What actually makes a lead hot? An honest look inside the score.
Lead scoring isn't a black box. It's a handful of signals weighted by what historically predicted a closed deal on your site.
When people hear 'AI lead scoring' they imagine a mysterious oracle. The reality is more boring and more useful. It's pattern matching on behavior, calibrated against what actually closed in the past.
The signals that matter most
Across thousands of B2B and DTC sites, the same behaviors keep predicting intent. None of them is decisive alone. Stacked together, they paint a picture.
- Pricing page views — especially repeat visits on different days.
- Time on page above 60 seconds on solution or feature content.
- Return visits within 7 days, ideally from a different referrer.
- Scroll-to-bottom on long pages — they're actually reading.
- Demo, contact or signup CTA hovers without click — hesitation is interest.
- Frustration signals (rage clicks, dead clicks) — friction we can fix.
How the model learns your site
On day one, the score uses a generic baseline that works reasonably well for most B2B and e-commerce funnels. From the moment you start marking leads as won or lost in the dashboard, the model retrains on your data. After a few hundred labeled visitors it knows your patterns better than any generic rule.
Logistic regression under the hood — not because it's fancy, but because it's transparent. You can see which feature pushed the score up and by how much. No mystery.
Why we show 'why this score'
A score without context is just a number you can't argue with. Every lead in Catch before they bounce comes with a plain-language breakdown: viewed pricing twice on consecutive days (+18), bounced from the demo page (-7), came back from a direct URL (+9). You always know what's driving the prediction.
What it doesn't do
It doesn't predict who you should hire, doesn't replace conversation, and doesn't know about the 6-month enterprise cycle nobody on your site is tracking. It tells you who, based on behavior alone, is most likely to convert in the next 14 days. That's it. That's the whole product.
Used that way — as a signal, not a verdict — a good score will free up half your sales week.
Ready to see Catch before they bounce?
Score every visitor 0–100. Spend your week on the 20% who already decided.
Begin quietlyKeep reading
Stop chasing everyone. Chase the 20% who already decided.
Most pipelines are noise. A small slice of your traffic is genuinely ready to talk — and you're treating them the same as the tire-kickers.
Session replay without the creep factor.
You can watch how people actually use your product without watching the people themselves. Here's the line.
