Data Enrichment
What is intent data? Types, buying signals and when it's worth paying for
B2B intent data explained: first-, second- and third-party intent, how it differs from buying signals, and how to combine it with fit to prioritize accounts.
Founder, OretonGTM · ·13 min read
Intent data is one of the most hyped categories in B2B sales. Vendors promise to show you “who’s in market right now”, and founders buy it hoping it will fix a weak pipeline. Sometimes it helps. Often it becomes an expensive dashboard nobody acts on.
The problem is rarely the data itself. It’s that “intent data” gets used to mean three very different things, it gets confused with buying signals in general, and it gets used on its own instead of alongside fit. This guide covers what each type of intent data actually measures, how it differs from buying signals, and how to combine intent, signals and fit into a system that tells your team who to contact and why now.
What intent data actually is
Intent data is a record of research behavior. It tracks activity like content consumption, repeat website visits, pricing-page views and review-site comparisons, then uses that activity to estimate whether an account is actively researching a solution.
Two words in that definition matter:
- Record of behavior. Intent data doesn’t tell you what a company plans to do. It tells you what people associated with that company did: read articles, visited pages, compared vendors.
- Estimate. Most intent data, especially third-party, is inferred. A vendor sees activity, attributes it to a company and decides whether it counts as “intent”. Every step in that chain has error.
Many third-party platforms compare an account’s current activity against its own historical baseline instead of using one fixed threshold. That’s sensible: a 20,000-person enterprise will always generate more research activity than a 15-person startup, so a relative change (“this account is reading about cold email deliverability three times more than usual”) is more informative than an absolute number.
Intent data comes from three sources, and each one tells you something different.
First-party intent data
First-party intent comes from your own properties and systems:
- Website visits, especially pricing, comparison, integration and case study pages
- Content downloads and webinar registrations
- Demo requests and free-trial or sign-up activity
- Email engagement: link clicks and, most importantly, replies
- Product usage, if you have a free tier or trial
- CRM history: past opportunities, closed-lost deals, churned customers
This is the highest-fidelity intent you’ll ever get, because the account is interacting with your company, not just your category. A head of sales at an ICP account who visits your pricing page twice in a week is a better lead than any topic surge a vendor can sell you.
The trade-off is reach. First-party intent only covers accounts that have already found you.
The other common problem is that first-party intent sits unused. Website visits live in analytics, email clicks live in the sequencer, product sign-ups live in the database, and nobody connects them to the CRM record of the account. Fixing that is mostly an automation problem: identify the company behind the activity, match it to the account in your CRM, and alert the owner fast. I cover the response-time side of this in speed to lead and the alerting side in Slack alerts for sales.
My take: Before you buy third-party intent, make sure you act on your own. Most startups I work with have their best intent signals sitting unnoticed in Google Analytics and their sequencer.
Second-party intent data
Second-party intent is another company’s first-party data, shared with you. The common sources are:
- Review sites and software directories, which can show that an account viewed your profile, compared you with a competitor or browsed your category
- Marketplaces and integration directories, such as app stores of platforms you integrate with
- Publisher and community partnerships, where a media site or community shares engagement with sponsored content
This data carries more commercial context than a generic topic surge. “Viewed a comparison of you vs. a competitor” is a much stronger statement than “read three articles about sales automation”. Its value depends heavily on two things: how much relevant activity the platform actually gets in your category, and how accurately it identifies the company behind a visit.
Second-party intent is worth testing if your buyers genuinely use review sites to shortlist vendors. In many technical and developer categories they do; in many services categories they don’t.
Third-party intent data
Third-party intent is behavioral data aggregated across publisher networks and websites outside your own properties. Vendors track content consumption across thousands of B2B sites, attribute that activity to companies (usually through IP and other identity signals), map the content to topics and flag accounts whose research on a topic is above their normal baseline.
Its strength is reach: it can surface accounts that have never interacted with you. Its weaknesses are the ones you’d expect from an inferred, aggregated dataset:
- Attribution error. Remote work and VPNs make company identification from traffic harder than it used to be.
- Topic ambiguity. Someone researching “email deliverability” could be a buyer, a student, a marketer writing a blog post or your competitor.
