Just fix it
May 2019
It’s easy to slip into the mindset that you should test everything - especially when you sell A/B tests for a living. But there are plenty of improvements you don’t need to test; you should just make them. Here’s a partial list, broken down by part of the site.
Forms
You can, and should, optimize the forms on your site. But if you’ve got broken validation, misaligned form fields, illegible labels, fields so narrow the text is hard to read, or a hard-to-find Submit button, just fix it.
Navigation
Navigation is optimizable too! You’ll need plenty of traffic, and it’s key that a significant proportion of visitors who convert actually use the navigation, but it can be done.
However, if you’ve got 15 items in your top navigation bar and your analytics tell you that visitors are only using 3 of them, go ahead and move the rest to the footer. If your labels are confusing, and you’ve heard this from user testing or customer complaints, just fix them.
Mobile experience
If your page isn’t responsive, if buttons aren’t reachable, or if you have paragraphs and form fields outside of the viewport, just fix it. The more mobile visitors you have, the more urgent this fix becomes.
Page speed
Test your site out on Google PageSpeed Insights. If you score a big fat F minus, take steps to address your page performance. An 80/20 approach is ideal - Google will tell you which fixes will deliver the biggest bump in load speed. So compress images, load some files asynchronously (not your A/B test snippet though!) and stop making your visitors wait.
Metrics
If your analytics install is broken, first of all please know that lots of analytics installs are broken. Almost all of them are at least a little broken. Just fix it.
In each of these cases, you could set up an A/B test to measure the impact of your fixes. There are two likely outcomes there:
- You measure lift, pat yourself on the back, and say “Yep, fixing our broken form validation helped.” Sure, but your form isn’t optimized; it’s just no longer broken.
- Your results are inconclusive, and you say “I guess the cluttered nav wasn’t such a problem after all.” You’ve improved your site from a usability and aesthetic perspective, but you’re unable to feel good about it because you didn’t get a win.
Neither of these results is as satisfying as taking decisive action, fixing what’s broken, and moving forward with optimization. So just fix it!
How about you? Are there items on this list that you think it’s better to test first? Are there other changes that you routinely make without testing? Hit reply and let me know.