Reviving Losing Split Tests To Increase Sales Conversion Rates

Unfortunately, not all tests can be winners.  If this were the case then our company would not exist.  But it is important to understand that although some variations may be losers, you shouldn’t count them out completely.

Today we are going to look at a specific case study where this was the case.  How small changes in button design on a previous “losing” variation helped increase sales conversion rate.
We already know that buttons play a big role in your conversions and changing them can increase sales conversion rates.  Of course, buttons aren’t the only thing you need to test but they often have a big impact.

In this particular case, we had been running our optimization plan on a specific landing page for a client.  We had already seen some nice improvements (60%+ increase in a free-trial take rate) and in an effort to further the improvement, because optimization is never done, we redesigned the landing page.

The previous tests showed the new design had promise but ultimately didn’t win the race as the results showed only slight improvement. Not enough to declare it a true winner.

This was for a free trial to one of their best selling products and is actually a big revenue driver for the company.

So How Did We Revive This “Losing” Variation To Increase The Sales Conversion Rate?

As mentioned we know that buttons play a big part in getting website visitors to take action and previous tests on the pages for this client showed improvement from button changes.

To revive this design to increase sales conversion rate we tested 4 new buttons to see if this slight change could bring the client a nice lift.


The buttons are in order from top to bottom starting with the control, button 1, button 2, etc…

Can you guess which one helped increase the sales conversion rate for this particular landing page?

The results are actually quite interesting.  Button 1 decreased sales conversion rate by almost 65%, button 2 showed a marginal increase, Button 3 and 4 showed a 20%+ decrease, while button 5 showed to increase sales conversion rate by 28.8%.

You will notice that we tested multiple button phrases here as well.  This was done on purpose since we had tested many different elements already we wanted to see a bigger impact to gather data to set us up for more test ideas to run to increase sales conversion rate even more.

We have a little bit of time left on this test to get conclusive statistically significant results but we are close.
The next steps here will be to test language and different variations of the “winner” such as gradients, wording, and highlights to see if we can further increase sales conversion rate.

The big lesson here is to never count out a losing variation in a test.  Markets are ever-changing. What didn’t work a couple of months ago, might work today.  This is why you constantly have to be testing and optimizing.  This case study proves that even though we “lost” before, doesn’t mean we can’t win today.

Be sure you are constantly testing and even revisiting old tests to test again in an effort to increase sales conversion rates.

Be sure to like, comment, share and subscribe if you liked this case study.

If you would like help figuring out what you should be testing and how to improve your overall results, be sure to reach out to schedule your no-obligation conversion acceleration session.