- No person attached. You usually learn which company is researching, not who. You still need to figure out which buyer to contact.
- Research isn’t budget. An account researching your category is not necessarily researching you, working with an approved budget or ready to talk to sales.
None of this makes third-party intent useless. It makes it a prioritization input, not a trigger by itself.
Intent data vs. buying signals
The terms get used as synonyms constantly, which causes real problems in how teams build their targeting.
A buying signal is the broader category: any observable event or behavior tied to a specific account that gives you a reason to reach out now. Intent data is one type of buying signal. Others include:
- Hiring signals: a company posting jobs that your product supports or replaces (see hiring signals)
- Funding signals: a new round that brings budget and growth targets (see funding signals)
- People signals: a new VP or head of a function, or a former customer champion who moved to a new company (see job change signals)
- Technology signals: adopting, dropping or outgrowing a tool in your ecosystem (see technographic targeting)
- Company events: expansion to a new market, a new product line, a merger, a compliance deadline
The key difference is the type of evidence:
| Third-party intent data | Event-based buying signals | |
|---|---|---|
| What it measures | Inferred research interest | A confirmed, dated event |
| Example | “Account X is surging on ‘sales automation’” | “Account X hired a VP of Sales on Oct 3” |
| Verifiable? | No, it’s a model’s estimate | Yes, you can check the source |
| Tells you who to contact? | Usually not | Often (the new hire, the hiring manager) |
| Gives you a message angle? | Weakly | Strongly (“new VPs usually rebuild X in the first 90 days”) |
| Freshness window | A few weeks | Weeks to a few months |
| Coverage | Wide, including accounts you don’t know | Depends on the signal; public signals are wide |
Why the confusion causes real problems
Treating “intent” and “signal” as the same thing leads to three common mistakes.
1. Buying a third-party intent tool when you needed signals. A founder wants “to know who’s in market”, buys an intent platform, and gets a weekly list of surging accounts with no contact, no context and no message angle. The SDRs don’t know what to do with it, so they send the same generic sequence they always send. Meanwhile, the company that just hired its first sales leader (a verifiable, contactable, message-ready signal) goes unnoticed.
2. Over-trusting inferred data. Intent scores look precise (“84/100 on Sales Engagement”). Teams start treating them like facts, routing accounts to reps and building messaging around them. When reps open with “I noticed you’re looking into…”, they’re wrong often enough to sound creepy and inaccurate at the same time.
3. Ignoring fit. An account surging on your topic but outside your ICP is still outside your ICP. A 10-person agency researching “enterprise sales automation” will not become a good customer because its research spiked. Intent and signals tell you when. Fit tells you whether.
Your outbound can generate its own signals
One underused idea: a well-designed outbound campaign creates first-party intent.
Every send is a small experiment. Opens are unreliable these days, but clicks, replies and website visits from your emails are real behavior from accounts you chose deliberately. A few ways to use that:
- Offer-based segmentation. Send two or three different offers to a segment (for example, a benchmark report, an audit and a direct meeting request) and track which accounts engage with which. The engagement itself tells you what that account cares about.
- Soft “not now” replies. A reply like “not this quarter, we’re mid-migration” is a dated, specific signal. Log it, set a follow-up date and treat it as warm when the date arrives.
- Website visits after a send. If you can identify companies visiting your site, a visit to the pricing page within days of your email is one of the strongest signals you’ll see. Route it to the account owner immediately.
- Referrals. “I’m not the right person, talk to Ana” tells you the buying committee structure and gives you a warm introduction.
This is why I’m skeptical of teams that buy intent data before their own outbound is tracked properly. Your campaigns are already producing intent. You just need to capture it. Classifying replies automatically is usually the first step.
Why neither intent nor signals is enough on its own
Intent without fit gets you busy accounts that will never buy. Signals without fit get you timely conversations with the wrong companies. Fit without signals or intent gets you the right accounts at random moments, which is how most outbound runs and why reply rates stay low.
