Alternatives to bidstream-based intent data, source by source

This guide is for the revenue operations, marketing or data lead who licenses intent data built on the advertising bidstream and is weighing what could replace it or sit beside it. It sets out the alternatives to bidstream-based intent data by what each source observes, at which level, and what evidence reaches you, as each vendor's own documentation describes it. By the end you can name the kind of source behind any score, put the right questions to its vendor, and test a candidate on your own closed opportunities.

No source is ranked here: each kind observes a different thing, and two can run side by side. The comparison of company signals and third-party intent data treats the two broad methods at length.

What bidstream-based intent data observes

Bidstream data is what bid requests carry in programmatic advertising. The IAB Tech Lab, which publishes the OpenRTB standard and uses the word bidstream for that flow, describes real-time bidding as a way of transacting media in which an individual ad impression is put up for bid in real time. Its OpenRTB 2.6 specification says that for each inbound ad request, bid requests are broadcast to bidders. Among the fields it defines for a bid request are the address of the page where the ad will show, the referring address, the browser's user agent string, the IP address closest to the device, the device's location, and user IDs specific to the exchange and to the buyer.

An intent vendor with access to that flow reads it as a record of pages being read. Demandbase states on its intent page that it pairs its own technology for identifying unknown accounts with direct access to the bidstream, and that intent is drawn from the content of the page. An article on its site, dated 8 January 2026, sets out the steps: each content consumption event holds the visitor's IP address, an anonymised cookie, a timestamp and geo-location information, the traffic is tied to an account with the cookie and IP information, and the content of the page is classified by natural language processing.

So the observation is a page loading where an ad slot was offered, the unit observed is a browser or device on a network, and the account is resolved from it. Each alternative changes at least one of the three.

What are the alternatives to bidstream-based intent data?

There are four kinds of alternative to bidstream-based intent data. First-party activity records what accounts do on properties you run. A publisher co-operative records reading on its member sites. A software review site records research on its own pages. Public company signals record what a company itself does: a tool adopted, a role opened, a filing made.

The table sets out the four.

The level says whether a person or a device stands between the observation and the account, which is where consent and identity come in. The evidence says what a rep or a model reviewer can open when asked why an account is on the list. The guides to Bombora alternatives, 6sense alternatives and Demandbase alternatives take those products in turn.

Alternatives to bidstream intent data by signal: first-party, technographic, product usage and community

The signals that can stand in for bidstream intent data fall into two families. First-party signals come from properties you run: your website, your product and your community. Public signals come from the company itself: its technology stack, which is technographic data, its job postings, its filings and its announcements. Each records an act with a date.

First-party web activity

First-party data starts with a tag on your own site: the tag records visits, and its vendor resolves each visit to a company, in the ways the provider table further down sets out. Read this source as complete for the accounts that visit: its record opens with an account's first visit.

Product usage and community signals

Product usage is the record of what signed-in accounts do inside your own product: a trial started, a seat added, a feature first used. The account is known from the sign-in, so no visit has to be resolved to a company. Community signals come from the places you run where practitioners discuss your category: a forum you host, a public code repository you maintain, your events. There the unit is a member's handle, so the work is tying a handle to an employer.

Technographic and other public company signals

Here the observation is the company's own act, so no visit and no reader enters the record. Technographic data records the tools a company runs on its website and when each was first and last seen. Job postings name the roles a company is hiring for and the tools those roles use. Filings and announcements record capital raised, launches, partnerships and leadership changes. BuiltWith presents itself as the way to find out what websites are built with and offers lists built on technology signals.

The Fokals intent dataset delivers a weekly score from 0 to 100 for each company and topic, with a surge flag and the dated evidence behind every score, written once after the week closes. Read any signal for what it records: a detection shows presence on the website, and a posting records an intention to hire.

Intent data sources that do not rely on cookies or a publisher co-operative

Three kinds of intent data source work without a publisher co-operative: first-party activity on your own properties, research a review site or a single publisher records on its own pages, and public company signals. Whether a source also works without cookies is answered in each vendor's documentation. Public company signals involve no visitor, so no cookie arises.

A regulator's wording is wider than the word cookie. The UK's Information Commissioner's Office says in its guidance on storage and access technologies that the UK regulations known as PECR apply to any technology that stores information on a subscriber's or user's device or accesses information stored there, with cookies, tracking pixels, fingerprinting techniques, scripts and tags among its examples. Its guidance on identifiers names IP addresses and cookie identifiers together among online identifiers, with advertising IDs, pixel tags, account handles and device fingerprints, and says organisations must consider whether such identifiers can distinguish one user from another. Other jurisdictions set their own rules, and your counsel decides which apply to you.

