B2B database definition

Most pages about B2B databases rank vendors by size. Size says little about whether the rows are right for your market this month. Below: how a database is built from its sources, why every record starts ageing the day it is collected, what else you can pull from it besides names, and where EU law limits what you do with the result.

What is a B2B database?

A B2B database holds two linked kinds of record: companies and people. A company record carries firmographic data such as industry, size, location and website. A person record carries a name, a role, the company they work at and, when available, a work email and phone number. You search it with filters and export the matches as a list.

Review sites file these products under sales intelligence, next to tools that add buying signals and company research on top of the raw records. The database is the base layer. Everything else, from enrichment to scoring, works on the rows it returns.

Where does the data in a B2B database come from?

From many sources merged into one row per company and one row per person. No single source is complete, so the work is in combining them.

1. Collect. Public business registers, job portals and company career pages, company websites, news, and licensed data partners each supply part of the picture.
2. Aggregate. Records from all sources are pooled. Leadspicker combines over 500 sources this way.
3. Resolve. Entity resolution decides which records describe the same company or person. Acme GmbH in a register, acme.de on the web and Acme in a job ad become one row. A stable identifier such as the D-U-N-S Number exists precisely because this matching is hard.
4. Refresh. The whole process runs again on a schedule, and changed records are updated.
5. Pull. You filter the result and export a list.

A publicly listed data vendor's annual report to the SEC describes the same engine at large scale: many sources, machine-learning models that score each fact, and human review for the records that matter most. The mechanism is the same at every vendor. The sources and their regional depth differ.

A listed data vendor's Form 10-K for fiscal 2024 is a public description of how a B2B data engine works: aggregation from many sources, machine-learning scoring, and human review. Dun & Bradstreet's D-U-N-S Number shows why entity resolution needs a stable company identifier. G2's sales intelligence category defines the product class. The EDPB Guidelines 1/2024 set the legitimate-interest tests for using the data in the EU.

How current is the data, and why does it decay?

Every record starts ageing the day it is collected. People change jobs, get promoted and leave. Companies hire, shrink, merge, rebrand and change domains. A work email that was valid at collection bounces once its owner leaves.

Two things limit the damage. The first is refresh frequency: the Leadspicker lead database is refreshed every month, so a record is at most a few weeks behind its last source. The second is checking at the moment of use. Verify emails right before a sequence starts, not when the list was exported weeks earlier.

Coverage decays unevenly too, because sources differ by country. Where public registers are rich, company records can be checked against an official source. Where they are thin, the database relies more on websites and partners.

Per the Leadspicker knowledge base, August 2026, pending clearance.

What can you pull from a database besides a list?

Attributes that decide whether a company is worth contacting. A list of names answers who. Enrichment answers whether they are relevant.

• Company attributes: company size, department sizes, technologies in use, and what the company says about itself on its website.
• Hiring data: open job positions and positions advertised in the past, a direct signal of growth and of which functions are being built.
• Lookalikes: companies most similar to a few existing customers, ranked from the whole database.
• Contact details: verified work emails and direct phone numbers, usually bought per record from specialised providers.
• Change over time: management changes and people joining or leaving, used as trigger events.

The point of the extra columns is a narrower list. Leadspicker's knowledge base illustrates this as a team processing the 60 most relevant companies instead of 500; the figures are an illustration, not a measurement.

What are the legal limits on using it for outreach?

In the EU, B2B cold outreach to individuals usually rests on legitimate interest under the GDPR, and that basis has conditions. The EDPB Guidelines 1/2024 on legitimate interest set three cumulative tests: the interest must be legitimate, the processing must be necessary for it, and it must not be overridden by the person's rights and interests. Direct marketing can be a legitimate interest, but it is not one automatically.

In practice that means outreach that is proportionate and relevant to the recipient's professional role. A narrow, well-matched list is easier to defend than a broad one. Keep the source of each record, honour objections at once, and check national rules on electronic marketing, which can be stricter than the GDPR.

Frequently asked questions

Where does the data in a B2B database come from?

From many sources combined: public business registers, job portals, company websites, news and licensed data partners. Entity resolution merges their records into one row per company and per person.

How often is a B2B database updated?

It depends on the vendor, so ask how often the data is refreshed and when each record was last checked. Records still age between refreshes, so verify emails right before a sequence starts.

What rules apply to using a B2B database for cold outreach in the EU?

The EDPB Guidelines 1/2024 set three cumulative tests for processing under legitimate interest: a legitimate interest, necessity, and a balance with the person's rights. National e-marketing rules apply on top of the GDPR and differ by country. Keep outreach relevant to the role, record the source of each record and honour objections.