# Write a discovery brief that finds the right companies

A brief that works names four things: who the organisation is, where it operates, what would make it a fit that a public source could actually prove, and what disqualifies it. The third is the one most briefs skip, and it is the one that decides whether research produces evidence or adjectives.

- **Status:** Available
- **Audience:** both
- **In the app:** #/prospects
- **Last verified:** 2026-09-10
- **Canonical:** https://connectbyjbrh.com/docs/how-to/write-a-discovery-brief/

## Why a vague brief produces confident rubbish

Discovery runs through collection, public-source research, evidence capture, scoring and a compliance check before anything is contacted. Every one of those stages needs something to test against. A brief that says 'growing manufacturers who need better customer service' gives the qualification stage nothing checkable, so it scores on impressions — and a claim without evidence behind it does not reach the record at all, which is how you end up with a long list and an empty pipeline.

## The four parts

**Who** — The kind of organisation, in the words a public source would use: the sector, the rough size, and where in it your buyer sits. 'Mid-size CNC machining shops' beats 'manufacturers'.
**Where** — The location scope, explicit. Research cost is routed to where it can change a decision, and an unbounded geography spends it everywhere.
**What makes it a fit** — A condition somebody could verify from a website, a filing, a listing or a public announcement. This is the test the qualification stage runs.
**What rules it out** — The disqualifiers. These are cheaper than they look: they stop deep research being spent on candidates that were never going to qualify.

> **Note** Write the fit condition as something you could point at on a page. If you cannot imagine the sentence on a public source that would prove it, the scoring stage cannot either.

## A brief rewritten

| Part | Weak | Rewritten |
|---|---|---|
| Who | Businesses that need our product | Independent equipment hire firms with their own workshop, roughly 10–80 staff |
| Where | Anywhere | Named cities and their surrounding districts, stated as a list |
| Fit | Growing and forward-thinking | Publishes a services page listing on-site repair, and has a contact route other than a form |
| Rules out | (not stated) | Franchise branches of a national chain; firms already in the workspace as customers |

The rewritten fit condition changes the whole run. It is checkable, so evidence attaches to it; it implies a research target, so cost routing knows what depth is worth buying; and it produces a reason for contact that a first message can be built on rather than a generic opener.

## Running it and reading the result

1. Enter the brief on Prospects and run a small discovery first.
   - Result: A small run tells you whether the fit condition is finding evidence before a large one spends research budget proving that it does not.
2. Open three candidates and read the evidence attached to each claim.
   - Result: If a claim has no evidence, it is not on the record — so a thin evidence list is a brief problem, not a data-source problem.
3. Check the contactability split.
   - Result: Connect does not pattern-generate an address, so a prospect with no discoverable contact route stays a researched prospect with a different next action. This deliberately costs volume, and the split tells you how much.
4. Check how many were set aside as existing relationships.
   - Result: Identity resolution stops an existing customer being prospected. A high number here means the brief is describing your current customers, which is useful information about the brief.

> **Careful** Headline counts on a prospecting workspace were once derived from a fixed slice of the newest rows, which reported a round number where the truth was larger and zero where it was not. Counts are computed over the whole set now — but if a figure looks suspiciously round, the counting page is worth reading before you draw a conclusion from it.

## Questions

### How specific is too specific?

When a small run returns almost nothing, the brief has over-constrained. Widen the location before you widen the fit condition — geography usually costs less precision than loosening what qualifies.

### Can I write the brief as a paragraph rather than as fields?

You can, but the four parts still have to be present and separable. A paragraph that never says what rules a company out leaves the disqualification stage with nothing, and every borderline candidate then gets the expensive research pass.

### Why do some qualified prospects still have no email address?

Because no address was discoverable and none is invented. That prospect is kept with a different next action — a call, or a named person to research further — rather than being contacted at a guessed address that would damage your sending reputation.

## Related

- [Describing who you want to reach](https://connectbyjbrh.com/docs/prospects/discovery-criteria/)
- [Evidence on a prospect](https://connectbyjbrh.com/docs/prospects/evidence/)
- [Connect does not guess email addresses](https://connectbyjbrh.com/docs/prospects/no-guessed-email/)
- [Qualifying a prospect](https://connectbyjbrh.com/docs/prospects/qualification/)
- [Location and scope](https://connectbyjbrh.com/docs/prospects/location-scope/)
- [Evidence-first prospecting without guessed emails](https://connectbyjbrh.com/research/evidence-first-prospecting/)

## What this page is based on

- Connect source pack: channels (docs-source/sources/CHANNELS.md §4 — prospecting)
- Connect capability registry (docs-source/facts.py — prospect_discovery, no_guessed_email)