Scoring list quality before you send

Score a cold list before the first email leaves, because every list defect converts into a campaign defect at a worse exchange rate. A workable scorecard has four checks — verification rate, ICP fit, completeness of the fields your copy uses, and provenance — and a list that fails any one of them goes back for repair, not into the sender. Ten minutes of scoring is cheaper than three weeks of sending into a list that was never going to work.

Why score a list you just paid to build?

Because the sending system cannot distinguish a bad message from a bad list, and by the time reply data tells you something is wrong, you have already spent your two most limited resources: weeks of calendar time and your domain's reputation. Bounces above a few percent damage deliverability for every campaign that follows, not just the one that caused them. A weak list also corrupts your learning: we expect roughly 4% positive replies from a working campaign and treat below 3% as a signal to fix something — but if the list is bad, you will rewrite perfectly good copy chasing a problem that lives in the data.

Scoring is the checkpoint between building and sending. The full build process — ICP definition, sourcing, enrichment, verification — is laid out in the B2B Database Building Guide; scoring is the acceptance test at the end of it, and it applies whether you built the list yourself or bought it.

What are the four checks?

Verification is non-negotiable and mechanical: run every address through a verifier, send only to valid, and quarantine catch-alls and unknowns for a lower-volume, watched send. ICP fit is the one most firms skip because it requires human eyes — pull a random sample and check each record against the ICP as written, not as remembered. Completeness only matters for fields the sequence uses; a half-empty column your copy never references costs nothing. Provenance — where each record came from and when — takes seconds to record at build time and is painful to reconstruct later; it is also the backbone of the record-keeping described in prospect data and UK GDPR.

How does the scoring pass actually run?

The mechanism:

1. When the list arrives, run full verification first — there is no point fit-checking addresses that bounce. Remove invalids, quarantine unknowns.
2. When verification passes, draw a random sample of 20–30 records and score each against the written ICP: firm size, sector, geography, and the role you sell to. Two or more misses in the sample means the sourcing filters are wrong — fix the filters and rebuild, because the sample fails in proportion to the whole.
3. When fit passes, check completeness on merge fields only. Any field the sequence references must be populated and plausible on every record it will touch; a blank first name in one row becomes "Hi ," in one inbox.
4. When completeness passes, confirm provenance: source and date logged per batch, suppression list applied against the import.
5. When all four pass, the list is cleared to load — and the score gets written down, so next quarter's comparison has a baseline.

The whole pass takes well under an hour on a list of a thousand and fails fast: most bad lists fall at step one or two.

What does a fit-check actually look for?

Not just the hard filters. A record can match size, sector, and title and still be a poor prospect — the firm is dormant at Companies House, the "managing director" runs a one-person shell, the website shows a business winding down rather than growing. The sharper version of the fit check looks for evidence the firm is in a buying posture, which is a topic of its own: readiness signals in 5–50-staff service firms covers which observable signs actually correlate with a firm being worth contacting now rather than existing in general.

Isn't this over-engineering for a cold list?

It is less engineering than most firms apply to far cheaper decisions. The pattern I see repeatedly is asymmetric care: weeks spent polishing sequence copy, zero minutes spent auditing the list it will be sent to — then, when replies do arrive, the same asymmetry repeats and interested prospects get two follow-up touches when deals typically need five or more. The discipline is the same discipline at every stage: check the input before blaming the output. A scored list will not rescue weak copy, but good copy cannot save bad targeting — and scoring is the only point where you find that out for the price of a coffee rather than a quarter.

---

Next step: the Growth System Audit — £450, seven days, credited against any build — maps where your growth system leaks and what to build first.