Where European B2B data actually comes from

Every outbound campaign rests on data someone had to source. Knowing where it came from is not bureaucracy, it is what makes the campaign defensible.

Public business registries

The base layer across most European markets: company existence, registration details, sector and often size. Highly reliable for firmographics, useless for finding the right person. Good for building the universe of companies, not the contact list.

What the registries look like market by market

Every European market runs its own registry with its own quirks. Germany has the Handelsregister and the Bundesanzeiger for filings. Poland runs KRS for companies and CEIDG for sole traders, which is where most Polish B2B lead generation list building starts. Lithuania's business register publishes company data cheaply and in a usable format. France, the Nordics and Benelux each have their own equivalents, with varying openness and cost.

Two practical consequences follow. First, cross-border list building means stitching together several registries with different field names, languages and identifiers, which is why pan-European B2B data projects take longer than single-market ones. Second, registry data tells you a company exists and roughly what it does. It rarely tells you the company is worth writing to this quarter. That timing signal comes from elsewhere: hiring pages, funding announcements, tenders, expansion news.

Company websites

Published contacts, team pages, role descriptions. Accurate when current, and genuinely public, but coverage varies enormously by company size. Small companies often publish everything, large ones publish almost nothing.

LinkedIn

The best source for matching a real person to a real role, which registries cannot do. Roles are current and public. It does not give you an email address, which is where the next step comes in.

Specialist databases

Useful for filtering at scale and for firmographic enrichment across several markets at once.

Quality varies widely, and the important question is where the provider sourced it. A database that cannot explain its own provenance is a liability you are renting.

Inference, and why it needs checking

Many tools infer addresses from a name and a domain pattern rather than confirming them. That is a legitimate technique, but the output is a guess, which is exactly why verification is not optional.

The compliance question

In Europe you should be able to say where a contact came from and why you have a legitimate reason to write to them about their professional role. Public, business-related sources make that straightforward, though the bar is higher in some markets than others, as the guide to cold email law in Germany sets out.

Large bought lists of unclear origin make it impossible, which is the practical reason to avoid them, quite apart from their poor accuracy. See the GDPR guide.

Combining sources beats any single one

In practice the workable approach is layered: registries for the company universe, LinkedIn for the person, websites and tools for the address, verification before sending. No single source does all four well.

A worked example: building a 500-contact DACH list

Say you sell warehouse automation and want operations directors at mid-sized German logistics firms. The layered approach looks like this:

1. Company universe. Pull logistics companies with 50-500 employees from registry and database filters. Expect the raw candidate count to run several times larger than the list you want before any manual review.
2. Qualification. Cut companies that do not fit: freight brokers with no warehouse, firms in insolvency, subsidiaries of accounts you already serve. A third of the raw pull often goes here.
3. Person matching. Find the operations or logistics lead on LinkedIn for each surviving company. At smaller firms this is often the managing director.
4. Address sourcing. Combine published addresses, pattern inference and enrichment tools, then verify every address before it enters a sequence.
5. Documentation. Record the source for each contact. One column in the sheet, thirty seconds per record, and your compliance position is written down instead of assumed.

Done properly this takes days, not hours. That is the honest cost of a list that lands, and it is why quality beats quantity in every campaign worth running.

Common B2B data mistakes, and the fixes

• Buying one big list and sending to all of it. A high bounce rate damages sender reputation fast, and the damage starts well before the number looks alarming. Verify first, send to the clean segment only.
• Treating data as a one-off purchase. Contact data decays as people change jobs; a list that was accurate in January is noticeably stale by summer. Budget for refresh, and read up on list decay and maintenance.
• Filtering only on firmographics. Industry and headcount produce a broad list. Adding a timing signal, hiring, expansion or new funding, produces a list of companies with a reason to reply.
• Ignoring the suppression side. Existing customers, open opportunities and past opt-outs must be removed before every send, not after the first complaint.

Questions to ask any data provider

Before paying for access, ask these and listen for hesitation:

• Where does the data come from? A serious provider names its sources. "Proprietary methods" with no detail is a red flag.
• When was this record last verified? Coverage numbers mean little if half the records are two years old.
• What is the guaranteed bounce ceiling? Good providers stand behind a figure and credit bounces above it.
• Can you handle deletion requests? Under GDPR you may need to prove a contact was removed. The provider should support that, not shrug.

What accuracy you can expect

The honest version is comparative rather than numeric, because the exact percentages depend on the market and on how old the data is. A carefully built and verified list bounces at a rate you barely notice. A purchased database that you verify again on your side bounces measurably more. An unverified bought list can bounce badly enough to burn a warmed domain in a week. Measure your own bounce rate on the first send and use that as the baseline instead of a borrowed benchmark. Registry firmographics are reliable; role data is only as fresh as its last check; inferred emails are guesses until verified. Build your process around those realities and the numbers hold. Then feed the clean data into a properly segmented list so the copy can match the audience.