A B2B contact database provider sells searchable access to a store of companies and people with work contact details, firmographics and sometimes intent signals, metered by seat and export credits. Providers differ most on regional coverage, how they verify, and how often records refresh; across the leading providers, 10 to 22% of exported emails are invalid at export time, whatever the claimed accuracy, so the choice is about fit for your segment and the verification step after export.
Record counts and claimed accuracy are vendor-reported. Invalid at export is measured only where BounceZero has an export study (Apollo 5.8% invalid plus 18.4% catch-all, ZoomInfo about 22%); the other rows are estimates from the finder accuracy range of 10 to 25% invalid, not measurements. Prices are list prices at the time of writing and change often; treat them as order-of-magnitude.
| Provider | Strongest region | Records (vendor-reported) | Claimed accuracy | Invalid at export (measured or estimated) | Entry price (vendor-reported) |
|---|---|---|---|---|---|
| ZoomInfo | US enterprise | 260M+ professionals | 95% | 12 to 20% | Custom, five figures a year |
| Apollo | US and global SMB | 275M+ contacts | 91% | 5.8% invalid + 18.4% catch-all | Free tier; from $49 a seat a month |
| Cognism | UK, EU, DACH | 400M+ profiles | 98% on verified phones | 10 to 18% | Custom, from about $1,500 a year |
| Lusha | US, Western Europe SMB | 100M+ contacts | 81% | 15 to 22% | Free tier; from $36 a seat a month |
| Kaspr | France, DACH via LinkedIn | 500M+ profiles | Not published | 12 to 20% | Free tier; from about $49 a seat a month |
| RocketReach | US, academic, government | 700M+ profiles | 90 to 98% | 15 to 25% | From $39 a month |
| Clearbit (HubSpot Breeze) | US tech, inbound enrichment | Not published | Not published | 12 to 20% | Bundled with HubSpot tiers |
Reviews with dated sources: ZoomInfo, Apollo, Cognism, Lusha. Export studies: Apollo, ZoomInfo.
ZoomInfo and Apollo grow their stores partly from customers' own mailboxes and CRMs, exchanged for credits. Broad and fast-growing, but a contributed address is as old as the mailbox it came from. The "verified" badge records when the provider last saw it, not whether it works today.
Cognism runs human-checked phone data and compliance screening against do-not-call lists. Phones are the strength; email accuracy sits below phone accuracy because mailboxes are checked less often than numbers are dialled.
Kaspr, RocketReach and Lusha lean on LinkedIn profiles plus the company's address pattern. Works for people never seen before; fails on catch-all domains, which accept any guess, and on firms with mixed formats.
Providers measure accuracy at indexing time against their own checks. Your export happens later, and B2B addresses decay fast, with only 19.2% of addresses still valid 90 days after verification in the BounceZero decay study as people move. A record verified eight months ago has a one in seven chance of being dead on the day you download it, with the badge still showing.
The second gap is catch-all domains. A provider's SMTP check sees the server accept the address and marks it good; the server would have accepted anything. In the BounceZero corpus, 2.3% of non-freemail (business and ISP) addresses were catch-all overall and 12.4% on .org domains, and those addresses count as "verified" in most provider interfaces.
The third gap is unknowns. Business gateways refuse probes 27% of the time in our 2026 data. A provider that shows no unknown state has folded those records into valid. A verifier that labels them separately lets you route them to phone or LinkedIn instead of burning a domain on them.
Region and company size decide more than feature lists. The corpus rates by country domain show why: the same provider exports very different lists for different markets.
| Your segment | First choice | Second choice | Corpus rate for the region |
|---|---|---|---|
| US enterprise, 1,000+ staff | ZoomInfo | Cognism | .com 16.4% invalid, 14.7% unknown |
| US and global SMB, price-sensitive | Apollo | Lusha | .com 16.4% invalid; freemail-heavy lists 17.6% |
| UK and Ireland | Cognism | Apollo | .uk 17.0% invalid, 10.6% unknown |
| Germany, Austria, Switzerland | Cognism | Kaspr | .de 41.2% unknown (gateway refusals), 13.9% invalid |
| France and Benelux | Kaspr | Cognism | .fr 12.1% invalid; .nl 21.9% invalid |
| Australia and New Zealand | Apollo with heavy verification | ZoomInfo | .au 36.5% invalid, 6.3% catch-all |
| Academic, public sector, non-profit | RocketReach | Apollo | .edu 32.8% invalid; .org 12.4% catch-all |
Rates: BounceZero corpus, April to October 2026, in the lead list quality benchmarks. High unknown rates in DACH are gateway behaviour, not provider error; plan phone-first there.
