B2B intent signals

In business-to-business, a purchase rarely begins with a call. It begins with someone inside a company having a problem and starting to read: technical articles, vendor pages, comparisons, price information, review sites. Weeks later a shortlist exists, and only then does anyone get in touch. Until that point the entire process is invisible to the vendor.

Intent signals are the attempt to make that phase visible. A signal is an observable action that allows an inference about current interest: the repeated visit to a pricing page, the download of a datasheet, a search for a product category, the opening of a comparison page. Individually these actions mean little. In accumulation, over a short period, and from the same company, they form a pattern.

Why this works differently in B2B than in retail

In consumer business one person usually decides, often within a single session. In business-to-business almost nobody decides alone: the specialist department, IT, purchasing, data protection and management all bring different questions, research at different times, and frequently from different devices. What connects them is the company, not the session.

That is precisely why intent signals in B2B work at company level. The question is not “who was that?” but “which company is currently occupied with our topic?” And that is at the same time the decisive limitation: a signal names an account and a period. It names neither the person nor the intention to buy, and it replaces no qualification.

This sobriety is missing from many accounts of the subject. A large part of the vendors and the professional literature comes from the Anglosphere, where data protection is regulated differently. The procedures described there – particularly those working at person level – cannot be adopted one to one.

What this series covers

The topic falls into four parts that answer different questions. Anyone new to it reads them in order; anyone already working with it jumps to the relevant one.

Fundamentals clarifies the terms. What distinguishes a signal from a data record, what distinguishes first-, second- and third-party data, and how do intent data differ from firmographic and demographic details? It also sets out what a signal explicitly does not say.

Use cases describes what you do with it: sales prioritisation, account-based marketing, ad targeting, content personalisation and monitoring existing customers. Plus the question of what an approach looks like that does not come across as uncomfortable.

Technology and integration goes into the technical side: how does the mapping from IP address to company come about, why does it only work partially, what belongs in the data model, and how do the signals get into CRM and marketing automation?

Trends and strategies places the development in context: the third-party cookie phase-out that did not happen, AI answers in search results and their consequences for the visibility of research, and the question of which advanced procedures actually hold up.

A procedure that is built first and reviewed afterwards gets expensive. Three points need clarifying in advance, with legal guidance rather than in passing:

Whether and from what point the collected data is personal data. Mapping an IP address to a company initially stays at company level; as soon as data is linked to individuals, the assessment changes.

Which legal basis the processing relies on and what of it requires consent. The German TDDDG governs access to information on the end device; the GDPR governs the processing. The two have to be examined separately.

How the processing is documented – in the record of processing activities, in the privacy policy, and in the contracts with the service providers involved.

How this connects to the rest of your measurement

Intent signals are not a separate universe. They build on the same data collection as regular web analytics and suffer from the same gaps: missing consent, ad blockers, bots. Anyone who does not have measurement on their own website under control will not extract dependable signals from it either. The order is therefore: measure cleanly first, then enrich.

The operator of this website develops Webmetic, a tool of its own in this field. It identifies companies from anonymous website traffic without setting cookies.

What getting started looks like

A small, clearly bounded trial makes more sense than a platform rollout. One segment, one use case, a defined period, and a metric that afterwards shows whether it achieved anything. What needs to be in place for that is on the use cases page.

If you want to judge whether the topic is worth it for your business at all – with very small target markets or very low order values it often is not – drop us a line. You can also book a slot directly there.