B2B DATA PROVIDER BUYER GUIDE
The largest database is not automatically the best source for your market. Compare coverage, evidence, freshness and the work your team must still perform.
Specify target countries, company types, buyer roles, required fields, acceptable uncertainty and delivery format. Include exclusions and CRM constraints. A provider cannot be judged fairly against an undefined output.
Static or searchable databases can work when their coverage matches your segment and your team can qualify and maintain the records. Ask how records are collected, refreshed, corrected and removed, rather than treating a field as permanently current.
A research service starts from your brief and investigates the specific companies and buyers that fit it. That can be a better match for narrow sectors, unusual buying roles or requirements where source context matters more than a large contact count.
Before buying, request a sample in your target segment and ask how duplicates, invalid emails, changed roles, suppression requests and unsupported matches are handled. Confirm privacy, compliance and outreach responsibilities with your own advisers.
| Dimension | Evidence to request | Decision |
|---|---|---|
| Geography and roles | A sample in your countries, company types and buyer roles. | Count usable matches; global database size does not answer this. |
| Freshness | Source dates and refresh method for roles and contact details. | Identify what needs rechecking before use. |
| Verification | Definitions for valid, invalid, accept-all and unknown outcomes. | Set an explicit acceptance policy for uncertainty. |
| Provenance | Where each field came from and what can be retained/exported. | Hold claims that cannot be traced. |
| Use and correction | Licence, data-handling terms, suppression and correction process. | Resolve your use case and responsibilities before activation. |
| CRM/API fit | A representative import with IDs, evidence and status fields. | Verify deduplication and field preservation. |
| Credits and support | Written quote, failed-lookup rules and support commitments. | Compare cost per accepted record including review time. |
Some sources supply a searchable catalogue, some find or verify contact details from known companies, and some help identify account activity. Those inputs can complement custom research but do not perform the same job.
Examples checked on 18 September 2026: Cognism describes phone-verified contact data; Hunter offers domain-based contact discovery; and Leadfeeder’s workflow example starts with identifying visiting companies and enriching them. These are vendor descriptions, not independent tests or evidence that a particular buyer has enquired.
Use the same scorecard if you shortlist Apollo, ZoomInfo, Lusha, UpLead, Lead411, RocketReach or LeadIQ. Their inclusion here is a set of research candidates, not a claim of verified coverage, compliance or current pricing. Ask each for current documentation and a sample that matches your brief.
Illustrative evaluation; no provider result is implied. Suppose you need UK contractors that maintain industrial roofs and an operations-level buyer. Give each provider the same exclusions and field requirements. Ask reviewers to inspect company service evidence first, then buyer relevance and contact status. Keep domestic-only firms, uncertain service coverage and missing buyers as separate rejection/hold categories.
Imagine one 50-record sample produces 20 accepted records, 15 domestic-only firms, ten missing buyer identities and five uncertain contacts. The accepted yield is 40%. The result is useful because it identifies what further work is needed; describing it as 50 qualified leads would hide that work. Do not generalise this fictional calculation into a rating for any vendor.
Calculate data charges plus research, verification, integration and review costs, divided by accepted records. Add refresh and replacement costs when comparing ongoing use. Preserve the date, brief, sample size and reviewer criteria alongside the calculation so another operator can reproduce the decision.
Custom research can investigate requirements that a catalogue does not encode. That benefit still needs proof in your own brief: inspect an included record, a rejected record and a partial-yield explanation. See custom researched B2B leads and the tools comparison for the operating choices.
Ask which fields, market, sample, dates and method it covers. Email contactability and company relevance are different measures; neither establishes buyer interest.
Only promise a field where the evidence and permitted source support it. Agree how missing or uncertain contact routes affect acceptance and pricing.
Confirm the correction and suppression flow across research, exports and CRM, including future re-imports. A refreshed database row should not silently override a recorded opt-out.
Use the job-based tool comparison to frame the shortlist, then review custom researched B2B leads as a separate operating approach.
APPLY KEYSTONE TO YOUR MARKET