Fundamentals

Terms around intent signals get mixed up regularly, and because most sources come from vendor marketing, the limits of the method blur along with them. This page sorts out the terms and names what sits behind each one. How to work with the signals is on the use cases page; how they come about technically is under technology and integration.

Signal, data, score

An intent signal is a single observed action with a timestamp: a visit to the pricing page, an opened datasheet, a search query, a click on an ad, a visit to a comparison page. It is an event, nothing more.

Intent data is the collective term for the processed set of such events, usually aggregated across several sources and over a period. Individual actions become a trajectory.

An intent score is the compression of that trajectory into a number. A score is always a modelling assumption: someone decided that a pricing page visit counts for more than a blog visit. That weighting is a hypothesis about your business, not a law of nature, and it belongs under review as soon as enough closed deals exist to check it against.

Anyone who keeps these three levels apart understands most vendor promises better. A high score means: the model considers this account interesting. No more than that.

Explicit and implicit signals

Explicit signals arise when someone actively discloses something: a completed form, a demo request, a newsletter subscription, a question in a chat. They are unambiguous and rare.

Implicit signals arise in passing, from observed behaviour: which pages were opened, for how long, how often, in what order. They are frequent and ambiguous. The entire effort around intent data revolves around extracting a pattern from many ambiguous observations that comes close to the informative value of an explicit signal.

Three origins

First-party data comes from your own channels: website, shop, newsletter, CRM, support. It is the most precise data you can get, because it relates directly to your offering and you control the collection. Its disadvantage is the narrow view: you only see who has already been with you. Whoever is currently reading at a competitor does not appear.

Second-party data is someone else’s first-party data that you use with their agreement. Typical sources are review and comparison portals or trade media where your category is researched. They deliver hints from the phase in which vendors are compared, but are bound to the reach of that one source.

Third-party data is compiled by providers across many websites and sold as a data stream. The appeal lies in early detection: an account can stand out before it has ever been with you. The flip side is a lack of verifiability. You see the result, not the collection, and you are liable for the processing regardless. Origin, legal basis and currency need clarifying before purchase, not after.

Origin	Phase	Strength	Limit
Third-party	early research	broad view, accounts before first contact	collection not verifiable, usually only coarse
Second-party	vendor comparison	concrete category relevance	bound to one source
First-party	concrete evaluation	precise, controllable, current	only those already there

In practice, almost every sensible project starts with first-party data. It costs nothing but care, its legal basis can be established, and it shows immediately whether there is enough movement to derive anything from at all.

Company instead of person

In German-speaking markets this is the most important distinction in the whole discipline. Methods that map website visitors to companies work at the level of the organisation: they establish that an access came from the network of a particular firm, not who was sitting there.

That is not technical modesty but the point at which the assessment is decided. As long as the analysis ends at the company, it has a different character from one that recognises individuals. As soon as signals are linked to names, email addresses or CRM contacts, you leave that level. Whether and under what conditions that is permissible is a legal question and belongs settled in advance.

For practice it also means: you learn that a company is occupied with a topic. Whether that was the specialist department, IT, an intern or a competitor is not part of it.

Distinguishing from firmographic and demographic data

Firmographic data describes the organisation: industry, headcount, revenue band, location, legal form. It answers whether a company fits your offering at all. These details are stable and go out of date slowly.

Demographic data describes people: role, department, level of responsibility. It answers who should be approached.

Intent data answers the third question: when. It is volatile and loses value within weeks.

Only together do the three form a picture. A fitting company without current interest is an address for later. A strong signal from a company that does not fit the offering at all is noise. Handling either one wrongly costs sales time.

What the quality of a signal is measured by

Currency. A hint from yesterday is something different from one from three months ago. Research phases are short; arrive too late and you find a decision already made.

Frequency. A single visit is chance. Several accesses from the same company within a few days are a pattern.

Depth. Opening the homepage says less than opening prices, technical specifications or references. Which pages indicate depth in your case depends on the offering and can be derived from closed deals.

Fit. Without a comparison against the ideal customer profile, every signal is worthless. That comparison is the first filter, not the last.

What a signal does not say

It does not say that someone wants to buy. Research also happens out of curiosity, for a university paper, for market observation, or because a competitor is checking what you cost.

It does not say who can be approached. Between the company and an actual contact person lies work.

It says nothing about budget, timing or authority. No data record takes that qualification off your hands.

And it is not complete. A considerable share of accesses cannot be attributed to any company – from home offices, over mobile networks, through providers with large shared address ranges. The technical reasons for that are on the technology page. Anyone promising ninety per cent coverage should be able to explain how they arrive at that figure.

The precondition that is often missing

Before intent data delivers anything at all, your own measurement has to be right. If page views are not recorded cleanly, internal accesses are counted in, or bots slip through, what you get is not signals but artefacts. How a dependable collection is built is on the web analytics page.

The B2B intent signals entry page gives an overview of the whole series. If you want to judge how much substance sits in your own data, drop us a line – you can also book a slot directly there.