SuperAPI AI Directory

lead-scoring

Marketing & growth

View repository

Updated 10-16 06:06

About this skill

Lead Scoring

Score a lead base against a defined ICP with a 100-pt rubric. The rubric must be adapted to your ICP and approved before you run it on a real base — swap the example bands below for your own firmographics, geos, industries, signals, and buyer titles.

Hard rule

The methodology (weights + thresholds) must be approved BEFORE scoring a real base. If it isn't confirmed, say so and score only as a labelled DRAFT preview. Never silently run on unapproved weights — that produces wrong scores.

Procedure

  1. Anti-ICP gates first. If any fires → Disqualify, skip points. Define your own hard disqualifiers (e.g. wrong geography, strong in-house capability, deal size below floor with no recurring revenue, service you don't actually deliver, culture/policy block). Any gate hit = out, regardless of other points.
  2. Score the categories (100 pts max). Pull each ONLY from provided/verifiable data. Missing → 0 + data-gap flag. Never guess firmographics, signals, or titles.

Example weighting (tune per ICP — weights should sum to 100):

Category Max Bands (example)
Deal potential 20 T1 20 / T2 10 / T3 5
Geo 15 focus region 15 / follow-up region 10 / else 0
Industry 10 core-fit industry 10 / adjacent 5 / none 0
Headcount 5 sweet-spot size 5 / near 3 / too small 1
Services match 10 do-best 10 / understand-no-proof 5 / none 0
Service-need signal 15 active buying signal 15 / capability gap 10 / legacy tooling 5 / leadership change 5 — take HIGHEST, don't sum
Digital maturity 5 mature 5 / mid 3 / immature 0
DM title 10 true decision-maker 10 / influencer 5 / specialist 0
Engagement & Intent 10 high-intent page 10 / newsletter 10 / visit 2
  1. Tier from firmographics (industry + size + service fit), not assumption.
  2. Bands → action: High 75–100 call · Medium 60–74 conditional (core dimensions must match) · Low 40–59 nurture · <40 or gate = Disqualify.
  3. Each signal carries source + date. Freshness matters: recent hiring signals are hottest, tender/RFP wins stay warm ~2 weeks, most signals decay by ~30 days.

Output

One row per account: Company · Tier · per-category points · Total · Band · Action · Data-confidence(H/M/L) · Missing fields. Plus a data-gap list of accounts needing enrichment before a reliable score.

Don't

  • Don't round up on doubt — take the lower band.
  • Don't invent missing firmographics, signals, or titles.
  • Don't score a real base on unapproved weights.