Intent data is guesswork. Public signals are evidence.

The B2B sales industry has spent a decade treating intent data as the gold standard for identifying in-market accounts. The problem is that most intent data does not tell you a company is ready to buy. It tells you a company's employees have been reading about your category on the internet. Those are not the same thing.

Cameron Evans · Founder & CEO, Haku

14 June 2026 · 8 min read

When Bombora launched the concept of third-party intent data in 2014, it was genuinely innovative. The idea was simple: aggregate anonymous content consumption signals from thousands of B2B publisher websites, identify which companies were reading heavily about specific topics, and surface those companies to vendors as likely in-market buyers. For the era of spray-and-pray outbound, any signal that helped prioritise who to contact first was valuable.

The problem is not that third-party intent data is useless. In the right context, for the right team, it still has a place. The problem is that it has been sold as something it is not — as a reliable indicator of genuine buying intent — and that teams have built significant budget commitments around signals that are far weaker than they appear.

At Haku, we made a deliberate decision not to use intent data of this kind. Not because it does not exist, but because we believe publicly observable signals — the things a company is actively doing and announcing in the world — are a fundamentally more reliable basis for outbound than inferred content consumption. This post explains that distinction, and why it matters for every outbound motion you run.

The three types of intent signal — and what each actually tells you

The intent data landscape is usually described in three tiers. Understanding what each one actually measures — and what it does not — is the foundation for making sensible decisions about which signals to build your outbound motion around.

First-party signals are the strongest because they reflect a company that has actively engaged with you. Second-party signals are strong because they reflect active research on a platform with genuine buying intent — a company comparing products on G2 is further down the funnel than a company reading a generic industry article. Third-party signals are the weakest because they are inferred — nobody at the flagged company raised their hand, and the signal is anonymous, company-level, and based on the assumption that reading about a topic correlates with being ready to purchase a solution for it.

The specific problems with third-party intent data

Bombora is the market leader in third-party intent. Its co-op of 5,000+ B2B publisher websites aggregates anonymised content consumption signals and sells them as Company Surge scores — indicating which companies are researching a given topic more than usual. ZoomInfo bundles Bombora intent into its Advanced and Elite tiers. 6sense and Demandbase build their scoring models around similar data.

None of this is without value for the right use case. But the limitations are significant, and they are not always prominent in the sales conversation when these platforms are purchased.

It is company-level, not person-level

Bombora signals tell you that someone at a company has been reading about your category. They do not tell you who. They do not tell you whether it was the CFO evaluating vendors or an intern doing desk research. They do not tell you whether the person with budget authority is involved at all. For outbound purposes, a company-level signal without contact-level context is a very blunt instrument.

It has inherent latency

Bombora refreshes Company Surge scores weekly. In practice, the signal you act on can reflect behaviour from up to seven days ago — which, in a competitive outbound environment where buying windows are narrow, is a meaningful disadvantage. The accounts flagged as surging this week may already have shortlisted vendors, booked demos, or moved on entirely by the time the signal reaches your rep.

The inference is fragile

Content consumption is a weak proxy for purchase intent. A company whose employees are reading about your category might be evaluating vendors. They might equally be tracking a competitor, producing their own content, training a new hire, or conducting academic research. The signal does not distinguish between these cases. Teams that act on every surge alert without further qualification end up with large volumes of outreach to accounts that have no live buying initiative at all.

“Intent data tells you what a company might be thinking. Public signals tell you what a company is actually doing. Only one of those is evidence.”

What public signals are — and why they are different

Public buying signals are observable, verifiable events. They are not inferences drawn from anonymous browsing behaviour. They are things that happened — announced, filed, published, or posted — that you can independently verify and that carry a direct implication about a company's current situation.

Haku monitors a wide range of public signals across every data source that companies make publicly available. The signals that matter most are not generic — they are specific to each client's ICP, modelled against what has actually preceded purchase in their customer base. But the categories of public signal that most reliably indicate a buying window are consistent:

Signal quality comparison: third-party intent vs public signals

Why the outreach hook matters as much as the signal itself

There is a practical dimension to this distinction that goes beyond signal quality — and it is one that every rep immediately understands. The outreach you write from a public signal is fundamentally stronger than anything you can write from a third-party intent flag.

A Bombora surge tells you a company is reading about your category. The best outreach you can write from that signal is something like: "I noticed [Company] has been researching outbound tools recently...". This sounds presumptuous, often inaccurate from the prospect's perspective, and fails the basic test of relevance — the prospect has no idea what you are referring to, and their immediate reaction is suspicion rather than curiosity.

A public signal gives you a specific, verifiable, congratulatory opening. "Saw you've just brought on [Name] as VP Sales — great hire. At the stage you're at, building the outbound motion underneath a new sales leader is usually the next challenge...". That message references something real, something the prospect is proud of, and something that directly connects their current situation to the problem you solve. It lands differently. It earns a reply at a meaningfully higher rate.

The signal is not just a trigger for outreach. It is the substance of the outreach itself. That is only possible when the signal is a verifiable public event — not an anonymous content consumption score.

How Haku builds custom signal models for each client

Not all public signals are equally predictive for every ICP. A funding round is a strong signal for a team selling to growth-stage SaaS companies. A Companies House filing showing revenue growth is more relevant for teams selling into established UK mid-market businesses. A new VP of Sales hire is one of the strongest signals for Haku itself — a new sales leader is precisely the person most likely to evaluate a new sales intelligence platform.

Before Haku surfaces a single account, it ingests your CRM data, your closed-won analysis, your case studies, and your product information to understand which signals most reliably preceded purchase in your specific customer base. The signal model it builds is not generic — it is tuned to your ICP, your category, and the specific triggers that have historically indicated a genuine buying window for what you sell.

Haku then monitors those signals continuously across every publicly available data source — LinkedIn, job boards, Companies House, Google Maps, news, forums, product databases — surfacing accounts when multiple signals align and the context suggests a rep's time will be well spent. The result is a list that is small, specific, and grounded in evidence rather than inference. Every account on it has done something you can verify. Every outreach built on the back of it has a specific, relevant reason to exist.

That is what signal-based selling actually means. Not a dashboard full of surge scores. A targeted list of companies where something real has happened — and where a well-timed, well-crafted message has a genuine chance of starting a conversation.