5-Minute Experiment That Improved My Conversion Rate by 53%

Hey all,

Recently, I’ve been experimenting with ways to improve blog conversion rates for my clients.

As promised, whenever I get results (good or bad) I’ll share them with you.

Here's one that's showing positive signals (+53% from session to trial sign-up).

💡

Disclaimer: I’ve done my best to set up these experiments in a way that gives confidence in the outcomes…but the same results are not guaranteed. So I encourage readers not to treat these results as gospel—but as inspiration for things to test.

The Background

I installed a click mapping tool on a client's site recently and noticed something interesting.

The article in question was a “Best X Software” listicle (e.g. something like ‘10 best productivity apps for iPhone’).

It was performing pretty poorly (0.85% session conversion rate) and still could be better.

The click map tool showed a significant number of visitors were clicking internal links in the introduction. These links took the reader to top-of-funnel content like “why productivity matters”.

Recreation for the sake of clarity (while keeping client confidentiality).

My hypothesis:

I hypothesized that having links too early in the content causes some visitors to leave the page before learning about our product.

Might they be getting distracted and failing to return? They may have started reading other content and ultimately decided they didn’t need a product like ours anymore. Or they may have simply got bored and moved on to another task. Or, none of this “logic” was true, it just sounded good in my head.

The Experiment

To test my theory, I simply removed those links from the intro. Everything else was kept exactly the same.

I wanted to completely rewrite the intro and some other bits of the article. It pained me not to. But for the sake of the experiment I literally just deleted the hyperlinks.

The Results

I had made some other unrelated changes around the 7th of Jan, so I monitored the week between Jan 8th and Jan 13th as the control group.

After the changes were implemented the session conversion rate ("The percentage of sessions in which any conversion event was triggered.") increased by 53%.

Not bad signals for a 5-minute change.

Potential Issues With The Experiment:

Volumes are low. The conversions moved from 17 in the control week to 28 per week in the following weeks. This could be a natural fluctuation or a random lucky coincidence. But the fact that the following week (15th to 20th) and the week after (21st to 27th) held steady at 1.3% is promising.
This is a before/after experiment. I’m working on being able to A/B test these things in parallel because we can’t really be confident in the results if the experiment isn’t happening at the same time. For example, the first week could’ve had a lower conversion rate because people were busier at work after Christmas.

I’ll update this page after two more weeks. Until then, happy experimenting.

—Benny