Every provider offers a trial or a free tier. Use it to measure export accuracy for your segment, not to admire the filters.
Same titles, company sizes and countries you will buy for. A US enterprise sample tells you nothing about a UK mid-market purchase.
Run it through bulk verification. The free tier covers 100 checks a month; a Starter pack covers 2,000 for $6. Note invalid, catch-all and unknown separately.
Better than the regional invalid rate is a good provider for that segment. Worse means the provider is weak there, however strong elsewhere.
Same segment, same day. Two samples make a decision; one sample makes an anecdote.
Divide the annual cost by the confirmed-valid records you will export, not by the records the provider claims. A cheaper provider with 22% invalid can cost more per reachable contact than a dearer one at 12%.
For UK and EU contacts you need to show legitimate interest, name the source in the first email and honour opt-outs. Cognism supplies do-not-call screening; most others leave the assessment to you. Build it into your privacy notice before the first send.
Credits that expire monthly push you to export more than you verify. Prefer annual or rolling credits, and budget verification credits alongside them: BounceZero credits never expire.
Ask how often a record is re-checked and what triggers a re-check. "Continuously" is not a cadence. Providers that publish a refresh interval tend to export cleaner lists.
Most contracts prohibit keeping exported data after cancellation. Your own verified records with consent history are yours; a provider's records are not. Keep the system of record in your CRM.
Anatomy, decay and the maintenance loop for your own base
Evaluating a one-off file instead of a subscription
The two largest providers head to head
US breadth against EU compliance tooling
How databases are built and where they decay
All sources ranked by data quality
None of the leading providers export better than about 10% invalid in our customers' 2026 data, and most run 10 to 25%. Accuracy depends more on your region and company size than on the provider, which is why a 200-row sample in your own segment beats any published claim.
For US enterprise coverage, org charts and intent data, ZoomInfo is deeper and costs several times more. For SMB, global breadth and built-in sequencing, Apollo covers most needs at a seat price. Both export 10 to 25% invalid addresses, so verification after export is the same step either way.
Cognism for UK, EU and DACH coverage with compliance screening, Kaspr for France and German-speaking markets through LinkedIn workflows. Expect high unknown rates in Germany because the dominant gateway refuses verification probes; plan phone or LinkedIn for those records.
Yes. The provider verified at indexing time; you export months later, after roughly 22% annual decay, and catch-all domains pass the provider's check while accepting anything. Verify every export before import, then before every send and at least monthly.
Divide each annual cost by the confirmed-valid contacts you will actually export, measured on a verified sample from your segment. Credits, seats and record counts hide the real unit; reachable contacts are what you pay for.
Export a trial sample from each provider, verify both the same day, and buy from the one that beats the corpus rate for your region. 100 free checks a month, unknowns refunded.
Ayoub built BounceZero's 5-stage validation pipeline, its dedicated BGP-announced IP infrastructure, and the Patroni HA PostgreSQL cluster behind every verification. Previously built high-volume email delivery infrastructure. Trained at 1337 Benguerir (École 42 network, 2019). Open-source: bgp_analyzer.
Where lists come from, what bounces, and what to check before you pay
Buying an email list for lead generation in 2026
A B2B leads database is only as good as the share of its contacts you can still reach
Email list websites for lead generation ranked for 2026
Sellers compared, with sample-check rules
How databases are built and where they decay
Thinking of buying a B2B email marketing list? Corpus data from 7.93M verified addresses shows what...
Continue through related topics