It was Tuesday, I remember... a guy named Dev sent me his demo page. He sells scheduling software to dental offices, mostly in his postal ranges. The whole page was three lines. I counted twice because I didn't believe it. Line one said "Request a Demo." Line two said "Fill out the form and our team will be in touch." The button said "Submit." "Be in touch" could mean a quick call. It could also mean nine emails from a guy named Kyle. The visitor has to guess, and people don't fill out forms when they have to guess. So we fixed it live, one small change at a time. Change 1. say what happens The headline became "See Chairtime work in your office." The subtext became "Book a 20-minute walkthrough with our team." Now you know what you get and how long it takes. But it still talks about the software. Nobody wakes up wanting to look at software. Change 2. talk about the person The new headline: "See how your front desk could stop losing 6 no-shows a week. Takes 20 minutes." The subtext: "We'll use your real schedule. No slideshow." That last bit is on purpose. Everyone who books a demo is scared they'll get a scripted pitch from someone reading off a screen. Say the fear out loud and it shrinks. Change 3. the button "Submit" became "Book my walkthrough." Submit sounds like taxes. Book my walkthrough sounds like you're getting something. Here's what I noticed after all that. The old page was about what Dev wanted, which was a lead. The new page is about what the visitor gets, which is 20 minutes that are actually about their office. Same form, same product. There was a fourth change, and I've been thinking about it for a week. Dev deleted one field from the form. Bookings went up. It was the field every sales book tells you to keep, and I still can't fully explain why it worked. Which field do you think it was? I'll tell you if someone guesses right.

"Submit sounds like taxes" is going to stick with me. Change 2 is the one most of us builders miss: we describe features instead of the front desk's real problem (6 no-shows a week). I build POS and business tools, so I'm taking this straight to my own demo pages.

The bot once scored “just browsing” as hot.

Early scoring leaned too hard on page visits alone someone reading the pricing page out of curiosity got flagged as ready to buy. Rebuilt the scoring around actual answers, not just clicks.

→ A page visit is curiosity. An answer is intent. They're not the same signal.

Switching the model to weight answers over page views really separates curiosity from intent, so you catch the buyers before they slip through.