Here's a realistic scenario: it wasn't so long ago that your website was riding high. You were attracting plenty of visitors, building up a following for your blog or newsletter, and - most importantly - your visitors were converting well into customers. Recently though, your upswing is turning into a downswing. Traffic is down, new content doesn't seem to be gaining the traction it once did, and actual sales are on a steady decline. Taking an honest look, maybe the site is looking a bit old and tired - might it be time for a revamp?
Before diving into a redesign and rebuild of your website from the ground up in an attempt to claw back missing business, it's best to consider a more measured approach, evolving your site rather than revolutionizing it.
Knowledge is Vital
Even with the best analysis and analytic tools it's possible to end up with only a partial picture about user interactions throughout your site. Sure, you can measure page views, traffic funnels and bounce rates, but all too often you have little more than guesswork to go on when it comes to determining why a particular part of your site is working well or otherwise.
Unless you have a very clear idea about what needs to change, a site rebuild will often merely substitute one set of unknowns for another. Of course, you might get lucky with a new design and see sales take off again, but wouldn't you rather take a more scientific approach with a greater chance of long term success?
Evolution = Tweaking, Observing & Repeating
If you take a look at some of the web's largest sites in their older incarnations from just a few years ago, you'll often find some similarities in design and user interface patterns (check out Wayback Machine or Google another website archive service).
Cisco.com in 2013:
Cisco.com in 2016:
The modern design might be a bit slicker, the layout might be bigger, bolder and cleaner, but in general you won't see major sites abandoning and replacing their entire look-and-feel overnight. Drastic changes can alienate existing users with no guarantee of attracting new ones and it can rule out two important and powerful methods of enhancing and improving your site's performance: A/B split testing and its more advanced version, multivariate testing.
Enhancing with A/B Split Testing
With A/B testing, you can measure one page (version A) against another (version B) then compare performance and serve up the winning page as part of your site. A/B split tests are pretty common because they remove the burden of "guessing" if a new page design will work.
Enhancing with Multivariate Testing
Multivariate testing on the other hand, allows several different page elements like on-page copy and interactive features to be tested at the same time. Generally, this is done using clever algorithms or prototypes to separate out the effects of each change.
Both of these techniques use software to deliver different versions of webpages to individual users and over time your statistics can show which version is the more popular - or conversely, if the change you've just made is turning people off. If you're curious to learn about the differences between A/B splits and multivariate testing, including their advantages and limitations, Optimizely (a web personalization tool) does a great job. Business2Community also shares a fantastic overview on both tests in their Landing Page Fundamentals, Part 3: Landing Page Optimization series.
Become an Evolutionary Champ
As site designs become older or modern features and functions make their absence felt, most websites inevitably experience a tailing-off of conversion rates over time. When this happens to your site, it’s tempting to go for a full site redesign with all the bells and whistles, but sometimes it can be mistake. Taking a more measured approach by testing and deploying minor changes can result in more reliable and longer lasting improvements. Over time, you may well find you've arrived at a substantially new design albeit with the changes being checked for improvement every step of the way. There may be times when a full redesign is the only way forward, and in some cases, the slow and steady approach can provide more control and even longer term success.