Person Level Intent Data: What It Is and Who Really Has It

Person level intent data is the biggest oversell in B2B sales intelligence right now.

The idea behind it is sound. Account-level intent has a known flaw: it names the building, not the desk. Person-level closes that gap by attaching the signal to an individual.

The problem is how many vendors claim to do it. A large share of what gets sold as person-level intent is account-level intent with a contact lookup stapled on top. The price gap between those two things is enormous.

Below: how person-level intent gets built. What it costs. The match rates nobody prints on a pricing page. Where compliance blows the whole thing up. And five questions that expose a fake claim on a vendor call.

What is person level intent data?

Person level intent data ties a research behavior to a named person, not to a company.

Instead of "Acme Corp is researching CRM software," you get four things. A name. A title. The topics that person engaged with. The date.

A rep can open that record and write. No guessing at titles. No forty minutes of LinkedIn archaeology.

It sits one layer below B2B intent data as a category. Same plumbing. Finer resolution.

The three parts of a real person-level signal

1. A resolvable identity.A name that maps to a work email, a LinkedIn profile, or a record already in your CRM. "Anonymous visitor at Acme" fails this test even when the company matched.

2. The behavior itself.Which topics. Which pages. Which comparison queries. Broad clusters like "sales technology" tell a rep nothing they can use in an email.

3. A per-person timestamp.Not a weekly rollup. The day, ideally the hour. A signal aggregated into a seven-day window has already lost value before you open it.

Drop any one of those and you have something weaker sold under a stronger name.

Person based, people based, contact level: are these different products?

No. They describe the same idea.

• Person based intent data is Demandbase's naming. Their feature matches keyword signals to known contacts, then ranks them inside a buying group.
• Contact level intent data is the term sales-led vendors use, along with much of the visitor identification market.
• People based intent data comes from ad tech, where platforms target individuals rather than companies.

One caveat before you treat them as interchangeable. "People-based" often means an ad-targetable record, not a name your SDR can call. Different jobs. Different prices. Ask which one the contract covers.

Account level vs person level intent data

The table hides the part that costs money. Account-level intent creates a second step. That step eats the timing advantage you paid for.

Here is the sequence.

1. The signal lands. Acme is surging on your topic.
2. A rep opens the account and guesses at three plausible titles.
3. Two of those people left the company months ago.
4. The rep sources contact details for the third, then writes.

Call it two to three days at a realistic pace across a full territory. By then the buyer has read four more vendor pages.

This is why buying triggers that predict deals get graded on freshness before anything else. A stale trigger is trivia.

The four ways person level intent data gets built

Four methods produce it. They are not equally reliable, and vendors rarely volunteer which one they use.

1. Identity graphs and cross-site pixel networks

A provider runs pixels across a network of publisher sites. Browsing behavior gets logged against a persistent identifier. That identifier links to a hashed email or device ID inside an identity graph. The graph resolves it to a person.

Two matching methods sit underneath this:

• Deterministic matching joins on an exact shared identifier. A hashed email. A user ID. A phone number. High confidence.
• Probabilistic matching scores patterns across IP, device, user agent, and location. A prediction, not a fact.

Vendors blend both and report one number. Ask for the split.

The honest limit: accuracy depends on cookie persistence and on the graph staying current. People change jobs. Browsers block trackers. The degradation happens quietly, and no vendor publishes a decay rate.

2. First-party website de-anonymization

Someone visits your own site. A tool matches the session to a person. It uses an identity graph fed by newsletter signups, form fills, and data partnerships.

This is the highest-quality signal in the category. The person came to you. They are already partway through the B2B buyer journey.

The numbers worth knowing:

• Between 97% and 98% of B2B website traffic stays anonymous. Form fills alone leave nearly all of it invisible.
• Company-level identification resolves 30% to 65% of visitors through reverse IP lookup.
• Person-level identification resolves 20% to 40%. RB2B reports 70% to 80% US person-level identification on the traffic it can see, which is a narrower claim than it first reads.

So for a site doing 4,000 monthly visitors, person-level resolution might surface 60 to 100 usable names. (That range swings hard by industry, so read it as a shape and not a benchmark.)

Which is fine, until someone on the exec team divides the contract value by 60 and asks a question you cannot answer.

3. Declared intent from publisher networks

TechTarget and similar media businesses run this model. Someone downloads a whitepaper or registers for a webinar and hands over their details willingly.

The buyer declares the signal. Nobody infers it.

• Cleanest consent story of any method here.
• Highest cost per contact. TechTarget Priority Engine runs roughly $60,000 to $180,000 a year.
• Coverage maps to that publisher's topics. If they do not cover your category well, you get nothing.

