Software Growth

Lead scoring

Lead scoring gives each lead a number based on how well they fit your ideal customer and how they behave, so you know who to contact first.

Lead scoring assigns points to every lead so you can sort them by how likely they are to buy. It replaces gut feel with a rule, and it gives marketing and sales one shared answer to the question "which leads are worth a call today."

It pays off when you have more leads than time. If you get five signups a week, read each one yourself. At fifty or five hundred, a score keeps you from spending your best hours on the wrong people.

The two kinds of points

HubSpot's lead scoring guide splits scoring into two parts, and most models use both.

  • Explicit (fit) scoring. Data the lead gave you or you looked up: job title, industry, company size, tools used. It answers "are they the right kind of customer?" See ideal customer profile.
  • Implicit (behavioral) scoring. What they do: page visits, email clicks, content downloads, demo requests. It answers "are they interested right now?"

HubSpot also recommends negative scoring, which subtracts points for signals like a free webmail address on a B2B form or obviously fake form data. Its sample point values give 20 points for a demo request, 15 for a pricing page visit and 5 to 10 for a content download. They are illustrative, not a benchmark.

A simple scoring example

Take a B2B tool that sells to agencies. You might score like this:

  • Agency with 5 to 50 employees: +20
  • Decision-maker title (founder, head of): +15
  • Pricing page visited twice: +15
  • Demo request: +25
  • Webmail address: -10

Add them up:

If your threshold for becoming a marketing qualified lead is 50, this lead qualifies, and a human should follow up the same day.

How to set the weights

Start from history, not theory. Pull your last 30 to 50 customers and note what they had in common before they bought: company type, role, pages visited, source. Give the strongest signals the highest points. Then test the score: do leads above the threshold close at a higher rate than leads below it? If not, change the weights. Revisit every quarter.

Mistakes to avoid

  • Over-weighting low-value engagement such as email opens, which are noisy.
  • Letting scores only go up. Subtract points for inactivity so old leads fade.
  • Building a 30-rule model with 20 leads a month. You do not have the data to tune it.
  • Ignoring in-product behavior. If you offer a trial, usage data predicts better than page views, which is the basis of a product qualified lead.

For small teams

You can score in a spreadsheet or CRM with five or six rules. The goal is to separate the top 10% of leads, so you can contact them first. The model does not have to be right, only better than reading the inbox in order.

Sources

Back to the SaaS glossary