Qualifying a prospect, end to end

The chain, stage by stage#

1. Trigger — an organisation has been discovered and researched, and there is now enough on the record for a judgement to be worth making.
2. External event — none. Nothing about qualification depends on the prospect doing anything; they do not know they are being considered.
3. Authentication and workspace resolution — criteria, scores and the prospect records themselves all belong to one workspace, and every read below happens inside it.
4. Request — the discovery criteria are the standard being applied. A score is only meaningful against a written statement of what a good customer looks like for this business.
5. Canonical record — claims about the organisation, each with the source it came from. A claim with no evidence attached does not reach the record at all, so it cannot quietly influence a score.
6. Reasoning — scoring compares the evidenced claims against the criteria and produces a fit judgement with the reasons legible beside it.
7. Knowledge, memory and rules — identity resolution runs here: an organisation that is already a customer, or already a relationship, is not prospected. Freshness decides whether the research behind the score is still worth trusting.
8. Autonomy and approval — qualification decides nothing about sending. It ranks work; the readiness checks and the channel's autonomy decide separately whether anything leaves.
9. Action — no provider is called and nothing is sent. This whole flow is inward, which is why it is safe to run at volume.
10. Result — the prospect carries a decision and a reason: worth approaching, not worth approaching, or not yet decided because the evidence is thin.
11. Relationship, timeline and memory — a prospect is not yet a relationship. It becomes one when a person is engaged with, which is a later and deliberate step.
12. Audit, usage and Needs You — research spend is metered, the decision is recorded with its reasons, and anything a person should look at is ranked into Needs You rather than emailed to them.

Evidence first, or the score means nothing#

Every claim on a prospect record carries the source it came from, and a claim without evidence does not reach the record. That rule sounds bureaucratic until you have watched a scoring model produce a confident number from three plausible-sounding assertions that nobody can trace.

The practical consequence is that a score is always openable. "Why is this organisation on my list" has an answer made of specific claims with specific sources, which is what makes disagreeing with it productive: you are arguing with a stated reason rather than with a number.

It also constrains what a thin record can produce. An organisation about which little is publicly evidenced does not become highly qualified because the gaps were filled optimistically; it becomes a prospect whose evidence is thin, which is a different and more honest state.

Three outcomes, not two#

The third outcome is the one that saves money. Research spend is routed so that a deeper pass runs where it can change a decision and not where it would merely add detail to one already made. An organisation that is obviously outside your criteria does not earn a second look, and one that is obviously inside them does not need one either.

Overriding it#

A person can disagree with any qualification, and should. The score is a ranking aid built from public evidence; you may know that an organisation is about to open a second site, or that its owner is a former colleague, and no amount of public research will find either.

1. Open the prospect and read the reasons rather than the score.

   Result You are checking whether the evidence is wrong or whether the criteria are — two different corrections.

2. If the evidence is wrong, say so on the record.

   Result The correction sits with the claim it contradicts, so the next person to open the record sees both.

3. If the criteria are wrong, change the criteria rather than the prospect.

   Result One edit fixes every future run. Overriding prospects one at a time is a treadmill that leaves the underlying standard as wrong as it was.

4. Promote it and continue.

   Result The override is recorded with its author, so a list nobody can explain six weeks later is not one of the ways this goes wrong.

What can go wrong#

The headline numbers disagree with the list

Summary figures were once derived from the newest five hundred rows, so a workspace with more of either was told it had exactly five hundred. On a workspace of 540 contacts whose 40 already-contacted ones were the oldest, that read as 500 organisations — really 530 — and none already contacted, really 40. Contacts are counted in batches now and organisations are counted by the database.

Everything scores about the same

The criteria are too general to discriminate. Criteria that would admit most businesses in a city produce a ranking that is really just a discovery order.

A current customer appears as a prospect

Identity resolution is what prevents this, and it reads the records it can see. A duplicate record, or a lifecycle stage never moved after a win, is what gets past it.

The score is built on stale research

Freshness governs re-research. A judgement made from a two-year-old page is not wrong so much as out of date, and it is worth re-running before spending outreach on it.

Questions#

Related#

• Qualifying a prospect
• Evidence on a prospect
• Why this prospect
• Describing who you want to reach
• Contactability
• Running deep research, end to end
• Resolving a prospect to an existing relationship
• Evidence-first prospecting without guessed emails