4. Public behavior signals

Job changes. LinkedIn engagement. Review-site activity. Conference attendance. A named person did something visible, and you can reference it directly.

This is the quiet workhorse of the category.

• Tracking job changes surfaces someone who just walked into a role with budget and a mandate. That beats a topic surge in a lot of scenarios.
• G2 review-site intent signals sit here too, though G2 reports at account level for everyone below the top tiers.
• Funding rounds, leadership changes, and headcount growth all belong in this bucket.

I lean on this category more than the first one. Public behavior is verifiable. Identity-graph inference is a black box you rent (and re-rent every year, on their terms).

Let me soften that. Identity graphs hold up fine at ad-targeting scale, where a few points of error wash out across a million impressions. One-to-one outbound is where the error rate bites, because a rep gets one shot per name.

How to spot fake person level intent data

This is the part I care about.

The pattern works like this:

1. The vendor buys account-level intent from a co-op.
2. An account surges on a topic.
3. The platform queries its own contact database for people at that account with titles matching your ICP.
4. It returns them ranked, with a confidence score.
5. The marketing says person-level.

Those contacts did not generate the signal. They just work there and hold the right job title.

Sometimes that is fine. Ranked guesses beat unranked guesses. But know what you bought. A ranking model, not an observation. Do not pay observation prices for it.

Five questions that expose the difference

1. Did this person generate the signal?Or does a model infer it?Watch how fast they answer. Inference is not a flaw. Dodging the question is.

2. What is the timestamp on the individual signal?Real person-level data carries a per-person event time. Inferred data carries an account-level surge window, usually weekly or monthly.

3. What does the confidence score measure?If the score reflects title fit rather than observed behavior, you are looking at a ranking dressed as a probability.

4. What is the deterministic versus probabilistic split?Ask for the percentage. A vendor running mostly probabilistic matching should say so.

5. What is your resolution rate by region?Ask for the US, the UK, Germany, and India separately. The spread will tell you more than any aggregate number.

What to do: Run all five questions before the pilot. Then ask for a sample export of 50 records with per-person timestamps included. A provider doing real observation will send it. A provider running inference will offer a demo instead.

Shortlisting helps here. The differences between intent data providers worth comparing come down to this distinction more than to topic taxonomies. Bombora's intent data co-op reports at account level by design and says so. ZoomInfo's intent data offering names a contact only when it can.

What person level intent data costs

The market splits into two price bands, and they barely overlap.

Band one: website visitor identification. First-party, your traffic only.

Band two: third-party and publisher intent. Broader reach, enterprise contracts.

Two things jump out of that list.

First, only TechTarget sells true contact-level resolution at the enterprise tier. It is also the priciest line on the sheet. That is not a coincidence.

Second, the cheap end of person-level costs under $200 a month because it only looks at your own traffic. Different product, different problem.

The compliance problem nobody puts in the deck

Person-level identification means naming a private individual from their browsing behavior. Regulators treat that differently than they treat company identification.

Why the EU effectively blocks it

The legal split is clean, and it explains the entire product map:

• Company-level identification usually rests on legitimate interest under GDPR Article 6(1)(f). You learn which business visited your business website. Low intrusion, corporate data, favorable balancing test.
• Person-level identification profiles an identifiable individual. Between GDPR and the ePrivacy rules on device storage, that pushes you toward opt-in consent.

Consent rates for that kind of tracking sit far too low to build a product on. So no major vendor attempts person-level identification in Europe. RB2B runs it on US traffic only. Others do the same and say less about it.

A vendor claiming person-level identification of EU visitors without consent should worry you, not impress you.

If your ICP is European, get your B2B consent management posture straight before you sign anything.

The US rules are looser, not absent

The US works as a notice-and-opt-out regime. Person-level identification is workable here, with conditions:

• Give notice at collection.
• Honor opt-outs, including Global Privacy Control signals.
• Put vendors on service-provider terms.

One detail teams miss: the CCPA B2B exemption expired on 1 January 2023. Work email addresses count as protected personal information now. Review the CCPA compliance requirements if you have not looked since then.

Where person level intent data breaks

Three places, and the pitch deck skips all of them.

Volume. Person-level surfaces far fewer records. A hundred named people you can act on beats two thousand accounts you cannot. That holds only if your team can work a hundred properly.

Early-stage awareness. When you need breadth for an ABM intent data program, account-level is the better buy. Person-level is an execution layer, not a targeting layer.

