How I increased our landing page registration rates from 18% to 36%
In 2019, Venngage had ambitious revenue goals, and acquisition was determined to be one of the levers that we can pull. After surveying the lay of the land, it became clear that our landing pages (LPs) were an untapped opportunity. Our keyword ranking was driving excellent traffic to our landing pages, but our registration rates remained curiously low.
I was tasked with figuring out why. The initial perception was that the LP design was "stale" and needed a "refresh." But after a bit of digging, the reason was entirely different.
Establishing a baseline
The first thing I sought out to do was to figure out what was actually happening versus what we thought was happening. I started with a few basic questions:
•
What's the traffic like?
•
What are the current registration rates?
•
When was the last time we redesigned them?What happened?
•
What does the user journey look like?
Finding True North
Turning to Google Analytics I got a set of baseline metrics for traffic and registration rates on all of the LPs in a defined 7-day period. There was quite a range in registration rates. I needed to put this into context.
Comparing the registration rates of the best and worst performing LP against the average
Desk research
To put this into context, I did a quick round of desk research to know whether this is "good" or "bad". Unbounced, a thought leader in the space, published a
Landing Page conversion rate benchmark report, and they deemed these rates quite effective.
Lead generation conversation rates by industry from the 2017 Unbounce report
Looking at a myriad of industries, the median was between 2.8-6% while the best performers were between 11-25%
At an average 18% conversation rate, we could've easily left our LPs as is, but it was clear to me that from our data there was room to grow.
The last redesign
Being asked to "redesign" something without a clear goal is always a red-flag for me. I wanted to know if we've ever attempted that before and how it performed. I did a little bit of digging and realized that 1 of the 50+ LPs does in fact have a completely different redesign. We were running an A/B test all along!
Again, turning to Google Analytics, I looked at a 4-week period before and after the redesign. The results were surprising.
Average registration rate before that LP was redesigned
Average registration rate after that LP was redesigned
The first week after the redesign, the registration rates were ever so slightly higher but then quickly plummeted to a low of 11.7% 😱
With a baseline established, I was set on figuring out what worked and what didn't.
Mixing quantitate and qualitative data
I decided to mix both quantitative and qualitative data to finally understand the problem at hand. I conducted a
5-second Usability Test and a
heatmap study using Hotjar to understand what users
thought of our LPs and what they
actually did. I also dug into Mixpanel to better understand our funnel.
The 5-Second Usability test
With a landing page, the goal is to convince users as quickly as possible of its value. The 5-second usability test is often cited as an effective and quick measure of user's first impressions—the main driver of whether a user will stick around or not.
A 5-second usability test asks users what they think of a website after only showing it to them for 5 seconds.
A 5-second usability test asks users what they think of a website after only showing it to them for 5 seconds. I ran a 5-second usability test with 5 users on three landing pages picking a mix of high, low, and average performing sites.
The results came back really positive:
Word cloud of the words users used to describe the LPs
Heatmap study
In parallel, I ran a heatmap study with Hotjar to understand what users clicked on and how they scrolled.
More than 50% clicked on the hero CTA
Hotjar heatmap of the "Brochure Maker" LP
And at least 75% of users didn't scroll past the hero
Hotjar scroll map of three LPs
What the funnel?
While everything seemingly pointed me to concluding that nothing was wrong with our landing pages, I decided to look at the entire user journey.
I looked at conversion rates at every step of the way and quickly found a massive drop in conversion.
A funnel looking at the conversion of each step of the user journey
It's all in the funnel
It became evidently clear that our goal—registration rates—was being significantly affected by our funnel. Before embarking on a "refresh" I suggested test a hypothesis: What if we asked users to sign up earlier?
What if we asked users to sign up earlier?
What the flow looked like before:
What I was proposing:
Testing my hypothesis
To test my hypothesis, we deployed this change to only 1 of our landing pages and monitored it for 1 week.
Variant registration rate increased from 15.01% to 24.67%
Given the promising results, we released the change to the remaining landing pages—all 50+ of them.
Average registration rates increased 18% to 34% for 50+ landing pages
