B2B Intent Data Explained: First, Second and Third Party

The short answer

First-party intent data is the most specific and the easiest to trust, because it is about your company and you know how it was collected, but it only covers buyers who reach you. Second-party data covers buyers researching your category on someone else's site, usually at company level. Third-party data covers the widest audience and says the least: that a company appears to be reading more than usual about a topic. Start with the first-party signals you already have, and test any bought data against your own closed deals before you act on it.

What intent data is, and what it is not

Intent data records behaviour: pages read, content downloaded, comparisons viewed, forms filled in. A tool then infers from that behaviour that an organisation may be in the market. That makes it different from two other kinds of B2B data it is often sold alongside:

Fit data (firmographics and technographics) describes what a company is: its sector, size, location and the software it runs. It tells you whether a company could buy, not whether it is looking.
Contact data describes people: names, job titles and email addresses. It tells you who to talk to, not whether they are interested.

Intent data answers a narrower question: is anyone at this company showing interest right now? Three cautions apply to every kind.

It is inferred. Someone reading about a topic might be buying, studying, writing a report, applying for a job or checking on a competitor.
It is usually company level. Most intent data names an organisation, not a person. Ten signals from one company could be one person or ten.
It decays. A burst of research is most useful while it is happening. A signal from two months ago is history.

The three types at a glance

	First party	Second party	Third party
Who collects it	You	Another organisation, which shares it with you directly	An aggregator, from many sites it does not own
Where	Your website, forms, emails, product, events, CRM	A review site, publisher or event partner	A network of publisher and business websites
What it shows	Interest in your company and specific pages	Interest in your category, your product and competitors, on that site	Rising interest in a topic, across the network
Usual unit	Person (if they gave details) or company (if identified from a visit)	Company	Company
Main strength	Specific to you; you know how it was collected	Shows research before buyers reach you	Widest reach, including companies that have never heard of you
Main limit	Only covers buyers who already found you	Only covers that one site's audience	Least specific; method is often opaque

First-party intent data

First-party intent data is everything your own systems record about how organisations engage with you. You already hold most of it, often spread across several tools.

How it is collected

Hand-raisers: demo requests, contact forms, pricing enquiries, chat conversations and replies to emails. These are the strongest signals any team has, because a person has asked for something.
Known-person behaviour: once someone has given their details, marketing automation can record which pages they view, which emails they open and which events they attend.
Anonymous visits, resolved to a company: most visitors never give their details. Visitor identification tools look up each visit's IP address and attach a company name where one can be found. How that lookup works, and why it misses so many visits, is covered in how IP-to-company lookup works.
Product and customer data: free trial use, logins, support tickets and renewal conversations, for existing customers.

What it can tell you

Which pages a company or person read on your site, in what order, and how recently.
Whether that reading includes high-intent pages: pricing, integrations, comparison pages, case studies in their sector, the demo page.
Whether an existing customer is reading pages that suggest expansion, or cancellation.

What it cannot tell you

Anything about buyers who are researching your category but have not reached your site.
The company behind most anonymous visits. Home broadband, mobile networks and VPNs hide the employer, so identified visits are a partial, larger-company-biased sample.
Why someone visited. A job applicant, a supplier and a competitor read the same pages as a buyer.

Second-party intent data

Second-party data is another organisation's first-party data, shared with you directly under an agreement with that organisation. In B2B the common sources are review and comparison sites, specialist publishers that run content syndication programmes, and event or webinar partners.

How it is collected

The partner records activity on its own property and passes you the part that concerns your product or category. A review site is the clearest example. G2's documentation, mentioned here only to show what such data contains, describes signals including when a buyer views your product profile, a category page that includes your product, a comparison page that includes your product, or a pricing, profile or compare page for a competing product in a shared category. It reports these as company-level information and summarises each buyer with a buying stage (awareness, consideration or decision) and an activity level¹.

Content syndication works differently: a publisher promotes your guide or report to its audience and passes you the details of people who chose to download it, under the terms the publisher showed them.

What it can tell you

That a company is comparing products in your category, sometimes before it has visited your own site.
Which competitors it is looking at alongside you, where the source records comparisons.
A rough stage of research, as the source defines it.

What it cannot tell you

Anything outside that one partner's audience. If your buyers do not use that site, it shows little.
The partner's method in full. How a company is identified, and how a stage is assigned, is the partner's own model; ask before you rely on it.
Whether a content download meant buying interest. A person who downloaded a report to read on a train is still counted.

Third-party intent data

Third-party intent data is gathered by a company that observes reading across many websites, none of which belong to you or to it, and sells the resulting signals. It aims to show which organisations are researching a topic anywhere in its network, including companies that have never visited your site.

How it is collected

The details differ by provider, but most third-party intent data goes through the same four steps:

Observe reading across a network. Participating websites run the aggregator's tag or tracking code. Bombora, for example, describes a co-op of publishers, brands and B2B media sites that run its proprietary JavaScript tag, and calls the resulting data set "consent-driven"². HubSpot describes its research intent feature as powered by a network of websites that have its tracking code installed and have opted into intent data sharing³.
Classify the content. Each page read is mapped to one or more topics from the provider's taxonomy, so a page about data warehouse migration counts towards that topic.
Resolve the reader to a company. The visit's IP address, and sometimes a cookie or other identifier, is used to decide which organisation the reader belongs to. This step has all the weaknesses of IP-to-company lookup.
Compare with a baseline. A company is flagged when its reading on a topic rises above its own normal level. HubSpot, for instance, sets its research intent levels by comparing research activity in the last 30 days with the 30 days before³.

