Lead Qualification Score

1) Qualification is a decision, not a leaderboard

A lead score is useful only if it explains a bounded decision. If one number mixes fit, activity, authority, timing and serviceability, a highly active but commercially impossible lead can outrank a quieter lead that can actually buy. This lesson builds Lead Qualification Evidence Scorecard v1.

2) Separate interest from suitability

Salesforce Trailhead separates scoring signals such as engagement from grading signals that describe prospect fit. Keep that separation visible. Intent answers “how much current buying activity is visible?” while fit answers “how closely does this prospect match the customer profile we can serve?” Neither substitutes for the other.

3) Run hard gates before weighted scoring

Evaluate hard disqualifiers and serviceability first. An excluded geography, prohibited use case, unsupported delivery requirement, or explicit non-serviceable state cannot be repaired by more clicks, downloads, or webinar attendance. A gate failure is evidence, not a low score waiting to be averaged away.

4) Authority is its own field

Interest is not buying authority. Record whether the buying path contains an accountable decision-maker or an evidenced route to one. If authority is explicitly absent, use NURTURE; if the authority field is required but unknown, use HOLD until the evidence exists.

5) Timing is not urgency theater

Use an explicit timing signal such as target purchase window or committed decision date. Do not infer a near-term purchase merely from repeated activity. A strong-fit account with a distant timing window can remain NURTURE without being treated as a bad lead.

6) Weighted score is advisory and versioned

After gates pass, a numeric score can summarize declared dimensions. In this lesson the example weight set is Fit 40%, Intent 30%, Authority 20%, Timing 10%. These are MAKNOW design values, not Salesforce standards. Record score_version and weight_set; a silent weight change makes old and new totals incomparable.

7) The scorecard fields are the audit trail

The row must show Lead, Fit, Intent, Authority, Timing, Serviceability, Disqualifier, Weight Set, Score Version, Missing Fields, Confidence, State, Rationale, Review Trigger and Evidence Ref. If the row cannot explain the state without hidden reasoning, it is incomplete.

8) State vocabulary prevents false precision

Use QUALIFIED when all hard gates pass and the declared versioned criteria are met. Use NURTURE when evidence is valid but the buying state is not ready. Use DISQUALIFIED for explicit hard-gate failure. Use HOLD, PARTIAL, DISPUTED, STALE or BLOCKED when the evidence contract itself is not ready for a positive decision.

9) Worked pair: identical activity, different commercial reality

Lead A matches the ICP, is serviceable, has authority and a 60-day timing window; its engagement is moderate. Lead B has very high activity but is currently unserviceable. Lead A may be QUALIFIED; Lead B is DISQUALIFIED even if its raw engagement score is higher. The ordering comes from evidence gates, not excitement about activity.

10) Wrong approach: double activity without changing fit

Doubling intent activity increases the advisory score but does not improve fit, authority or serviceability. If fit remains below the declared minimum, the state remains NURTURE. The correct explanation identifies the unchanged limiting dimension instead of celebrating the larger number.

11) Missing evidence is not a negative fact

Unknown authority is different from “no authority.” Unknown required evidence produces HOLD, because the system does not yet know enough to decide. This prevents a guessed zero or guessed yes from contaminating the qualification record.

12) Version changes and stale evidence invalidate comparability

If weights change from W1 to W2 without a new score_version, mark the result DISPUTED. If decisive evidence is older than the declared freshness limit, mark it STALE. Historical evidence can remain stored while the current decision is withheld.

13) Lab: eight fixtures and 64 deterministic assertions

The offline lab tests a qualified baseline, high-intent poor-fit lead, serviceability failure, missing authority, unversioned weight change, stale evidence, the executive-trigger transfer, and a newly serviceable lead that still lacks buying authority. Each fixture exposes eight deterministic outputs.

14) Transfer: recompute changed evidence, do not copy labels

The low-intent ICP lead receives an executive trigger event: intent and timing improve, so recompute under the same score version and gates. The high-intent lead becomes serviceable but still lacks buying authority: serviceability is repaired, but the authority field still prevents QUALIFIED; its bounded state becomes NURTURE, not success by momentum.

15) Delivery checklist

Fit, intent, authority, timing and serviceability remain separate; hard disqualifiers run before advisory scoring; weights and thresholds are versioned; missing data and confidence are explicit; every state has rationale, review trigger and evidence reference; no activity metric can hide a gate failure; and changed evidence is recomputed before the state changes.