Intent data vs buying signals: what each one tells you

Intent data is evidence that people at a company are researching a topic, rolled up to the company as a whole. Gartner’s March 2024 note on intent programs says third-party intent data lets an organization “identify demand for their products by synthesizing signals from digital buyer behavior.” A provider collects reading and research activity across many sites, matches it to companies, and scores how far each company’s interest in a topic has risen above its usual level.

It is widely used, mostly to choose which accounts see ads and which accounts sales looks at first.

Two limits are built in. Intent data usually describes a company as a whole, and it comes from a model, so there is no document to open. It shows that a company is reading. It does not show who is reading, or why.

Intent data comes from four kinds of source, and each one sees a different slice of buyer behavior.

• First-party intent: visits and engagement on your own website and emails, from your own analytics. It is the only kind that shows interest in your own company.
• Third-party co-op intent: publishers and B2B media sites pool reading activity, and a provider flags companies reading more than usual about a topic.
• Review-site intent: visits to product categories, comparisons and competitor pages on software review sites. Forrester calls signals from owned properties like these second-party.
• Advertising-network intent: some providers also draw signals from online advertising activity. Ask how each signal is matched to a company before you rely on it.

The business model matters too. Forrester’s July 2023 review of 14 intent providers sorts them into four models: traditional data providers that sell third-party signals by volume, ABM platforms that bundle signals into the platform subscription, campaign execution firms that often price per lead, and walled gardens that sell second-party signals from their own properties, such as review sites. The model tells you how you will pay.

SalesOne’s signal library lists intent data and first-party engagement as relationship signals, the one group with no dated public event behind it; in this guide, “buying signal” means the dated event.

A buying signal is a dated event at a company that makes a purchase more likely, with a source anyone can open; intent data is an estimate of interest.

Gartner’s B2B buying journey research states that “99% of B2B purchases are driven by organizational changes”. A new COO, a plant expansion, an acquisition or a system on its way out is that change. It is usually public, in a press release, a filing, a permit or a company page, and it has a date. That is what SalesOne means by a buying signal, and it is the same definition used across this site.

Forrester drew the line in a 2019 post on account-based planning: “Intent is not a buying signal.” It called intent “a prioritization tool” and warned that intent signals do not mean an account is ready to talk to sales. That is still the clearest way to hold the two apart.

They answer different questions: intent data says which companies are reading about a topic, and a buying signal says what changed at one company and when.

Neither column wins on its own. Intent is broad and early, and it covers companies you would never have thought to research. A buying signal is narrow and specific, and it gives you something true to say. The useful question is which one you need for the job in front of you.

Intent data is an estimate, and we did not find an independent, published test of its accuracy with a stated sample, so this page gives no accuracy figure.

What can be said is where error enters. Between a person reading an article and a score in your CRM sit three steps: the activity has to be captured, matched to the right company, and judged against that company’s normal level. Each step is a model, and errors at one step carry into the next. Gartner’s March 2024 note on the subject is titled “Optimize Intent Data Use: Overcome 5 Potential Points of Failure”.

The honest test is your own. For one quarter, compare the companies a provider flags with the companies that later became opportunities, and with the ones your team rejected for fit. The overlap is the accuracy figure that matters for your market. Ask the provider, too, which sources feed each topic and how companies are matched. If the answer is vague, count that against the provider.

Intent data earns its cost when more companies fit your profile than your team can work, and you need a way to rank them week by week.

• A large market: thousands of companies fit, and ranking them is the hard part.
• Ad targeting: you want to spend on companies reading about your topic now.
• Inbound-heavy teams: intent helps decide which known accounts get a follow-up first.
• Less useful: a market of a few hundred companies you can research one by one, or buyers whose research happens mostly offline or through peers.

Timing is the catch: by the time a company is reading heavily, it may already have a favorite vendor.

If you already buy intent data, Forrester’s July 2023 guidance is to add a second provider only if its signals are “at least 30–40% unique” from what you have. To compare the platforms that sell it, see 6sense alternatives.

Use intent data to decide where to look first, and a buying signal to decide what to say.

1. Start from fit. Only companies that match your ideal customer profile go further.
2. Use intent to rank which fitting companies get researched this week.
3. Research each one for a dated event inside its window, with a source.
4. Write from the event, never from the intent. Telling a prospect you saw them reading reads as tracking.
5. If no event turns up, keep the company with marketing and out of sales sequences.

A sample account, with a fictional company

Say an intent feed shows Halvorsen Castings, a manufacturer in Toledo, Ohio, reading more than usual about supplier risk this month. That says the topic is live. It does not say why. Research finds two dated events: a new COO named on Aug 25, 2026, and a county permit dated Sep 12, 2026 for a 140,000 sq ft foundry expansion in Dayton. The first message names the Dayton expansion and the supply risk a plant that size creates for the VP of Supply Chain. The intent reading did one job: it moved Halvorsen to the top of this week’s research.

In an account scoring model, intent works best as a timing input next to fit; it cannot stand in for a dated reason.

Not long: SalesOne’s default window for topic research and for site engagement is 0 to 30 days, the shortest in its signal library.

Interest moves on quickly. A spike in reading can end when a project closes or the person researching moves to something else, and the score drops without leaving a reason behind. A past deal or former customer becoming active again lasts a little longer, with a default window of 0 to 90 days.

Events last longer because the work they create lasts longer. In SalesOne’s defaults, a new executive counts as a signal for up to 180 days, because reviews of budgets and vendors run about two quarters, and an acquisition for up to 730 days, because integration runs in phases. A dated event also stays true after its window closes. Only the timing expires, so the account can be checked again later.

In practice, act on intent within days or treat it as stale, and plan outreach from events over weeks. Signal-based selling explains how to set a window for each signal type.

SalesOne works from dated, sourced public events. It has no anonymous intent data and no website visitor identification.

S1 deep research agents search public pages for dated events, check each company they name against your ideal customer profile, and accept or reject it with a reason. Every buying trigger comes with its source and the date it happened. In our runs, the newest trigger on a typical account was about two months old. Each trigger has a response window, and SalesOne puts accounts with an open window first and shows the date to reach out by.

Every company S1 turns down keeps the rule it failed and the evidence, so you can see why it is not on your list. S1 reads public pages only. It never logs in and never gets past CAPTCHAs. See AI account research for how an account is checked.

If your team already buys intent data, keep it for ads and for ranking. SalesOne answers the other question: what changed at the company, and who owns the problem it creates.