, based on the placement of the retweet button.
But how do you perform a split test? You have two options. You can either do it manually, by changing the variable one week at a time and then tracking the results, or by using a piece of software and letting it rotate the variables for you.
The second option, which involves automated software testing tools, is usually the best one, and a couple of days I discovered a software you can use to do that. It is called phpA/B, and it is free as long as you use it for personal purposes.
I haven’t test it yet, but it looks promising. The main idea is to let you test anything within your site, with a simple setup interface. You can also integrate it with Google Analytics if you want.
Definitely worth a look. Once I test it here I’ll write a more complete review, so stay tuned.
Recommended Articles for You
18 Responses to “Perform Split Testing with phpA/B”
And why not use Google Website Optimizer for this? Its free.
As is phpA/B, it doesn’t matter so much what tool you use really, all A/B testing tools provide one big common element in their toolkit, they provide the ability to test between version A and B of a certain aspect in your site. Google website optimizer does this just as well as phpA/B does.
One might argue phpA/B provides a little more freedom as it actually provides you with tools that you can use to implement the A/B testing, where-as Google website optimizer only provides result verification.
- Rian Orie
I recently came across your blog and have been reading along
Superb concept. Will try it soon enough.
Will be waiting for your review though 😛
I’ve never make split testing. Now split testing in my blog may be stupid, because now I am doing a lot of other works on my blog, so results may be falsified.
I think about write a split tester in Ajax.
Keep an eye out for potential lopsided testing segments.
i have never done split testing, but it definitely sounds like something that I will be performing in the future, as I delve deeper into the depths of blogging.
If you do a review, could you compare it to google webmaster tools?
Tom | Oral Answers
Do you know if there is an easy way to implement this to alternate themes on a WordPress-powered site?
I noticed that you can add the phpAB code into a page, but I think it would take a LOT of those to change all of the CSS, images, divs, and everything else.
Thanks for letting us know about this.
Tom, combined with, for example, a theme switcher plugin this should be relatively easy to do.
I wonder though, how effective is A/B testing when you change the whole site.. doesn’t it defeat the purpose of A/B testing? The point of A/B testing is to increase results through fine-tuning small parts of your site.
Tom | Oral Answers
Thanks for the reply, Rian. I’ve been working on reducing my bounce rate and I think the theme plays a huge role in that. I would make the themes pretty similar, but change subtle graphics and colors to see if one is more attractive to visitors and entices them to stick around for a while.
While the theme may play a large role in sites, I think the content is the main attraction… It’s probably not as subtle as many of the other A/B tests that could be performed.
- Tom | Oral Answers
- Rian Orie
Do you know something, it’s funny, but I test sales pages and squeeze pages all the time…
But I never thought to test my blog!
Can’t see the wood for the trees sometimes until someone points out the obvious.
Will get onto it pronto!
Fab post, Sally 🙂
I didn’t do formal split testing but tried out various combinations of AdSense units for one week, then switched a week later and vice versa. Something that automates the split testing is much more beneficial as I had to track all the statistics manually.
Thanks for reminding me of this. I was doing a lot of split testing couple of month back. Then of course, I got comfortable with what I have. I will look at the link. However, tell us how it works with a full review.
Chat with you later…
PHP A/B is a great package, especially if you are hands down with coding and programming. It also offers you flexibility of testing just about anything.
However, if you are looking for easier alternatives (which don’t require programming), have a look at our product: Visual Website Optimizer http://visualwebsiteoptimizer.com It has WYSIWYG editor for creating variations and doesn’t require any coding or programming. It will be great if you could review it on your blog. I am sure your readers will find it interesting.
We also have a WordPress plugin for it, which makes running A/B tests a dead-simple task http://visualwebsiteoptimizer.com/split-testing-blog/how-to-create-an-ab-test-on-wordpress-or-drupal-sites-without-adding-a-single-line-of-code/
Founder, Visual Website Optimizer
I look forward to hear how it works out for you.
Nabeel | Create Your First Website
I agree, split testing can work wonders and not just in blogging, but in any website like a squeeze page, a sales page etc.
phpA/B looks like a novel concept and idea. I thought the only way to split test was the ‘this title for one week and the next title for the second week’ approach.
I wonder how how do you actually implement it on a blog/website?
Looking forward to the detailed review!
It is a php script you need to donwload and install on your server.
But it is easy to do it, and they have a guide too.
- Daniel Scocco
Comments are closed.