# Prospect prioritization
Rank a supplied prospect set so a team can decide where to spend human effort. Use explicit criteria and preserve uncertainty; a long dossier is not evidence of fit, and a public signal is not proof that a buyer wants help.
Choose the run and intake
State whether the request is a **supplied-list ranking** or a **sourcing run**. For a supplied list, rank every supplied record and do not silently substitute a new list. For sourcing, confirm the offer, market, geography, candidate-count target, sources in scope, and date or capacity constraints; ask one focused question if the population cannot be identified. In both modes, collect the ideal-customer criteria, exclusions, and evidence. State default weights or thresholds before scoring.
Build a sourcing evidence base
For an explicitly requested sourcing run, first write a run brief and discover candidates from relevant public sources such as company sites, directories, association lists, public careers pages, and reviews. Maintain a source ledger with source, query or filter, access date, result count, kept/dropped count, and blocked or unavailable surfaces. Normalize obvious duplicates before ranking and label incomplete coverage as partial.
Pre-filter only clear exclusions. For each observation, preserve its URL, observed date, current-status evidence, and status: verified, likely, unavailable, or deferred. A search result, old listing, or weak web clue is discovery evidence, not proof of fit or buyer intent.
For time-sensitive signals, distinguish posted, first-seen, updated, reposted, and currently-open evidence. Do not use a search query's date wording or a stale listing as proof of urgency.
Workflow
1. Define inclusion and exclusion rules that trace to the offer and capacity. 2. For every prospect, separate: fit, evidence quality, observed urgency, reachability, and disqualifiers. 3. Score or tier each dimension using only the supplied evidence. Do not reward document volume, guessed budget, or an unverified job posting. 4. Explain the recommendation: pursue now, research next, monitor, defer, or exclude. Include the precise evidence and the most important uncertainty. 5. If funnel data is supplied, report each stage with its denominator and as-of date. Do not turn a snapshot into a trend or claim messages were sent. 6. Return a short queue with the recommended next research or human action. 7. Use a priority tier only as a work-ordering label, never as a buying signal or forecast.
Output
Use a table containing prospect, fit rationale, evidence status, priority, unknowns, and next action. Include the scoring assumptions and exclusions. For a sourcing run also report raw, normalized, pre-filtered, qualified, and ranked counts plus source coverage. For every ranked or disqualified record, include the source URL, source access or observation date, qualification or disqualification status, and the evidence that supports that status. Keep proposed research, meeting preparation, draft-only outreach, and record-update handoffs separate from any action.
Boundaries
This method does not enrich a contact database, create records, send outreach, or represent a prioritization score as a buying signal or forecast. It does not bypass blocked sources or invent dates, headcount, authority, or contact data.