Two different things are sold as intent data. One measures what people at a company read: articles, reviews, comparison pages. The other records what the company itself does in public: a tool added to its website, a role opened, a filing made. This comparison sets the two methods side by side, using the public method descriptions of Bombora and G2 for the first and of Fokals for the second. By the end you can tell which kind a score is built from, what each kind is built to show, and how to use both in one account ranking.
Fokals is the worked example of the second kind. Its intent scores are evidence-backed: each is built from what a company itself does in public, and every score carries the dated signals behind it, so a rep, an analyst or a model reviewer can open the evidence and check it. The two methods observe different things, which is why this comparison ends with a join that runs them together.
Two methods under one name
Behavioural intent is third-party data about attention. A vendor observes content being consumed on sites it has access to, attributes the activity to a company, classifies the content by topic and reports where a company's reading has risen. Bombora's own definition is that intent data shows when buyers are researching topics across the internet, based on the content they consume. What is observed is a person reading. What is reported is the account.
Company signals are evidence of action. The observation is a dated change in what a company publishes, announces or files, recorded at the level of the company, and each one can be traced to its public source.
The two observe different stages of a purchase, which is the reason to tell them apart. Reading can rise with no budget behind it, and a company can commit to a purchase without anyone reading on a monitored site. Research shows interest in a subject. A public act shows a decision already taken.
How Bombora, G2 and Fokals describe their methods
Bombora's Company Surge is derived from its Data Co-op. Its site says research consumption is collected through a consent-driven tag on every site in the Co-op, and its Co-op page describes direct relationships with member publishers and brands, with engagement captured from gated, paywalled and open content. The site states that Bombora monitors company research activity against more than 25,000 topics. Its site describes scoring that is relative to the account: for every pairing of account and topic, the most recent three weeks of activity are compared with a 12-week historical baseline. Its guidance for customers treats a score of 60 or more on a topic as a spike.
G2 Buyer Intent comes from the research that buyers do on G2. G2's documentation describes it as data about buyers researching a vendor's product across G2, sent at the level of the product, and says it also includes research activity from Capterra, Software Advice and GetApp. A signal is triggered when a buyer views the vendor's product profile or its pricing page, a category page that includes the product, a comparison that includes it or an alternatives page, and a competitive signal records views of a competing product's pages. G2's documentation says two scores are calculated for each company, a Buying Stage (awareness, consideration or decision) and an Activity Level (low, medium or high), and that G2 updates both daily. It adds that the vendor's G2 plan determines which signal types it receives.
Fokals builds its intent dataset from a different kind of observation. A signal is a dated public action by the company, tied to one or more topics in a curated taxonomy of nine groups: a technology adopted or removed, a platform migration, a new market, language, currency or app, a role opened, a tool named in a job posting, a private capital raise reported in a regulatory filing, an announcement. The dated signals of the trailing 90 days, weighted by recency, produce a weekly score from 0 to 100 for each company and topic. Each score carries its five strongest signals as evidence, and a surge flag marks a score that at least doubles against the company's own twelve-week average. The methodology documents the model.
| Bombora Company Surge | G2 Buyer Intent | Fokals intent | |
|---|---|---|---|
| What is observed | Content consumption on the sites of its Data Co-op | Views of profile, pricing, category, comparison and alternatives pages on G2 | Company actions: a tool adopted, a role opened, a filing made, an announcement |
| How it is collected | A consent-driven tag on member sites | Buyer activity on G2 and, as its documentation states, on Capterra, Software Advice and GetApp | From first-party company sources and public records, processed in-house |
| Subjects | More than 25,000 topics, as its site states | The vendor's product, its category and its competitors | A curated taxonomy of buying topics in nine groups, from advertising and marketing to data, technology and business systems |
| Measure | The most recent three weeks against a 12-week baseline, for each account and topic | Buying Stage and Activity Level for each company, updated daily | A weekly score from 0 to 100 for each company and topic, with a surge flag and its five strongest signals attached as evidence |
| Delivery | Integrations with advertising, marketing, sales and CRM platforms, or a flat file or API | my.G2, CSV exports and integrations | REST API and bulk files in JSON, JSON Lines or CSV, delivered direct |
| Built for | Teams that use go-to-market tools: advertising, marketing, sales and CRM platforms | Vendors with a product on G2 | Revenue teams, funds and data platforms that need a dated, verifiable reason behind every score |
The delivery row for Bombora follows its integrations page, and the fields G2 returns, including the buyer's company domain, are in its data reference.
What each method can show
Behavioural intent can show interest before any public act. People read about a category before a role is advertised or a tool appears on a page, and sometimes neither ever follows. It can also separate narrow subjects. A taxonomy as large as the one Bombora states draws fine distinctions between subjects, and G2's signals are specific to one product: its profile, its pricing page, a comparison that includes it.
Company signals show action. A role advertised, a tool installed and a filing made are public acts that have already happened. They also show evidence that a person can check: each signal carries its source, its date and its kind, and each week is written once after it closes and never revised, so the score you ranked on is the score you find when you audit the ranking later.
Read each for what it records. A rise in reading records attention to a subject. A company signal records a public act, so it marks the point at which interest became a commitment: a tool in place, a role funded, a filing made. That is the dated reason a rep can open a call with, and the feature a model can be tested on point-in-time.
