“I love the idea of testing, but only after we are done with our redesign.” This quote during a recent show gave me pause. Was she right? Should you get your site into its new set of clothes before you tune it?
After hundreds of tests performed by our customers, my answer would be: “Not only is testing essential, but the site redesign may be the last your company will ever do.”
“I love the idea of testing, but only after we are done with our redesign.” This quote during a recent show gave me pause. Was she right? Should you get your site into its new set of clothes before you tune it?
After hundreds of tests performed by our customers, my answer would be: “Not only is testing essential, but the site redesign may be the last your company will ever do.”
I don’t mean to make this heavy. It’s nearly Thanksgiving, with its Black Fridays (and Black Mondays) to think about, after all.
Testing is definitely important any time you make significant changes to branding, merchandising, marketing, and messaging. Since a site redesign can encapsulate all of these, then it is really the most important time to test. At one company, their site re-launch decreased conversion by almost 30 percent! That is not a reasonable risk to take for an activity that can be done relatively easily these days.
But the question nagged at me. A redesign of a site hints at changes made to a location, a place where visitors come, shop, and leave. It hints at a site that is relatively static, one on which changes are made en masse, at a certain, specified time.
Let me go back to an article I wrote in April, in which I claimed that the site is dead. My point was that, as keyword prices rise and landing page optimization becomes an art form, the smartest marketers will lead visitors through a “landing session” -- a series of pages that specifically targets an individual visitor’s needs, with a single message driving the visitor through the sales funnel -- rather than a website.
Yes, there will be common pages throughout sessions that will be seen by most visitors, but the visit will more closely resemble an experience than a pit-stop. In other words, sessions are driven by campaigns, where visitors are targeted depending upon the ad that drove them to the site, the keyword for which they searched, the visitor's previous behavior on the site, and so on.
In an online world where the site is dead, and the campaign is an entity that is changed and optimized on an ongoing basis, testing is the rule, not the exception -- because when content is changed on a regular basis, it's simple to add tests at every turn to optimize ROI. Imagine, for example, a home page on which every section is designated a content- or image-slot. As you have new ideas for copy, images, promotions, branding features, you rotate them into a slot, run a test, watch results.
What you call a site absolutely must start to be a tuned set of pages that uniquely serve the customers driven by specific source of traffic. For general traffic, your “site�? may need to be designed to convey useful information to help your prospect decide. For more high-intention visitors, balancing branding with specific product or service offers is essential.
But at this point there is no longer a justification for a wholesale redesign of every page, as this will likely destroy far more than it creates.
So I thought about the question: Is testing appropriate for a site redesign?
And I came to a conclusion: No. Testing is not appropriate for a redesign, because a redesign itself is not appropriate.