You then usually choose a list of topics to watch, and receive the companies whose activity on those topics has risen.

What it can tell you

That a company appears to be reading more than usual about a topic related to what you sell.
Which accounts on a target list show any research activity at all, which can help decide where to spend limited outreach or advertising.

What it cannot tell you

Interest in you. A topic signal says a company is reading about a subject, not that it has heard of you or wants what you offer.
Who is reading, or why. One consultant researching for a client can make a company look active.
Small companies reliably. A small firm produces too little reading to rise clearly above its own baseline, and is harder to resolve from an IP address in the first place.
An exclusive edge. The same signals are available to every competitor who buys the same topics from the same provider.
What it misses. Research done on sites outside the network, in search engines, in communities or in conversations with peers does not appear. HubSpot states, for example, that it does not see what a company types into a search engine³.

Observed versus inferred

A useful way to judge any signal is to count how many inferences sit between the behaviour and the conclusion.

A demo request needs none: a named person asked to talk to you.
A company identified reading your pricing page needs one: that the IP address really belongs to that company.
A review site comparison needs one or two: that the visitor was identified correctly, and that the source's stage model is right.
A third-party topic surge needs at least three: that the pages were classified correctly, that the readers were resolved to the right company, and that a rise in reading reflects a buying project.

Each inference adds error. That does not make the weaker signals worthless, but it should decide how you act on them: a demo request deserves a same-day reply from a person, while a topic surge is better used to decide which accounts to watch or warm up.

How to evaluate an intent data source

Before you buy or switch on any intent data, ask the provider, or yourself for first-party sources:

Where exactly is it collected? Which sites or systems, and in which countries. A network strong in US technology publishing may say little about UK engineering firms.
How was the data obtained? What the people being observed were told, and on what basis it is shared with you.
How is a reader resolved to a company? IP address, cookie, login or form details, and what happens when the match is uncertain.
What is the unit? Company, office or person, and whether you can limit it to company level.
How is a signal scored? The baseline it is compared with, the time window, and how often it refreshes.
Can the topics describe your offer? Check whether the taxonomy has topics specific enough to separate your buyers from general readers.
Does it work on your history? Ask for, or run, a back-test: did the companies that became customers last year show signals before they contacted you, and how many companies with signals never did?

How a small team should use intent data

For a team of one to five marketers, the order matters more than the source.

Make the first-party signals reliable. Every demo request reaches a named person quickly, and the team knows which pages count as high intent. Our guide to seeing which companies visit your website covers how to set those pages and filters up.
Combine fit and intent. A signal from a company outside your target market is noise, however strong it looks. Act on companies that are both a fit and showing interest.
Decide in advance what each signal triggers. A hand-raise gets a reply; an identified visit to pricing from a target account gets a review by the account owner; a topic surge adds an account to a watch list. Our inbound lead follow-up process sets out those steps.
Test bought data before scaling it. Run it for a quarter against a fixed list, record what happened to the flagged accounts, and compare with accounts that were not flagged.
Use signals to choose, not to script. Intent data is good for deciding who to contact and what to talk about. Opening a message with what someone was reading tends to feel intrusive.

Privacy and data protection

All three types involve personal data at some point, even where the output is a company name: IP addresses, cookies and contact details can all be personal data in the UK. The questions differ by type. For first-party data they concern your own website and notices; for second-party and third-party data they also concern how the provider collected the data and what people were told. We do not give legal conclusions here. Our guide to visitor identification under UK GDPR sets out the rules that apply, and you should take your own advice on your own set-up.

Where Artiweave fits

Artiweave does not sell intent data. Its work today is on the website itself, proposing evidence-based changes that a marketer approves. Showing which companies visit, a first-party signal at company level, is in development and not available yet; when it is ready, it will show which companies read the pages being improved, not who the individual visitors are. For the wider topic, see our guide to website visitor identification.

Frequently asked questions

What is B2B intent data?

Behavioural signals suggesting that an organisation is researching a topic, a product category or a specific supplier. It is usually reported at company level and is an inference from reading and browsing, not a statement of a plan to buy.

Is website visitor identification a type of intent data?

Yes. It is first-party intent data: it attaches a company name, where one can be found, to visits to your own website. It shows interest in you specifically, but only for the share of visits that can be matched to a company.

Is third-party intent data accurate?

It can be useful, but it is an estimate built on several inferences: that a page is about a topic, that a reader belongs to a given company, and that a rise in reading means a buying project. Ask how each step works and test the data against your own closed deals before relying on it.

Does a small B2B team need third-party intent data?

Usually not first. Most small teams get more from organising the first-party signals they already have, such as demo requests, pricing page visits and email replies. Third-party data is worth testing once there is a clear target account list and someone with time to act on it.

Sources

Improve the pages your buyers read.

Artiweave continuously improves the website your B2B company already has, with every change approved by you. Showing which companies visit is in development. It is in private beta.

Join the waitlist