Reading the evidence behind a score
The difference is clearest when a rep asks why an account is on the list. With company signals the answer is a set of rows from Intent Scores.
select week_start, score, surge, signals, evidence
from company_intent_weekly
where company_id = :company_id
and topic = 'crm'
order by week_start desc
limit 4;Illustrative: Acme Robotics scores 58 on the CRM topic in the latest week, is flagged as a surge and carries three signals as evidence. The first is a CRM tool adopted on its website. The second is a job posting whose title names the same tool. The third is a new domain verification with a software vendor. Each signal states its date, its kind, its source and its weight.
Three public acts from three sources say more than the number does. The surge flag marks an intent surge: a strong score that at least doubles the company's own twelve-week average for the topic. For signals beyond the five strongest, the Company Signals dataset delivers every signal daily, each with its observation time, its kind, its source, its topics and its weight.
Using both in one ranking
If you license a behavioural feed as well, keep the two sources in separate tables and join them at the account. A behavioural feed may identify companies by domain, as G2's data reference does with its company domain field, while Fokals rows are keyed on the company ID, so the bridge is a table of domains, built by domain matching.
with domains as (
select distinct domain, company_id
from company_technologies
),
acts as (
select company_id, score, surge, evidence
from company_intent_weekly
where topic = 'crm'
and week_start = date '2026-09-28'
)
select b.domain,
b.reading_surge, -- from your behavioural feed
a.score, a.surge, a.evidence
from behavioural_feed b
left join domains d on d.domain = b.domain
left join acts a on a.company_id = d.company_id;Read the result as four cases.
| A company signal on the topic | No company signal | |
|---|---|---|
| Reading has risen | Attention and action agree. Work the account now and open with the public act | Research with no public act yet. Nurture, and watch Company Signals for the first act |
| Reading is flat | A public act without visible research. A decision is under way, so open with the evidence and learn where it stands | No evidence from either method. Keep the account in the plan |
A surge from one source is not comparable with a surge from the other, so do not average them. Keep both columns in the queue and let the rep see which evidence put the account there.
Test the combination before you route leads on it. For each account that opened an opportunity, record whether each source flagged it in the 90 days before, and do the same for a matched set that did not. Fokals weekly scores are written once, after the week closes, and are never revised, so the look-back reads each score exactly as it stood in the week you would have acted on it.
What each is built for
- Behavioural intent is built to observe research: buyers who read for months before they act, subjects as narrow as a single product, and views of your product's profile, its pricing or a comparison that includes it, which is what G2's signals record. The guide to Bombora alternatives covers that kind of source.
- Company signals are built for evidence: a reason that a rep or an analyst can verify at its source, weekly scores written once and never revised for a product or a model, and a point-in-time record that an investor can test as readily as a seller can act on it.
- Both are worth running when a wrong call is expensive, because agreement between attention and action is stronger evidence than either alone.
How Fokals delivers its signals
Scores are written weekly and signals arrive daily. Intent Scores, Company Signals and Company Funding are delivered direct, by REST API and as bulk files in JSON, JSON Lines or CSV, which you load where your scoring runs. The feed of signals runs oldest first from a time you set, so the last cursor is the bookmark for incremental sync. Scores are produced under a named, frozen version, and a breaking change ships as a new version with at least 90 days' notice, so a ranking or a model built on them stays stable.
Two further guides start from here. How Fokals intent scores are built follows one score from its signals to its row, and prioritising accounts with intent scores turns the scores into a weekly queue.
Frequently asked questions
What is third-party intent data?
Third-party intent data is evidence of buying interest gathered outside your own website and CRM and licensed from a vendor. Bombora defines it as signals collected outside a company's owned and paid experiences. The kind that Bombora and G2 describe is behavioural: it measures the content that people at a company consume, on a network of publisher sites or on G2, and reports a rise by topic. It reaches accounts that have never visited you, which first-party data cannot.
What is the difference between behavioural intent data and company signals?
Behavioural intent measures attention: what people at a company read, attributed to the account and scored by topic. Company signals record action: a tool added to the company's website, a role opened, a filing or an announcement, each dated and sourced. The first observes interest in what is read, often early and in narrow subjects. The second records a public act already taken, with a source a person can open and check.
What evidence stands behind a Fokals intent score?
Every weekly score carries its five strongest signals, and each signal states its date, its kind, its source and its weight. A rep can open the job posting or the announcement behind a score before making contact, and a model reviewer can trace a feature back to the public facts that produced it. The Company Signals dataset delivers the full feed daily, for teams that want every signal behind a score and not the strongest five alone.
Can you use Bombora or G2 intent data together with company signals?
Yes. Keep each source in its own table and join at the account, with the company's domain as the bridge. Rank first the accounts where reading has risen and a public act has been observed, treat a rise in reading alone as early interest, and treat a public act alone as a sign that a decision may already be under way. Test the combination against your own opportunities before you route leads on it.
How does an intent surge differ between the two methods?
Both compare a company with its own recent past, but they count different things. Bombora's site describes comparing the most recent three weeks of content consumption with a 12-week baseline for each account and topic. A Fokals surge is a strong score that at least doubles the company's own twelve-week average for the topic, built from public acts. A surge from one source cannot be compared with a surge from the other.
What does a company signal record, and about whom?
A company signal is one dated public act by a company. A record names the company, the topics, the kind of act, its source, the observation time and a weight, and a leadership change is recorded by the role concerned. It is company-level data throughout, drawn from what companies publish, announce and file. Behavioural sources start from the activity of people, so ask the vendor what is collected about individuals, on what basis, and what reaches you.
The queries and code on this page are examples to adapt. Test them in your own environment before you rely on them.
What this page says about the products it names was checked against their public documentation on 4 October 2026. Product and company names are trademarks of their owners. Fokals is not affiliated with them or endorsed by them.