Guide · 31

Lead scoring

Scores and ranks your prospects against your own JSON rules (budget, size, source, age…), with the reasons for each score.

In one line

node scripts/31-lead-scoring.mjs prospects.csv --rules scoring.json --top 10

What it's for

You have 80 prospects and time to call 10. This script scores them against your criteria (budget, size, source, age of last contact…) and ranks them, so you know where to start. The rules live in a file, so they are visible and editable: no black box.

Before you start

  • A CSV of prospects (one header per criterion: budget, employees, source, country…).
  • A JSON rules file.

Step by step

  1. Write your rules: each rule adds (or removes) points when a condition is true.
    { "grades": { "A": 70, "B": 40, "C": 0 },
      "rules": [
        { "field": "budget", "op": "gte", "value": 500000, "points": 40, "label": "high budget" },
        { "field": "email", "op": "empty", "points": -15, "label": "no e-mail" } ] }
    
  2. Rank:
    node scripts/31-lead-scoring.mjs samples/prospects.csv --rules samples/scoring.json --top 10 --out ranking.csv
    
  3. Call the "A"s first; read the reasons behind each score.

Reading the output

For each prospect: score, letter (A, B, C; a negative score gives -) and the list of rules that applied. Operators: eq, in, contains, gte, lte, gt, lt, empty, notEmpty.

Automating it

Re-run after each CSV update; then compare the scores with reality (who actually bought) to adjust the weights.

Pitfalls

  • A score reflects your assumptions, not the truth: check it against your past sales.
  • Do not score on criteria that would be discriminatory or that you are not allowed to use.

This guide goes with a free script (MIT license). Try it on a copy of your data first.