Intent data limits: when account-level is the right call
Published on 15 September, 2026 | Author: Digitalzone
Most arguments about account-level intent are really arguments about who is selling what. Here is the version without a product attached. Account-level intent data answers where you should spend. It does not answer who you should call. What follows is what a domain-level signal can and cannot support, plus a decision table you can take into your next budget conversation.
Buying rarely starts with a form. The median B2B site converts roughly 2-3% of visitors through forms, so most of the people researching you are invisible in your own systems.
What a domain match actually proves
Start with the mechanism, because the concession only means something once you can see how the signal is built. An account-level intent signal is the end of an inference chain, and every link is a probability rather than a fact.
Here is the chain. Someone reads an article, watches a video, or downloads a report on the web. A data co-op or publisher network observes that activity, matches it to an IP address or cookie pool, resolves the IP to a company domain, and reports it back to you as “this company is researching endpoint security.” Most co-op-based providers describe the approach the same way: aggregate engagement data across B2B publisher sites and map the surge back to target accounts. An honest description of a real pipeline, and also four inferences stacked on top of each other.
What the chain can support is genuine. If you need to know which target accounts are showing more research activity than usual this quarter, the domain is a valid unit of analysis. The signal tells you which companies to weight. That is a real answer to a real question.
What the chain cannot support is any statement about a named person. The signal resolves to a domain, not to a desk. It cannot tell you that the VP of Security is the one reading, or that the reader sits in procurement, or whether anyone researching has budget authority. The moment you need a name, the inference chain has run out. For how these signals are collected and where each type comes from, see our contact-signal-first intent data stack. For how domains get resolved to companies, see our explainer on identity resolution through the Journey Pixel.
Where the account is the right unit, and the cheaper one
Account-level intent is not a lesser signal. For a large share of real programs it is the correct choice, and paying for person-level resolution would buy precision you will not use. Four conditions where the account wins outright:
Your addressable market runs to thousands of accounts. You cannot run person-level plays against all of it, and you shouldn’t try. The job is to narrow the field before you spend on people. Account signals do that well and cheaply.
The decision is media allocation, not outreach. Nobody calls a person off a display impression. You’re buying reach against companies, and account signals price that reach.
The buying committee is small. When two or three people make the call, any reasonable contact is a sensible entry point. The committee is small enough that you’ll reach the right room quickly regardless.
The motion is inbound-led. If most of your pipeline comes to you and sales mostly responds, account signals are enough to prioritize follow-up and route the right rep. The buyer will surface themselves.
Here’s a worked example. Say you have 4,000 target accounts and a display budget of $200,000 for the quarter. Account intent tells you 600 are surging. You concentrate 70% of the budget on those 600 and hold the rest for the long tail. You never needed a single name to make that allocation better. Buying person-level data on 4,000 accounts to inform a media decision would be spending on precision the media buy cannot use.
That is the honest case for account-level, and it holds. Most programs should run this way for the top of the funnel.
The moment the inference breaks
The failure is not gradual. It happens at one specific decision: which person do we contact. Up to that point the account signal is doing useful work. At that point it goes silent, because it was never built to answer the question.
Forrester found that identifying specific contacts to target within accounts demonstrating intent was the top execution challenge for intent data practitioners. Not a data-quality complaint about one vendor. The whole category runs into the same wall at the same place: the signal identifies the account but cannot tell you who inside it to reach.
The scale is easy to underestimate. Forrester’s Buyers’ Journey Survey, 2025 reports that 73% of purchases involve three or more departments, with an average of 13 people inside the buyer’s organization and nine from outside shaping the decision. That is 22 people in and around one account. An account-level surge tells you the building is warm. It cannot point you to a floor. Forrester calls this buying group blindness: systems that see accounts but not the humans doing the buying.
Ask any of those 22 people who started the research and who holds the most sway, and most will say they did. In our 2026 buyer survey (n=1,500), a majority at every job level believed it initiated vendor research and carried the most influence. Gen Z buyers were the most convinced at 64%, compared to 24% of Boomers who said the same. Several people inside one account each believe they are the one who matters. An account-level signal cannot adjudicate between them. Neither can a rep guessing from a title list. Accounts don’t buy. People do. The account signal does not know which people.
The decision table
Five variables decide whether the account is the right unit or whether you need the person. Read each row as a testable statement, not a lean. Where your program disagrees with a cell, that’s the useful part.
| Tier | Signal threshold | SDR action | Timing | Follow-up |
|---|---|---|---|---|
| Hot | Verified buyer-role contact, 2+ signals in 14 days | Assigned SDR calls; unowned goes to round-robin | 24-hour first touch | Multi-channel sequence; disposition logged |
| Nurture | Buyer-role contact, 1 signal in 30 days | Enroll in behavior-referenced email sequence | Within 3 business days | Promote to hot if a second signal fires |
| Monitor | Signal older than 30 days or below depth floor | No direct outreach; watch for re-engagement | None | Re-tier when a fresh signal fires |
Most real programs sit in more than one row at once, and the rows can conflict. A company can have a broad TAM (account wins) and a $250K average deal with an eight-person committee (person wins). When rows disagree, the tie-breaker is the sales motion, because motion decides who acts on the signal. If a human is reaching out to a named individual, you are on the contact side of the line no matter how broad the TAM is. If no human reaches out, the account is enough. Deal size is the second tie-breaker: the more a deal is worth, the faster the math favors resolving the person.
What contact-level resolution changes
Once the unit of analysis becomes a person, the program does different things.
Routing gets a named recipient instead of a company and a shrug. Sequencing differs by role, so the security architect and the CFO get different first messages. Suppression works, because you can hold back people already in a deal instead of hitting the whole domain again. And the sales conversation starts from behavior, not from a topic: not “we saw your company researching,” but “you looked at the integration docs twice last week.”
The difference shows up in our campaign data. Accounts engaged by two or more tactics surface 72× more buying-committee members than single-tactic accounts, and convert a qualified lead at over 10× the single-tactic rate (100% vs. 9.6%). That gap is what contact-level resolution, combined with multi-channel reach, produces.
This is the side of the line Digitalzone works on. We read signals across tactics to reach the specific people inside an account, not the account as a whole. Not instead of your account platform. On the other side of the same boundary.
Deciding without a product in the room
Account-level intent tells you where to spend. It does not tell you who to call. Nearly every argument you’ll read about intent is really an argument about which side of that boundary the writer sells on.
Decide by the table, not by the pitch. If your motion is inbound, your TAM is broad, your committees are small, and your channel is media, the account is the right unit and the cheaper one. Stay there. If your motion is outbound, your deals are large, and your committees are crowded, the account gets you to the building and leaves you at the door.
For teams on the contact side of the table, talk to us about contact-level lead generation. For teams running broad account programs, our demand generation services start from where to spend. Most programs need both.
FAQs
What is the difference between account-level and contact-level intent data?
Account-level resolves activity to a company domain; contact-level resolves it to a named person. The first answers where to spend, the second answers who to call.
When is account-level intent data the right choice?
When your TAM is broad, the decision is media allocation, the buying committee is small, or the motion is inbound-led. Use the decision table above to test your own program.
Why can’t intent data tell you exactly who is researching?
The signal is an inference chain that ends at a domain. It never touches a named individual, which is why Forrester found that identifying specific contacts inside intent-showing accounts is the top execution challenge for practitioners.
Does account-level intent still matter if the committee is large?
For awareness, yes. For action, no. With 22 people shaping a typical purchase, the wrong first contact stalls the deal. You still need contact-level resolution to route correctly.