A lead scoring model built on fit and behavior

In a technical sale with a long, consultative cycle and a buying committee on the other side, most early signals are ambiguous. A content download from someone outside the market and a pricing-page visit from a buyer at a target account arrive as the same kind of record.

Without a shared score, lead quality becomes a judgment made record by record, and campaign results are hard to read in pipeline terms. A company selling a technical product into long cycles cannot hand every form fill to a seller.

I split the score into two halves and made both visible. Fit answers whether the company is worth selling to at all. Engagement answers whether this person is doing something that suggests a buying conversation now. A lead needs both, and a seller reading the record should be able to see which half is carrying it.

Three more decisions followed. The threshold would be set with sales, because a score only works if the people receiving the leads agree on what it means. Engagement would decay, because in a cycle measured in quarters, last month's webinar says little about this month's intent. And routing would live in Salesforce, where the sellers work, while scoring lived in HubSpot, where the signals land.

1. A signal lands. A visit to a high-intent page, an email click, a download, a form fill or an intent spike adds engagement points to the person.
2. The company is graded. Fit comes from the company record: industry, size and account tier, pushed down to every contact at that company.
3. The threshold is crossed. The lead becomes an MQL, with the grade and the engagement history that got it there on the record.
4. The owner hears about it. Slack and email alerts go to the owner the moment the threshold is crossed, not at the next list pull.
5. Salesforce assigns it. Routing on the Lead object puts it in the right queue by tier, where the BDR sees it in ranked order.
6. Reporting reads the same fields. The dashboards count MQLs, SQLs and pipeline by the same score and grade the routing used.

Once the score lived on the record, conversations about lead quality started from the same data. A lead that reached a seller carried its grade and its engagement history, and the seller could see why it had arrived. A good-fit account that had gone quiet and an engaged lead outside the profile looked different on the record.

The model sat between two other builds. The account universe supplied the fit half, and the account tiers set how much coverage each lead received once it qualified.

• Show both halves of the score on the record. Sellers trust a number they can take apart.
• Agree the threshold with sales before launch, then treat it as a lever: revisit it where pipeline is reviewed.
• Decay is required in a long cycle. Without it, scores only go up.