Heard of A/B testing? If you haven’t, you ‘re missing out!

What is A/B Testing?

A/B testing is a great way to get the most out of the information architecture of your website to ensure it is delivering to its potential.

Firstly you need to set goals for your website to measure its performance. Engagement metrics (time on site, pages per visit, bounce rate etc) are good, but more concrete stats even better. If you’ve got an online store or ecommerce platform, you should already be measuring sales performance against the traffic referral channels coming into your website. Alternatively, if your website is designed to generate leads (through online contact forms, phone calls, emails, etc) then you should track these too.

A/B testing involves reviewing the structure of your website and coming up with ideas for making it better  – increasing its performance in the key metrics we outlined above – and testing it. The reason we test is simple – we might be wrong!

So install the code for your favourite A/B testing software on your website, prepare your “Treatment” (to duke it out with the old page, the “Control”), and get testing!


Let Go of What You Want – It Doesn’t Matter

If everything you wanted and planned for your website design was the best way to connect with your customers and maximise sales, then you’d create the perfect website right out of the gate. This is pretty rare, and I’ve never seen it happen. This is because even if you undertake a thorough information architecture design phase, most client budget’s require certain assumptions and best practice layouts to be used in the process. In reality, there is no such thing as “best practice”, as every product, target market and website is different. The only way to really know how to get the most out of your design is to test.

Start by analysing your website – how to users get there? How do they progress through the website to the ultimate conversion goals? What pages, messages, images and design features do they navigate through on the way?

Once you’ve broken down how your website is actually working, look to spot key areas that could be improved. Button could be more obvious? Try a different placement, colour or shape. Got a long form? Try a short one. Missing opportunities to sell to your potential customers in key areas? Test some new copy ideas.

The key takeaway for this is to create a hypothesis to guide your tests – testing random variations of random elements might net you results, but is more likely to be a waste of time. This hypothesis should be guided by the analytical data you have about your site already, as well as your own intuition. At Uprise we have a Funnel Analysis and Testing Plan process to identify these key areas and really think about how we believe they can be improved before we do any testing at all.

Conversions and Engagement – The (potential) Customer is Always Right

Dilbert, March 8 2014

One obstacle people often get hung up on is their own notions of what works, what looks good, what is “on brand” etc for their website. I can certainly understand this – your website is your online presence, and you want it to reflect your business and its core values. From a marketing and return on investment perspective, however, there is only one thing that really matters – website performance.

From this perspective any change that results in an increase in sales is a positive one. If it doesn’t fit your brand or your ideals, serious thought should be given to adjusting your expectations as the testing strategy shows they don’t resonate as well with your customers as others, and you’re missing out on sales.


So what are you waiting for?

Do you want to know more about how customers use your website? Are you looking to increase the performance of your digital presence? Do you want to make more money online? Of course you do.

So why aren’t you A/B testing?