The growth team had been running pricing experiments for a quarter. They'd tried discounting, free trial extensions, and urgency messaging. Nothing moved the needle meaningfully. Trial conversion sat at 8% and felt stuck.
I was asked to take a fresh look at the problem before the team ran another experiment.
I started with heatmaps and session recordings, not hypotheses. The pattern was clear: users spent most of their time on the pricing page not comparing plans — they were re-reading the feature list, scrolling back up to the headline, and then leaving without converting.
That told me the problem wasn't decision fatigue or price sensitivity. It was uncertainty. Users weren't sure what they'd actually get.
In 14 user interviews, the most common phrase was some version of: "I'm not sure if this is worth it for me specifically." Not "it's too expensive." Not "I need more time." Just genuine uncertainty about fit.
Our pricing page was written for a feature comparison. It answered "what does each plan include?" but not "what will change for you if you upgrade?" Those are very different questions.
"Users needed to see themselves using the product — not a grid of features they'd have to mentally translate into their own context."
We changed nothing about the price, the plan structure, or what was included. This was purely a communication change.
Before — feature grid layout
Replace src="images/pricing-before.png" with your image file.
After — use-case scenario layout
Replace src="images/pricing-after.png" with your image file.
Left: the original feature-comparison layout. Right: the reframed version with use-case scenarios and role-specific social proof.
Relative lift measured over a 4-week A/B test.
Up from 8.0% — sustained 6 weeks post-launch.
No discounts, no plan restructuring, no new features.
The best growth interventions often aren't growth interventions at all — they're clarity interventions. When users don't convert, the instinct is to make the offer more attractive. Sometimes the offer is already attractive; you just haven't communicated why yet.
Starting with session recordings rather than a hypothesis backlog was the right call. The data showed me where users were struggling before I had a chance to assume I knew what the problem was.