I read an article recently from Roger Dooley at Forbes titled, “Never Redesign Your Website Again—Really!” Making the case for incremental site changes, Chris Goward of Wider Funnel is quoted explaining it is the only approach to fully and accurately determine the effect of each element change. Though a wholesale redesign, “throw the baby out with the bathwater” approach, is more popular, Dooley and Goward argue it involves more risk and produces less effective results. I read this article nodding in agreement the entire time. Yes! This measured approach is one every company should be thinking about applying.
The concern I have, however, is not with the logic of the approach to developing more effective customer experiences; it’s in how realistic the situation is. I imagine most people thinking, “Yeah, that’s great and all, but I don’t have the tools or the seniority to make this change.” The reality is that while iterative redesign makes sense in an ideal scenario, most enterprise-level companies have a lot of work to do before such a strategy becomes palatable. At Brooks Bell, we know this because we work with our clients through redesigns all the time and despite all efforts to stop an entire site redesign, the speeding train of change can rarely be slowed. It’s not for lack of detailed numbers to support an incremental approach. We have run countless custom reports to model the benefits of learning and iterating—demonstrating improvements to both customer satisfaction and revenue.
We can show how the only best-case scenario of a redesign is that conversion numbers do not drop. But more often, the process results in many wasted hours leading to the launch of a new site that underperforms its predecessor—and no clue as to which elements or changes cause the dip. Unfortunately, most traditional CMOs are apathetic to this reality. They understand the numbers but small incremental changes to the website don’t offer the visual splash they want for their next meeting with the CEO.
There are a lot of resources outlining the various ways to argue against a complete redesign. This strategy is unrealistic and impractical. You may not be able to stop a site redesign but you certainly can use it to your advantage. Here are 5 ways to be data-driven during a redesign:
1. Benchmark your Website Data
While the rest of your organization is scrambling to launch a new website, you need to measure every part of the current site before the switchover. This data is important for a number of reasons. When the new site launches you will see changes in your numbers. You need to understand what normal traffic and conversion metrics look like so that you don’t make an uninformed decision post-launch.
For example, one of our clients launched a brand new website right before the important Back-To-School shopping period. The launch was a huge success! Traffic was up from the last week of the old site! Sales were up too! It was surely a win for the creative team! The re-design had spiked sales! But when we looked at year-over-year numbers, we saw that the launch of the new website coincided with a similar traffic spike the year before. It just so happened that the re-launch coincided with a new advertising push and a time of year that had more site visits. And when we dug even deeper we discovered that the conversion rate declined year-over-year, proving the redesign was not as effective as originally thought. Having this historical data is crucial to making the case to restart your testing program immediately after re-launch.
2. Fix tool implementation issues during development
One of the benefits of a site redesign is the ability to implement your tools in a more effective way. Many times new analytics and testing tools are shoe-horned into an existing framework that may hinder site performance, slow page-load, limit the number of new tool features that can be utilized, or restrict where testing can be done on the site! A site redesign typically means brand new code and an opportunity to re-implement your tools the right way—a way built for greater flexibility moving forward. Make sure to communicate your current tool limitations early on in the process so that fixes can be worked into the proper requirement documents. We have seen optimized testing tool implementations increase a site’s page-load time, resulting in performance lifts.
3. Recode the new site so it’s ready for rapid testing
Another common point of frustration with old site designs is the disorganization of element IDs and DIV classes. While this may not have been a focus when the old site was created, a redesign typically means a recode! Grouping elements in code more effectively using IDs will make testing easier by enabling you to isolate key element groups for specific tests. Make sure to meet with the development team early and discuss which elements of the site you will focus on for testing. This will ensure the redesigned site will be coded in a way that maximizes your testing flexibility for the future.
4. Run tests using redesign elements on the existing site
One for the most important things you can do during a site redesign is to incorporate some of the new elements into tests on the existing site. This is incredibly valuable to the redesign process and to the testing program. By incorporating key redesign elements into the existing site experience you can see the effect that element has on your existing traffic. If it is a drastic design change, this incongruence of experiences may not be recommended but if there are other changes like taxonomy, checkout steps, or a new approach to upsell language, it can be tested. Testing these new concepts can validate the design change for the redesign or give the design team a chance to tweak their design before launch.
5. If the redesign fails, don’t call for a reversal
I get it. You may be against the redesign from the start. You may secretly hope it fails, to be able to say, “I told you so.” And you may even have a well-prepared PowerPoint ready to go to justify reverting back to the old website. But that’s not the answer. Undermining the hard work and effort of a redesign (even if decisions were gut-based, not data driven) will do more harm than good. Many redesigns do show depressions in conversion metrics out of the gate. Many have technical issues after launch. Data isn’t tracked. There’s the single user on IE7 that can’t render the code correctly. A redesign is no different than any other change in market conditions. The way to navigate them successfully is to have a testing program in place that can effectively and accurately identify the cause of the issue and optimize the experience for the future. The most important requirement for building a testing culture is to have the trust and support of your peers.
While incremental changes from iterative testing and optimization is conceptually a great idea, it doesn’t always mean lead to a site experience. You can influence behavior in a lot of ways using a lot of different tactics. In the purest form of our craft, this is effective. But in some cases it can create a Frankenstein effect as the overall site experience is pulled in different directions, patched with winners from different campaigns. The next thing you know, you have a website that is functional and performs well but is ugly. Some may argue this is fine! Ugly wins sometimes, after all. But in the eyes of marketers, C-suite executives, shareholders, and the ultimate boss—consumers—it may not be appropriate. It’s time we embrace the site redesign so that we may approach it in a smarter, more data-driven way.
Brooks Bell helps top brands profit from A/B testing, through end-to-end testing, personalization, and optimization services. We work with clients to effectively leverage data, creating a better understanding of customer segments and leading to more relevant digital customer experiences while maximizing ROI for optimization programs. Find out more about our services.