Lead scoring

How lead scoring works

A scoring model gives each lead a number. Points go up for what tends to come before a sale and down for what tends to come before a dead end. When the number crosses a threshold that sales and marketing agreed on, the lead goes to a rep. Below it, the lead stays in nurturing.

The number has two parts, and the scoring tools keep them apart:

• Fit is who the lead is: job title, company size, industry, region. HubSpot calls these fit scores and builds them from the values on a record.1
• Engagement is what the lead does: visits to your site, clicks, a booked meeting, a reply. Adobe's Marketo guide calls it behavioral scoring and names a visit to the pricing page as a sign of intent.2

Keep them apart even if you add them up at the end. A perfect-fit buyer who never opens anything and an intern who clicks every link can land on the same total, and they need very different next steps.

An example model

A team that sells to sales leaders at software companies of 50 to 500 people might start with this:

• Title. Head of sales or VP of sales, +20. Any other sales role, +5.
• Company. Software with 50 to 500 people, +15. Under 10 people, -10.
• Actions. Opened the pricing page, +15. Replied to an email, +25. Nothing in 90 days, -10.

Negative points matter as much as positive ones. Adobe's guide uses them for a poor fit and for long silence, and its example hands a lead to sales at 50 points.2

How to keep scores honest

• Build it with sales. Ask reps what last quarter's closed deals had in common, and score for that.
• Let old activity fade. A pricing visit last week says something. One from last spring says very little. HubSpot lets engagement points decay over 1, 3, 6 or 12 months.1
• Check the bar. If reps reject most of the leads they get, the threshold is too low. If they sit idle, it is too high.
• Show the reason. A rep acts on a score faster when it says why.

In GetSales you score each row with an AI column you define, before anything sends. It keeps its reasoning beside each answer with how sure it is, and you can try the column on one row first. A first line set to run on a good fit waits for the score.

Related

• Data enrichment: work emails and an AI score on each row, with its reasoning.
• Buying signals: the actions an engagement score counts.
• Signals: plays for website visits, posts and job changes.
• AI SDR: it reads replies, scores intent and drafts the first answer.

Sources

1. HubSpot Knowledge Base, Overview of the lead scoring tool, last updated 23 October 2026. Read 28 October 2026.
2. Adobe Experience League, Building a person scoring model (Adobe Marketo Engage), last update 19 August 2024. Read 28 October 2026.