The company
Diana (YC W24) gives small businesses an AI-powered HR and people-ops team, a human-in-the-loop service that handles compliance, payroll, benefits, and onboarding so SMB owners don't have to. Their buyers are busy business owners, which made every ounce of friction in the demo funnel expensive.
The problem
Diana's paid campaigns were doing their job. Clicks were coming in at a reasonable cost. The problem was everything after the click. Very few of those visitors ever booked a demo, and CAC kept creeping up quarter after quarter as the team scaled spend into a funnel that leaked.
When we dug in, the leaks were obvious in hindsight but invisible in the dashboards Diana was watching. Landing pages promised something different from the ads that sent people there. The booking form asked for too much. On mobile, a redirect in the booking flow was quietly killing most conversions before anyone even saw the calendar. And if a visitor didn't book on their first visit, that was it. No retargeting, no follow-up, nothing pulling them back.
More spend on top of that funnel just meant paying more for the same pipeline.
What we owned
Two numbers: cost per booked demo, and demos booked per month.
We ran one weekly loop, and every change was routed to Diana's team for approval before it shipped. Nothing went live that they hadn't signed off on.
What we actually did
First, a funnel audit: every step from ad click to booked demo, with a dollar cost attached to each drop-off. This mattered because where the money was leaking turned out to be different from where the team assumed. That audit set the order for everything else.
Then we fixed the leaks, biggest first:
Landing pages. Each page was rewritten to match the ad that fed it. Same message, same promise, proof near the top, one CTA. No more asking a visitor who clicked on one thing to care about five others.
The booking flow. Fewer form fields, calendar embedded directly on the page, and the mobile redirect gone entirely. This one change alone recovered a chunk of conversions Diana didn't know it was losing.
Follow-up. Visitors who engaged but didn't book started getting retargeted, with a short nurture sequence behind it. A first visit stopped being the only shot at a demo.
Spend. Once we could see which audience segments actually booked demos, not just which ones clicked, we moved budget toward them. Same total spend, pointed at the people who convert.
What happened
In the first 60 days, per Diana's own analytics:
CAC down 33%
Demos booked up 22%
Landing page conversion rate more than doubled
No budget increase. No new channels. Just a funnel that finally stopped leaking, and spend that went where the demos were.
If you’re ready to see what GrowthOS can do for you, get in touch.