In this post
Most lead-scoring setups treat the form fill as the moment intent appears. It isn't. It's the moment intent gets paperwork. The real decision — the part worth a same-day call — usually happens one or two visits earlier, quietly, with no form anywhere near it.
This post walks through the pattern we see across customer data, explains which signals show up before the form, and shows how to turn them into a scoring model your reps will trust.
What a buying journey actually looks like#
Here is a journey we see, with small variations, in almost every workspace:
- Day 1. Someone arrives from a search or a colleague's link, reads the homepage and one product page for about ninety seconds, and leaves.
- Day 4. They come back directly — no ad, no email — and go straight to pricing. They leave again after two minutes.
- Day 6. A second person from the same company visits the integrations page and the security page.
- Day 9. The original visitor returns, reads a case study, revisits pricing and fills out Talk to sales.
By any reasonable definition, that account made up its mind somewhere between day 4 and day 6. The second person checking security and integrations is a classic sign that a shortlist is being validated internally. The form on day 9 is a formality.
A form fill tells you someone is ready to talk. The visits before it tell you they were ready days ago.
Why the delay costs deals#
Scoring only the form means you learn about this account on day 9. A well-instrumented site could have flagged it on day 4 or 6. Four or five days sounds small until you consider what happens in the meantime: the buyer is also talking to two other vendors, and at least one of them is watching for exactly these signals.
In our own analysis of 2025 data across Team-plan workspaces, accounts contacted before their first form fill converted to an opportunity at roughly twice the rate of accounts contacted after it. Part of that is selection — the teams that act early tend to be better at sales in general — but the direction is consistent across company sizes and industries.
The signals that show up before the form#
Not every visit means something. The signals below consistently precede a form fill in the data, roughly in order of strength:
| Signal | Why it matters | Typical weight |
|---|---|---|
| Repeat pricing visits | Nobody reads pricing twice for fun. Two visits within a week is the single strongest early signal we see. | +20 to +30 |
| Return session within 72 hours | A direct return, with no campaign bringing them back, means the product is on their mind. | +12 to +18 |
| Security, integrations or docs pages | Evaluation has moved from “is this interesting?” to “will this work for us?”. | +10 to +15 |
| Several people from one account | Buying is a group decision. Multiple visitors from one domain usually means a shortlist. | +10 to +20 |
| Case study or comparison page | They're building the internal argument for a purchase. | +6 to +12 |
| Blog or general content | Useful for awareness, weak as a buying signal on its own. | +1 to +5 |
Two things stand out. First, the strongest signals are about repetition and recency, not volume: two pricing visits in a week beat twenty blog reads over a month. Second, several of the best signals — security pages, integration docs, a second visitor — have nothing to do with marketing content at all.
How much earlier you can act#
When customers switch from form-based scoring to behavioral scoring, we measure the gap between the moment an account first crosses the threshold and the moment it fills out a form (for accounts that eventually do). The median gap is 3.8 days. For accounts with more than one visitor it rises to about five days.
That window is where the advantage lives. A rep who reaches out on day 5 with “I noticed your team has been looking at our SSO and HubSpot docs — happy to answer anything before you decide” is having a very different conversation from one who replies to a form on day 9.
Behavioral outreach works best when it's helpful rather than creepy. Reference the topic someone was researching, never the exact pages or times. “Happy to help with the HubSpot setup” lands well; “I saw you visited pricing three times on Tuesday” does not.
Turning it into a scoring model#
You don't need a data-science team to act on this. A short, explainable model gets most of the value:
- Weight by intent, not volume. Give pricing, security and integration pages real weight, and keep blog and general content low.
- Reward recency. A repeat visit to a high-intent page within 24 hours should count for more; in Aurora it's worth 1.5× its normal weight.
- Let old interest fade. Scores should decay when nothing happens, so a burst of activity from three months ago doesn't look like a hot lead today.
- Keep the form as a signal, not the trigger. A form fill still deserves a big weight — it just shouldn't be the only way into the queue.
- Review after two weeks. Ask your reps which alerts weren't worth their time, then adjust.
The scoring documentation covers every signal type and multiplier in detail, and What we learned rebuilding the scoring model explains why we moved away from counting page views.
The takeaway#
Forms still matter. They're the moment someone explicitly asks to talk, and they should always reach a person fast. But if they're the only thing your scoring model listens to, you're always arriving a few days late to the conversation that matters most.