So ask more than whether a cookie is set. Five questions separate the sources.

1. Is the thing observed a person's visit or a company's own act?
2. On whose pages is it observed, and under what agreement with their owner?
3. Which identifiers tie it to a company: a cookie ID, an IP address, an advertising ID, a sign-in?
4. Is anything stored on a visitor's device or read from it, and on what basis?
5. Does anything about a person reach you, or the account alone?

How G2, Bombora and 6sense source B2B intent data: first-party research, a publisher co-operative or cookies

G2, Bombora and 6sense document three different sources. G2 Buyer Intent records research that buyers do on G2 and on Capterra, Software Advice and GetApp. Bombora runs a publisher co-operative, with its tag on member sites. 6sense combines keyword intent from B2B publisher content, activity on its customer's own website and intent from partners, Bombora and G2 among them.

G2's Buyer Intent documentation says a signal is triggered when a buyer interacts with a vendor's product profile, compares the product with a competitor or views alternatives in a shared category. Its data reference returns the buyer organisation's name, domain, country, industry and number of employees, with the country and region of the visitor who triggered the signal. We found no description in those two pages of how a visit is matched to an organisation, and no mention of cookies, so put both questions to G2.

Bombora's Co-op page describes direct relationships with member publishers and brands, its own JavaScript tag on their websites and engagement captured from gated, paywalled and open content, and it calls the data set consent-driven. Company Surge scores every pairing of account and topic against that account's own baseline. Its privacy policy lists the technical identifiers in its event data, cookie IDs and IP addresses among them, and says events are aggregated into one company profile where they identify the company a person represents.

6sense's Signalverse page lists keyword intent drawn from buyer research across B2B publisher content, and partnerships for topic intent with G2, TrustRadius, Bombora, PeerSpot and Gartner. Its FAQ on third-party cookie deprecation says its core advertising features rest on identifiers other than third-party cookies, notably IP, for company identification.

Which intent data providers use first-party, IP, firmographic or contextual sources in place of cookies?

First-party, IP, contextual and firmographic name four different steps in an intent data source, and one provider can use several. First-party is where the activity happens: on your properties or the provider's own. IP is how a visit is tied to a company. Contextual is how a page gets its topic. Firmographic data describes the company: fit, not intent.

Contextual is the remaining word. The IAB Tech Lab's Content Taxonomy is a common language for describing content, with contextual targeting among its typical uses. A contextual label describes a page, so it becomes account intent only when a method in the third column ties the reader to a company. Firmographic fields travel with several of these products, as in G2's data reference, and answer whether an account fits.

A worked example: testing a replacement on closed opportunities

The fair way to choose is a look-back test on your own pipeline. The example is illustrative: Acme Robotics is an invented company, and every figure is invented to show the arithmetic and describes no real source.

Assumptions:

• Acme Robotics opened 100 opportunities with target accounts in the last two quarters.
• Each is paired with a matched account of the same industry and size band that opened none, which takes its pair's opportunity date as its reference date.
• Acme holds the weekly flags of the feed in use and of a candidate source for the whole period, each as delivered in its week.
• An account counts as flagged when a source flagged it, on a topic Acme sells into, in the 90 days before the reference date.

The ratio is the first count divided by the second. Read the table in three steps.

1. Reach. The feed in use flagged 24 + 30 = 54 of the 100 opportunities beforehand, the candidate 24 + 16 = 40, and the two together 24 + 30 + 16 = 70.
2. What a swap changes. Replacing the feed with the candidate gives up the 30 opportunities that only the feed saw and gains the 16 that only the candidate saw, so reach moves from 54 to 40. Adding the candidate moves it from 54 to 70.
3. Agreement. Accounts flagged by both sources appear six times as often among opportunities as among their matches, 24 against 4. A flag from one source alone appears twice as often: 30 against 15, and 16 against 8.

On these figures Acme would run both and work the agreed accounts first. Other figures would support a swap.

Two conditions keep the test honest. Fix each source's threshold before you look at the outcomes, and read each flag as it stood in its week, so ask each vendor whether a past week is ever revised. Fokals writes each weekly score once, after the week closes, and never revises it, which makes the record point-in-time by construction. Then run a second test that needs no outcomes: take twenty flagged accounts from each source and count the flags for which a rep can open a dated source.

Frequently asked questions

What this page says about the products it names was checked against their public documentation on 7 October 2026. Product and company names are trademarks of their owners. Fokals is not affiliated with them or endorsed by them.