Digital marketing teams have long relied upon A/B testing as a well-established means of finding a website’s optimal layout. The approach is simple: make a change and monitor its impact on click-through-rates (CTR), conversion rates, and revenue. When it comes to A/B testing strategy though, typically these changes fall into two categories, each with their own advantages and drawbacks. However, there is an alternative.
Two Traditional A/B Testing Approaches: Iterative and Redesign
Traditionally, there have been two approaches to A/B testing: iterative and redesign. Both use the exact same methods of testing, comparing an original (or control) variation to new variations within a test audience. But their approaches to how these new variations are created differ greatly.
The Iterative Approach
The iterative approach is based on methodically testing different elements on a page. The variations in this tests are usually based on one or two element changes, so that their impact can be closely monitored with the A/B test. From these results, new decisions on what elements to change or keep are then made and tested again. So, finding that elusive combination of changes for maximum impact takes a lot of trial and error.
Pros and Cons
The iterative approach is the most pragmatic, methodical and precise way to measure the impact of each individual change. Unfortunately, it’s a slow process and its impact is often minimal due to the time it takes to finally identify that one change, or series of changes, that optimizes a page. Also, there are times when identifying the impact on a single change is problematic as audience preferences typically fluctuate daily and weekly.
The Redesign Approach
The redesign approach, in contrast, uses a completely different strategy. Instead of making small iterative changes, it combines many by retooling a page with new content, new images, fresh graphics, and perhaps new color themes. The strategy aims to make a bigger impact by combining many positive changes, as well as saving time by not pragmatically testing each change separately. Though an exciting and potentially rewarding strategy, it does have its pitfalls.
Pros and Cons
Obviously, a revamped design makes an impression on customers. It gets noticed and creates that all-important buzz. Unfortunately, redesigns aren’t always customer-focused. It’s common for an individual to drive that design based on their own personal preferences. It’s also risky. It takes a lot of time, effort and money to plan, initiate and create a new design; not to mention all the content upgrades and new imagery. With all that invested time and upfront work, once the new design is tested, it may perform poorly. At that point a tester has to make the tough choice of whether to continue down the new design path or go back to the drawing board.
A Better, More Balanced Approach
Our approach is different. It’s entirely customer-centric. We’ve tackled each of the drawbacks of the aforementioned approaches. Iterative was precise and systematic, but it was too slow and limiting for our customers. Redesigns were time-consuming and laborious due to the number of revisions and sign-offs that are common in the process of redesigning a website.
We needed a faster and agiler approach, one that provided our eCommerce, retail and fashion customers with the returns they deserved, but one where the risk of failure was minimized. The approach had to make an impact and create that all-important buzz. Well, we found that approach, and we call it the “Minimally Viable Redesign” or MVR, for short. So, what is MVR and how does it differ from a complete redesign?
Minimally Viable Redesign uses the approach of combining multiple changes to have a bigger effect. However, it minimizes the effort upfront by focusing on reusing materials and elements around the site. For instance, we don’t generate any new graphics. Often, a simple adjustment or a change in placement is all that’s required. We also avoid wholesale stylistic changes like buttons, layout, borders and fonts. These remain unchanged, and we focus on minimal CSS changes if needed. Another tactic is to hide entire page sections, or rearrange them with the A/B testing tool. Last, we don’t produce new content. We only tweak current existing content to fit our needs. These minimal changes combined together have a greater effect that mimics a complete page redesign.
Typical MVR Changes
For the MVR, a way to prevent us from going overboard is to focus on simple changes that have the highest potential for return. Usually, these fall under the umbrella of best practices, but not always. Overall, they key is to make sure changes have a low difficulty level so that if by chance a new variation doesn’t perform well, it can be easily tweaked and retested.
Some of these typical MVR changes include:
- Condensing the page and move eye-catching elements above the fold so that your audience is focused on what matters.
- Remove whole sections of a page that aren’t working
- Moving customer testimonials and other social proof elements from other parts of the website to areas that grab the most attention
- Making sure CTAs are clear and noticeable. Our approach to call-to-actions (CTAs) exemplifies this strategy
Ultimately, with MVR, you should have a page that feels completely different but still follows your current content and style guidelines. Here is a perfect example of the simplicity of this approach.
Venetian Luxury Suites Page (original and MVR):
- Added detailed subheadline (content from page)
- Added Forbes icon (trust badge)
- 18pt font size for intro paragraph (easier to read)
- Removed Featured Amenities section (too large)
- Separated out tabs of content (Floor Plan, Suite Details, Resort Amenities) into different sections
- Added special call-out section for Booking w/ Featured Amenities Image
- Hid Tour Resort
All of these changes were easy-to-implement, fast and efficient. Development time was minimal since it merely involved moving content from other portions of the page to areas more likely to catch the audience’s attention. We made simple adjustments to existing content already displayed, while removing redundant sections of content that didn’t add value.
Once tested, the new page design converted 14.6% better and generated 22.1% more revenue on Desktops. Not only were these simples changes successful, they also shortened the approval process since they were nothing more than tweaks to existing content and re-positioning images from other portions of the website. We were able to define the changes, get it approved and develop the new page within a few days.
For Elite SEM CRO team, MVR has proven to be a valuable approach to A/B testing that fulfills our customers needs to make a big impact with minimal upfront investment. The key is to focus on simple changes and prioritize those based on expected impact.
If you would like to see more about CRO check out our guide How to Convert: Top 10 CRO Tips.