How to unmask fake contact intent: a 10-question vetting framework

Every intent data vendor sells contact-level precision now. Few will describe the methodology that earns the claim. 

You’ve heard the pitch. “We identify intent at the contact level.” It sounds like the thing you actually need: not “someone at Acme is researching,” but “the VP of IT at Acme opened your comparison page on Tuesday.” 

The problem is that the line gets used the same way whether the vendor fingerprints individual behavior or relabels an account-level topic surge and attaches a generic contact to it. 

That gap is what this guide exists to close. Below is a four-dimension evaluation framework and a 10-question vendor scorecard that any contact-level intent data provider should be able to answer on a single call. 

Forrester puts it plainly: the term “intent data” gets used to market everything from specific purchase behavior to vague topic interactions, so marketers should understand the collection methodologies their provider uses. Vendors who can answer all ten are running a contact-level methodology. Vendors who dodge several are selling account-level data with contact-level marketing. 

Dimension 1: how a vendor defines a behavioral signal tells you everything. 

Start here, because every other answer depends on it. Ask the question directly: how do you define a behavioral signal at the contact level? 

A strong answer describes individual-level web activity, fingerprinted to a specific device or an authenticated session, then mapped to a contact record with verified title and company. That’s a person doing something, tied to a name you can hand to sales. The signal has an owner. 

A weak answer sounds like this: “Our platform identifies intent from account-level topic surge.” That’s account-level data, regardless of the label on the box. Topic surge tells you a company is warm. It doesn’t tell you who. 

When a vendor blurs that line, ask for a technical data sheet that documents the signal collection methodology. If one doesn’t exist, the methodology probably doesn’t either. 

This is also why the probabilistic-versus-deterministic question matters. Probabilistic matching guesses identity from device patterns; deterministic matching ties a signal to a verified record. Both can say “contact-level.” Only one survives a sales rep checking the name. 

Dimension 2: a TAL match test separates real data from a good demo. 

Canned demos always look great. Your target account list is the test that matters. 

Before you sign anything, run a match test. Give the vendor 50 accounts from your TAL and ask a specific question: which verified buyer-role contacts do you have, at these accounts, with active behavioral signals in the last 30 days? Then read what comes back. 

In our experience, a good result is a match rate around 70% or higher. Anything below 40% usually means the vendor is working from a thin contact database or padding account-level signals with unverified names. 

Each contact should carry a full title, company, the behavioral signal type, and a recency date. That’s a list sales can work on Monday. A red flag is a vendor who won’t provide a sample at all, or who returns account-level signals stapled to generic contacts with no specific behavioral evidence behind them. 

Forrester recommends the same move, and recommends scoping it tightly: request sample data and focus your trial on accounts you already know well, so you can judge accuracy against what your own team has already learned. You’re not testing whether the data looks impressive. You’re testing whether it’s right where you can check it. 

Dimension 3: a 90-day signal window is a stale signal window. 

Intent decays fast. Forrester flags this directly, calling data decay one of the biggest intent data mistakes in B2B: if you store historical buying signals without incorporating decay, it will soon appear that every company in your database is demonstrating intent. 

A contact who researched your category last quarter may have already bought, already stalled, or already forgotten. Recency isn’t a detail; it’s the difference between a live signal and a memory. 

Ask one question: what is the behavioral recency window for a contact to be flagged as in-market? In our experience, a 21-to-30-day window is a good answer. Signals older than 30 days should be deprioritized; anything past 90 days is stale enough to create false confidence in your pipeline. A 90-day window, or longer, means you’re working signals that may have gone cold weeks ago. 

Then ask the follow-up that catches the rest. How often does the data refresh, and do you suppress a contact when signal activity drops below threshold? Or do contacts stay flagged as hot indefinitely? A platform that never cools a contact down isn’t measuring intent. It’s maintaining a list. 

Dimension 4: CRM and MAP integration is where good data goes to die. 

Contact-level intent is only worth what reaches your CRM and MAP cleanly. The best data in the world generates nothing if it lands as a quarterly CSV nobody opens. 

So ask for a live demo of the integration, not a slide about it. Show me what a contact record looks like in Salesforce or HubSpot after it syncs from your platform. 

A good answer puts the signal type, the recency date, the behavioral event details, and the ICP role classification right on the record, where a rep will see it. A red flag is “we export to CSV” or “we sync to your CRM weekly.” 

