B2B Company Data in DACH: Which Sources Actually Work (2026)

Every outbound project in the German-speaking market runs into the same wall in week one: where does usable company data actually come from? The answer sold by most vendors is a single subscription that supposedly covers everything. In practice, the data that makes outbound work in DACH is assembled from four different layers, three of which most teams never touch. Here is the map, with what each layer is genuinely good for.

Layer 1: the official registers, and why DACH is unusually lucky

Most countries do not publish much about their companies. Germany, Austria and Switzerland do, by law, and this is the most underused advantage in the market.

HHandelsregister and the UUnternehmensregister give you the legally correct company name, legal form, registered seat, managing directors and authorised signatories, plus the history of changes. A change of managing director is a first-class buying signal, and it is public.

The BBundesanzeiger publishes annual financial statements. For anyone selling into the German mid-market this is the closest thing to a free firmographic gold mine: balance sheet totals, often employee numbers, and the ability to see whether a company is growing or shrinking. Filings arrive with a delay of several months and small companies file abbreviated accounts, so treat it as a directional signal rather than a live feed.

In Austria, the Firmenbuch covers the same ground; in Switzerland, ZZefix provides the commercial register index across all cantons, which is particularly useful because Swiss company data is otherwise fragmented by canton.

What these sources do not give you: contact people below board level, email addresses, phone numbers, or anything about technology and current projects. They are the skeleton, not the body.

Layer 2: aggregators, or paying someone to do layer 1 properly

Reading registers by hand does not scale beyond a few dozen accounts. Aggregators take the public filings, clean them, connect corporate structures and make them searchable.

NNorth Data is the pragmatic choice for register and filing data, with useful corporate network views and reasonable pricing. CCreditreform is the incumbent for credit and firmographic data in Germany and carries genuine depth on small companies that international providers simply do not have, at incumbent pricing. IImplisense focuses on German company data enriched with signals. International players like Dun and Bradstreet cover DACH but are usually overkill unless you already use them for credit checks.

The honest rule: pay for an aggregator when you need structure at scale, which means filtering thousands of companies by size, region, industry code or financial trajectory. For a target list of 300 companies, an analyst with register access and a few hours does better work.

Layer 3: contact data, where the disappointment usually happens

This is the layer everyone starts with and where DACH-specific reality hits hardest. Most contact databases were built for the US market, where professional profiles are richer, job titles are more standardised and privacy rules are looser.

Two practical notes that save money. First, LLinkedIn Sales Navigator remains the most accurate source for who currently holds which role in DACH, because the people maintain it themselves. It is not an export tool, but as the truth layer for roles and changes it is unmatched. Second, mobile phone numbers of employees are personal data in the strict sense, and the legal exposure of calling a mobile number obtained from a database is different from calling a company switchboard. Treat the two as separate categories.

The 50-row test. Before signing any contract, take 50 companies that genuinely represent your target segment, German mid-market, the right size band, the right industry, and have the provider enrich them. Then check three things: how many rows came back at all, how many titles are actually current, and what the bounce rate is on a small send. Every provider looks excellent in a demo built on well-known technology companies. The test that matters is a Maschinenbau company with 80 employees in Baden-Wuerttemberg.

Layer 4: signals, the only layer that changes results

Address data tells you where to send. Signal data tells you when, and that is the variable that actually moves reply rates. This layer is assembled rather than bought.

Job postings are the strongest and cheapest signal in the German market. A company hiring for a role in your area has a budget, a problem and an admission that the current setup does not suffice. Sources range from the company career pages themselves to SStepStone and aggregated job data providers such as TTheirStack or PPredictLeads. The refinement that most teams miss: a posting that has been open for eight weeks says far more than a fresh one.

Filing and register changes from layer 1, watched over time: new managing director, capital increase, new subsidiary, address change. Leadership changes on professional networks, tracked with tools such as UUserGems or TTrigify. Technology changes visible through BBuiltWith or WWappalyzer. Review activity on OOMR Reviews, CCapterra or GG2, where a company researching a category is doing so publicly. And employer reviews on Kkununu, which in the German market often describe operational problems in remarkable detail, from tooling to processes.

None of these require an expensive subscription to start. They require someone deciding which three signals matter for your offer and then watching them consistently.

The GDPR part, without the scaremongering

B2B outreach in DACH is legal when it is done properly, and the data side has three rules that matter.

Legitimate interest is a real basis, but it is not a blank cheque. Processing business contact data for direct approach can rest on Article 6 (1) (f), provided the contact is addressed in their professional role, the topic is plausibly relevant to that role, and the interest is documented. A list of private addresses fails all three tests.

The information duty applies. When you obtain personal data from a third party rather than from the person, Article 14 requires you to inform them, in practice within your first message: who you are, where the data came from, and how to object. One clean sentence in the footer handles this, and it also improves reply rates because it looks professional rather than shady.

Objection has to work immediately and permanently. A suppression list that survives system changes is not optional, and it is the single thing regulators check first.

Two things to simply avoid: consumer data, and bought lists whose origin nobody can explain. Beyond the legal exposure, purchased lists poison your sending domains, which is a slower and more expensive problem. We wrote about that in detail in the guide on buying B2B leads, and the country-by-country rules for cold outreach are in the legal guide.

What a working DACH data stack actually looks like

For a mid-market outbound build, the sensible combination is layered rather than monolithic:

1. Define the segment with register logic: legal form, size band, region, industry code, from an aggregator or the registers themselves.
2. Verify the segment against reality: 100 companies checked by a human, because industry codes lie constantly in the German mid-market.
3. Add the people: one contact provider as the base, professional networks as the truth layer for roles.
4. Add the timing: two or three signal sources that fit your offer, watched weekly.
5. Keep it clean: every bounce, every objection and every "wrong person" reply feeds back into the list, not into a folder.

The mistake almost everyone makes is spending the entire budget on layer 3 and skipping layer 4. A perfectly enriched list without timing produces a perfectly personalised message that arrives at a moment nobody cares. The signals guide covers the timing side in depth, and if you would rather have this assembled and operated than build it internally, that is exactly what we do as a lead generation agency in DACH, in your accounts and with the data staying yours.