Why Enrichment Is Hard to Evaluate

Enrichment demos are self-selecting. The vendor picks the account, the account is a household name, and the record comes back gleaming. Nothing about that predicts what happens on row 1,400 of your actual list: the regional distributor with a thin web presence and a name shared by three other companies.

The questions below are designed to be asked across a demo table. They split into four groups: coverage, depth, freshness, and fit. For each one, the interesting part is not the answer itself but whether the vendor answers with a mechanism or an adjective.

Coverage: Will It Work on YOUR List?

1. What is your match rate on a list like ours, and how do you measure it?

Ask for the definition before the number. A "match" that means "we found the domain" is a different claim from "we produced a usable record". Get the number for your segment, not their global average.

2. What happens to the accounts you cannot match?

The mediocre answer is silence: unmatched rows quietly return empty. The good answer names a mechanism: ambiguous companies get flagged, name collisions get resolved against the domain you supplied, and the output tells you which rows to treat with caution.

3. Run ten rows of our list, live, right now.

Not their sample. Yours, including the awkward ones. Any vendor confident in their pipeline will do this in the meeting. Hesitation is data.

Depth: Columns or Context?

4. Show me the full record for one mid-sized, unglamorous account.

Household names are easy; the median account is the product. Look at whether the record reads like research (what the company does, what is changing, why it matters) or like a form (codes, bands, and booleans).

5. What in this record would change what a rep says in the first thirty seconds?

This is the sales-ready test. If the vendor cannot point to the field that shapes the opener, the enrichment adds columns, not conversations.

6. Where does the research come from when the data is not in a database?

Firmographics live in databases. Context does not. A vendor producing genuine context can explain how they research an account, not just which datasets they licence.

7. Can the output drive personalisation directly?

The strongest test of depth: can the record feed a personalised page or asset without a human rewriting it? If enrichment and personalisation cannot connect, you are buying a better spreadsheet.

Freshness: When Was This True?

8. When was each field last verified, and can I see that per field?

A record is a snapshot. A good vendor timestamps the snapshot. A great one can tell you which fields decay fast (contacts, roles) and which hold (firmographics), and treats them differently.

9. What does a re-run cost?

Lists go stale. If refreshing a campaign costs the same as enriching it the first time, the pricing quietly assumes you will work with stale data. Ask how repeat runs on the same accounts are priced.

Fit: Will It Work in YOUR Operation?

10. How does the output arrive, and in whose schema?

If the answer is "a CSV export", every downstream step is your integration problem. Ask what it takes for the enriched record to land where your team actually works.

11. What do you charge for: rows processed, or usable records delivered?

Per-row pricing means you pay for the misses. Per-usable-lead pricing (CPL) means the vendor's incentive is your usable output. The pricing model tells you what the vendor optimises internally.

12. What does a bad month look like, and what do you do about it?

Every pipeline has bad months: a source degrades, a segment underperforms. A vendor who claims otherwise is hiding the ball. A vendor who describes their detection and remediation process has been through it and built for it.

How to Run This

Do not send these as a questionnaire; vendors have teams for questionnaires. Ask them live, in order, against your own list. Take the answers that named mechanisms and weight them. Discount the answers that were adjectives.

And start with the list itself: the quality of an evaluation is capped by the quality of the target account list you test with. Our campaign TAL template is the exact seven-column format we ingest, with notes on which columns matter most for enrichment depth. Fill it with fifty real accounts, including your hard ones, and take it to every demo. The comparison will make itself.