This is a walkthrough of a homepage rewrite we ran late last year for an anonymized client (B2B SaaS, $7M ARR, developer-tooling adjacent). I’ll skip company-identifying detail; the framing and outcome are what’s useful to share.
The starting state
Homepage hero copy at engagement start (lightly paraphrased): “The modern platform for [category]. Build, deploy, and monitor — all in one place.”
Three sub-features below. A logo wall of customers. Generic CTA: “Get started free.” Conversion rate from visitor to free-tier signup was around 1.4%. Conversion from free-tier to paid was running at 6% over a 90-day window.
The framing problem was visible immediately: the hero didn’t say who this is for or what specifically it does. A reader who didn’t already know the product had no way to self-qualify. The CTA was generic. The logo wall had 12 logos — too many to scan, too few to overwhelm. The middle of the page was a list of features rather than a story.
The three hypotheses we ran
Rather than a single rewrite, we ran three discrete hypotheses sequentially, each with a 2-week measurement window:
- Hypothesis 1: specificity helps qualification. Changed the hero from category-level framing to specific buyer-job framing. “The [thing] platform for [specific buyer profile dealing with specific named pain].” Made the page less broadly appealing on purpose — narrower target.
- Hypothesis 2: social proof works better as a story than as a logo wall. Replaced the logo wall with a single customer’s named outcome and a real quote. Risk: lost the impression of breadth. Bet: increased the impression of fit.
- Hypothesis 3: feature list is the wrong middle. Replaced “three feature bullets” with “three before/after moments” — short narrative scenes describing what changes in the user’s workflow when they adopt.
Results, with caveats
The combined six-week rewrite delivered:
- Visitor → signup rate: 1.4% → 1.9% (+36%)
- Signup → paid (90 days): 6% → 8.4% (+40%)
- Average time on page: 38s → 71s
- Bounce rate: 64% → 49%
Caveats I want to be honest about:
- Sample size was modest — we didn’t have the traffic for a clean A/B test, so the comparison is sequential, before vs. after. Some of the lift could be regression to the mean or seasonal.
- The signup-to-paid uplift is partially driven by better qualification at the top of funnel — fewer total signups but better-fit ones. The absolute count of new paying customers per month went up about 50%, but if you’re a board member who scores on free-signup volume you’d have been mad in week 3.
- We made unrelated improvements to the onboarding flow during the same window. I can’t isolate them perfectly from the homepage changes.
What I learned
The biggest single contributor was hypothesis 1 — specificity. Narrowing the hero copy to a specific buyer profile lost some breadth but dramatically improved the quality of qualified traffic. The before/after scenes (hypothesis 3) was second-biggest contributor. The social proof rework (hypothesis 2) had the smallest measurable effect, though it might have had a longer-tail brand effect we couldn’t measure inside the window.
If I could only do one of the three on the next engagement, I’d run hypothesis 1.
— Liu
