Every B2B data provider leads with a database size. Two hundred million contacts, five hundred million, seven hundred million.
The number tells you almost nothing. What matters is how many records the provider can return for the people you specifically want, how accurate those records are, how recently they were checked, and what you pay for the ones that miss.
Here is how to compare them properly, and how to test before committing.
The four things that decide value
1. Match rate, in your segment
Match rate is the proportion of requested records the provider can actually return. It varies enormously by segment, and aggregate figures hide that.
A provider with excellent coverage of US software companies may return a third as much for German manufacturing or Indian services firms. Seniority matters too: VP-level contacts at mid-market companies are well covered almost everywhere; operational roles at small companies are not.
Ask for match rate on a sample of your actual target list, not the headline figure. Any provider that will not run a sample against your list is telling you something.
2. Accuracy, which is different from coverage
A returned record is not a correct record. Two failure modes:
Stale. The person left eighteen months ago. The email bounces or reaches a successor who has no idea who you are. This is why refresh cadence matters more than most buyers realise — a database refreshed every 30 days behaves very differently from one refreshed annually.
Wrong. Pattern-guessed addresses that were never verified. Some providers infer firstname.lastname@company.com and present it as found. It is a guess, and guesses bounce.
Ask whether emails are verified at the point of delivery or at the point of collection. Real-time verification before you spend a credit is the meaningful version.
3. The credit model
This is where cost diverges most between vendors and where buyers get caught.
Things to establish:
- Do you pay for records that return nothing? Some providers charge for a lookup regardless of result.
- Are different data types priced differently? A mobile number costs considerably more to source than an email and is usually priced accordingly.
- Do company records or filters consume credits, or only contacts?
- Do unused credits roll over, or expire monthly?
- Is verification charged separately from the reveal?
A provider charging $0.02 a record that bills for misses can easily cost more than one charging $0.04 that does not.
4. Regional depth
If you sell internationally, test each region separately. The pattern across the industry is broadly:
| Region | Typical coverage quality |
|---|---|
| US and Canada | Strongest across all providers |
| UK and Western Europe | Good, with GDPR-driven gaps in some countries |
| Nordics, DACH | Variable; German coverage is often weaker than expected |
| India, SEA | Improving quickly, still inconsistent |
| LATAM | Thin outside Brazil and Mexico |
| Middle East | Thin |
| Japan, Korea | Thin, and email conventions differ |
A provider that is excellent for a US-only motion may be close to useless for a European or APAC one. This is the most common mismatch in the category.
How to run a proper test
Do not evaluate on a demo. Run this instead, and run it on every provider you are considering with the same input.
Step 1. Build a list of 100 people you can verify independently. Existing customers, people you have met, contacts you already hold accurate data for. This is your ground truth.
Step 2. Strip it back to name and company. Remove the emails and phone numbers.
Step 3. Run it through each provider. Record the match rate: how many returned anything at all.
Step 4. Check accuracy against your known-good data. How many of the returned emails are correct? How many phone numbers?
Step 5. Test a segment you do not know. Take 100 records in your actual target market where you do not have ground truth, run them, and send a verification-only check. Bounce rate on that sample is your real accuracy proxy.
Step 6. Price it. Total credits consumed across both tests, including misses, divided by correct records returned. That is your true cost per usable contact, and it is frequently two to three times the advertised rate.
The whole exercise takes an afternoon and routinely changes which provider people choose.
The categories of provider
Large aggregators. Broad coverage, strong in North America, usually with sequencing attached. Convenient, but data quality varies by region and the sequencing is typically weaker than dedicated platforms.
Enterprise data vendors. Highest accuracy on large-company records, with firmographic and technographic depth. Priced accordingly, usually on annual contracts with seat minimums.
Waterfall and enrichment platforms. Query several sources in sequence to raise match rates. Excellent coverage, but you pay per source queried and the cost model takes work to predict.
Built into outreach platforms. Data included in the tool you already send from, usually credit-metered. Fewer records to move between systems, and no export step before a campaign.
Specialists. Phone data, specific regions, specific industries. Worth layering on top of a general provider for the segments where the general one is thin.
Most teams past a certain size end up with two: a primary provider covering the bulk, and a specialist filling a gap.
Questions worth asking a vendor
- What is your match rate on a sample of my list? (Not aggregate.)
- How often is the database refreshed?
- Are emails verified in real time, or at collection?
- Do I pay for lookups that return nothing?
- How is phone data priced relative to email?
- What is your coverage like in [the regions you actually sell to]?
- Where does the data come from, and what is your compliance position under GDPR?
That last question matters more each year. Ask how contacts are sourced and what lawful basis the provider relies on, and get an answer you would be comfortable repeating to a prospect who asks where you got their details.
The short version
Database size is marketing. Match rate in your segment, accuracy against ground truth, the credit model’s treatment of misses, and regional depth are what you are actually buying.
Test with 100 records you can verify, then 100 you cannot, and calculate cost per correct record including the misses. That number is usually two to three times the advertised rate and it is the only figure worth comparing.
And if you sell across several regions, test each one separately, because providers are far less consistent geographically than their homepages suggest.
Next: Apollo alternatives, or how to find someone’s email address.
Frequently asked questions
What is a good match rate for a B2B data provider?
It depends on the segment. For US mid-market and above, 70% or better is reasonable. For smaller companies, operational roles or regions outside North America and Western Europe, expect considerably less. Always test on your own list rather than trusting an aggregate figure.
How do I test a B2B data provider?
Run 100 records where you already know the correct data, measure match rate and accuracy, then run 100 in your real target segment and use bounce rate as an accuracy proxy. Divide total credits spent by correct records returned.
Why do B2B email databases have bad data?
People change jobs constantly and some providers infer addresses from patterns rather than verifying them. Refresh cadence and whether verification happens in real time are the two things that separate good data from stale or guessed data.
Should I buy data separately or use what is in my outreach tool?
Included data avoids exports and keeps one credit balance, which suits most teams. A separate specialist provider makes sense where your primary source is thin — a specific region, or phone numbers.