Weekly isn’t fast enough for a 30-day signal. By the time the batch lands, a third of the window is gone. 

The 10-question vendor scorecard. 

Here is the full scorecard. For each question, the specific ask, the answer that signals a real contact-level methodology, and the answer that signals account-level data in a contact-level costume. 

1. Signal definition. “What did this specific person do, and how did you tie it to them?”  

Good: individual web activity fingerprinted to a device or authenticated session, mapped to a verified contact. Red flag: “account-level topic surge.” 

2. Identity matching. “Is identity matched deterministically or probabilistically, and how do you validate it?”  

Good: deterministic matching against verified records, with a documented validation step.  

Red flag: no answer, or “proprietary” used as a wall. 

3. Recency window. “How recent does a signal have to be for you to call someone in-market?”  

Good: 21 to 30 days. Red flag: 90 days or longer. 

4. Data refresh. “How often does the data refresh, and do contacts ever cool off?”  

Good: frequent refresh with suppression below threshold.  

Red flag: contacts stay hot indefinitely. 

5. CRM integration. “Show me what a synced contact looks like in Salesforce right now.”  

Good: signal type, recency, event details, and role classification on the record.  

Red flag: CSV export or weekly batch. 

6. TAL match test. “Can you run a 50-account match test before we sign?”  

Good: a sample with 70%+ match and full contact detail.  

Red flag: refusal, or account-level signals on generic contacts. 

7. Data provenance. “Where does your contact database come from, and where do the signals originate?”  

Good: named first-party sources and a clear collection method.  

Red flag: vague third-party aggregation with no origin story. 

8. False positives. “How do you keep from flagging people who aren’t actually in-market?”  

Good: thresholds, signal corroboration, and suppression rules.  

Red flag: no concept of a false positive. 

9. Pipeline attribution. “Can you connect a flagged contact to a closed deal?”  

Good: attribution that marketing and sales can both see.  

Red flag: engagement metrics with no revenue line. 

10. Data quality support. “When the data is wrong, who do I call and how fast does it get fixed?”  

Good: a named process and a human to call.  

Red flag: a ticket queue and a shrug. 

Score it simply. Pass means the vendor answered with specifics. Fail means they reached for the brochure. 

What separates real contact-level intent from the claim. 

Vendors who pass all ten questions are operating from a contact-level methodology. They can show you the signal, name the source, put it on the record, and tie it to pipeline. Vendors who can’t, or won’t, answer several of them are selling account-level intent data with contact-level marketing. 

This scorecard sits one layer above the model itself. If you want the methodology these questions test against, our breakdown of how our contact-level intent model works shows what a real signal looks like end to end. 

The Digitalzone Data Cloud is built on a proprietary ID graph and first-party data from our editorial brands, resolving identity across the full lead journey rather than guessing at it. The proof is in the numbers we’ll put in writing: 320M+ B2B decision-makers in our owned database, 6x campaign engagement lift over industry benchmarks, and 90% client retention because the data actually holds up in the field. 

Run the scorecard on us. Then run it on everyone else. 

Bring us your TAL and we’ll show you the contacts, the signals, and the dates. Demand better from your intent data. 

Frequently asked questions.

What is the difference between account-level and contact-level intent data? Account-level intent tells you a company is researching your category. Contact-level intent tells you which person is, with a name, a title, and a behavioral signal. One narrows your list; the other tells you who to call. Our breakdown of how contact-level intent actually works goes deeper into the methodology. 

How do I test whether a vendor’s contact-level intent data is real? Run a TAL match test. Give them 50 target accounts and ask for verified buyer-role contacts with behavioral signals from the last 30 days. Real contact-level data comes back with names, titles, signal types, and recency dates. Bring us your TAL and we’ll run one with you. 

What is a good behavioral recency window for intent signals? 21 to 30 days. Intent decays fast, so a contact flagged on a 90-day window may have already bought or moved on. Ask whether the vendor suppresses contacts when signal activity drops. The Digitalzone Data Cloud is built to refresh signals at this cadence. 

Why does CRM integration matter so much for intent data? Because data that doesn’t reach your reps cleanly generates no pipeline. Ask to see a synced contact record in Salesforce or HubSpot, with the signal type, recency date, and role classification on it. CSV exports and weekly batches are too slow for a 30-day signal. See how Programmatic Nurture™ connects intent signals to your activation stack.