Signal accuracy. Topic-based third-party intent carries real false-positive rates. Somebody researching "CRM migration" might be a consultant, a job applicant, or a competitor doing the same homework you are. Person-level narrows that guess. It does not remove it.

How to turn a person level signal into a meeting

Getting a name is the easy half.

Speed decides the outcome

Intent has a half-life measured in days.

• Signals lose half their predictive value inside 30 to 45 days on a typical B2B software cycle.
• Average B2B lead response time runs above 40 hours. Under a quarter of companies respond within five minutes.
• Qualification odds collapse between the five-minute mark and the one-hour mark.

So a person-level signal sitting in a queue for two days is an account-level signal with extra steps.

Fix the plumbing before you fix the data:

• Feed the signal into your lead scoring model as a weighted input, not a separate alert stream.
• Apply decay. Weight the last 7 to 14 days heavily. Discount anything past 30 days. A proper intent data scoring model does this automatically.
• Let lead routing software push the alert to a named rep with an hours-based SLA.

Teams running a real signal based GTM motion track routing latency as a metric. Everyone else treats it as somebody else's job.

A name you cannot reach is a dead record

This is where the category quietly fails.

Your platform surfaces a VP of Revenue Operations researching CRM migration three days ago. Excellent. Her email bounces and you have no mobile number.

Intent freshness and contact freshness are separate problems. B2B data decay rates mean a record sourced eighteen months ago has a real chance of being wrong today. No amount of signal quality fixes a dead phone number.

Verified B2B direct dials matter more here than in ordinary outbound. Narrow window, one person, one shot. Voicemail on a wrong number burns it.

This is also why CRM data enrichment on a schedule beats buying a list once and hoping.

One signal is not a buying group

Even a perfect person-level signal names one person out of a group.

Gartner's research on the B2B buying journey describes buying as a looping process. Groups revisit the same buying jobs instead of moving in a straight line. The same research puts 99% of B2B purchases down to organizational change, not a single trigger moment.

Forrester's buying networks research goes further. Buyers now lean on a wider network: the internal buying group, plus outside parties they consult for advice. Partners. Customers. Influencers. AI agents.

Read that carefully. The person you identified may be a researcher, not the buyer. Committees send people to do the reading.

So treat the signal as an entry point, then map the rest:

• Mapping the B2B buying group turns one name into a route into the account.
• Platforms that identify buying group members automate the step reps skip when they get busy.
• Buying group intelligence tells you which of those people hold budget.

Then layer the signals. A person-level signal plus a funding round plus a new VP hire builds a far stronger case than any one alone. That is the argument for stacking multiple buying signals instead of chasing one score.

Build the message around what the person read. That is what cold email personalization that works depends on anyway.

Where SMARTe fits

I will be direct about what SMARTe does and does not do. This article has spent 2,500 words arguing that vendors blur exactly this line.

SMARTe does not run a cross-site pixel network. It does not sell identity-graph resolution either. That is not the product.

SMARTe closes the two gaps that make person-level intent unusable in practice. Knowing who else sits in the room. Reaching any of them.

Buying Signals.Bombora intent sits inside the platform, alongside funding events, leadership changes, and headcount growth. SMARTe scores them so you can rank accounts in active buying mode instead of eyeballing a topic list.

Buying group discovery.SMARTe's AI Agents auto-discover the buying group per account. One name from an intent alert becomes the full set of people who have to agree before anything gets signed.

Reachability, which is the real bottleneck.

• 289M+ verified B2B contacts
• 75%+ US mobile and direct dial coverage
• 86% of US decision-makers reachable with a verified email
• 50%+ global direct dial coverage across 200+ countries
• 66M+ company profiles and 64K+ technologies tracked
• Strong depth in LATAM and APAC, where US-centric providers run thin
• Real-time verification at the point of use (not a batch refresh you hope ran last quarter)

Compliance you can defend.SOC 2 Type II certified. GDPR aligned. CCPA compliant.

Pricing.Free with 10 credits a month. Pro at $25 a month with no per-seat cost. A team shares credits instead of buying licenses for people who prospect twice a week. Enterprise from $15,000 a year.

If you already buy intent and your reps keep telling you the contacts bounce, the intent is not your problem.

See how SMARTe finds verified mobile numbers in your target accounts.

Conclusion

Person-level intent solves a real problem. The teams selling it have every reason to overstate how much of it they solve.

The signal tells you someone cared enough to read something. It does not tell you they hold budget. Or that you can reach them. Or that the other nine people in the room agree.

Better data narrows the guesswork. It never removes it, and any vendor promising otherwise is selling you the deck rather than the product.