I’ve been musing over the way we’ve historically designed websites since completing the ZDNet redesign last year. CNET has a great reputation for site design and many years experience but with some of the tools available now for minimal cost, there is so much more data available to make informed decisions. When I was in New York recently, I saw a presentation by clickdensity who run a great service that displays user activity as a heatmap (shown above) on the page.
Unlike the reporting behemoths like Omniture, these guys are focusing purely on where the user clicks, not the underlying mass of reporting data and costs as little as £2.50 per month depending on your traffic levels. There are other products out there that do similar things and it got me thinking about a more structured methodology for web design. Here’s my 10 step plan to a more analytical approach to web design: -
- First, look at the most popular pages across the site by page type. For example, news sites would probably find that story pages are the highest ranking, followed by home, search and then listing pages.
- Starting with the most popular page, use something like Clickdensity to analyse not just where the users are clicking but also where they are not.
- Make sure you look at whether the user clicked near to a link which gives you an indication of usability from an image/text link size perspective.
- Analyse the results - Is the user doing what you want them to at this point?
- If they are, great, if they are not then you need to look at changing the design to make it easier for the user to do what you want them to.
- Take away stuff that the user isn’t clicking on. Less is more, infinite choice=overwhelming confusion.
- Go back to step 2 and repeat until you get to the 10th most popular page type.
- In all likelihood, by now you will have redesigned the parts of the site that are responsible for over 75% of your site traffic.
- Measure, measure, measure! No two web audiences will be the same. Your site is unique, your users are unique, there is only so much you can learn from the “experts”. Make sure that you learn from every change, good or bad.
- Go back to step 1 and start the whole process again, web design is an iterative process and you’ll never finish it.
Good web designers will always be worth their weight in gold but however much you know about design or think you know about user behaviour, the data never lies. One of the challenges for all publishers is that the cost of redesigning an entire site is becoming prohibitive, focusing on the areas where you can make the biggest difference is therefore common sense. The other thing about this is that anyone can afford to be scientific now. You don’t need to have expensive analytical tools and eye-tracking focus groups to work out what’s going on.
As one of the presenters in New York said “The lab rats are always right”. The web allows us to learn faster and make adjustments more easily than ever before so put on your white coats, get into the lab and experiment.