You need all three, and they answer different questions:
| Input | Question it answers | Examples |
|---|---|---|
| Fit | Could this account become a good customer? | Industry, size, geography, business model, tech stack |
| Signals | Is there a reason to reach out now? | Funding, hiring, new leader, tech change, expansion |
| Intent | Is the account researching this problem? | Pricing-page visits, review-site comparisons, topic surges |
A framework for combining fit, signals and intent
Here’s the checklist I use to decide whether an account deserves outreach this week. An account should pass all five:
- Fit. Is it inside the ICP? If not, stop here, no matter how strong the signal.
- Relevance. Is the signal connected to the problem you solve? A funding round is relevant to a recruiting tool; it’s barely relevant to a payroll compliance product unless headcount is about to jump.
- Strength. How strong is the evidence? A confirmed event or first-party action beats an inferred surge. Two independent signals beat one.
- Recency. Is the signal still fresh? A demo-page visit from yesterday is hot; a surge from two months ago is stale.
- Actionability. Do you know who to contact and what to say? If the answer is “no idea”, the signal isn’t actionable yet. You need research first.
Each data type helps with different parts of the checklist:
- Firmographic and technographic data answers fit (see firmographic data).
- Event signals answer relevance, recency and actionability: they’re dated and come with a natural message angle.
- Intent data answers strength (when stacked with other evidence) and helps you prioritize within a large fit-qualified list.
How to build an intent and signal prioritization system
This is the practical version, the one I’d build for a seed or Series A startup.
Step 1: Start from your wins
Pull your last 15 to 20 closed-won deals. For each one, write down what was happening at the account in the 90 days before the first conversation. Did they just raise? Hire a sales leader? Switch CRMs? Visit your site several times? You’re looking for patterns that repeat across several wins, not one-offs.
These patterns become your candidate signals. They’re much more reliable than a vendor’s generic list of “buying signals”, because they come from your own buyers.
Step 2: Define the ICP filters first
Write down the fit criteria that gate everything else: industries, size range, geographies, business models, must-have technologies. I go through this in detail in the ICP template. Every signal and intent source should only be applied inside this filter.
Step 3: Instrument first-party intent
Before adding anything external, make sure these reach the CRM, tied to the account:
- Visits to pricing, comparison and integration pages (with company identification)
- Email clicks and every reply, classified
- Sign-ups, trials and demo requests from ICP accounts
Then set up an alert for the high-value ones. A pricing-page visit from an ICP account should reach the account owner the same day.
Step 4: Add two or three public signals
Pick the signals from Step 1 that you can detect reliably. For most B2B startups that means some combination of hiring, funding, leadership changes and tech-stack changes. Tools like Clay make it relatively easy to monitor these across a target account list (see the Clay tutorial).
Step 5: Score and decay
Combine fit and signals into a simple score. Keep it simple enough that a rep can explain why an account is at the top. Give each signal a weight and a shelf life: a pricing-page visit might be worth a lot for seven days and nothing after 30; a funding round might hold value for 90 days. I go deeper on structure and weights in the account scoring model guide.
Step 6: Only then, test third-party intent
If your TAM is large enough that fit plus signals still leaves you with more accounts than your team can work, run a time-boxed test of a third-party intent source:
- Split a fit-qualified list into a group with intent and one without.
- Run the same sequence to both.
- Compare positive reply rate and meetings booked after four to six weeks.
If the intent group doesn’t clearly outperform, you have your answer, and you saved yourself an annual contract.
How to use intent in your messaging (without being creepy)
Intent should shape your targeting and your angle, not your opening line.
Don’t write: “I saw your team has been researching sales automation tools.”
Do write: an email that leads with the problem the intent suggests. If an account is surging on deliverability topics and just hired two SDRs, a good opener is about the operational pain of ramping SDRs on new domains, not about what you “saw”.
The same applies to event signals, with one difference: event signals are public, so referencing them is normal. “Congrats on the Series B” is fine (though overused). “Teams usually rebuild outbound in the first quarter after a raise; here’s what we see break first” is better, because it turns the signal into a relevant point of view. More on this in how to write a cold email.
Common intent data mistakes
- Buying third-party intent before acting on first-party intent. Your own pricing-page visitors are better leads than any surge report.
- Using intent without fit. A surge outside your ICP is noise.
- Treating scores as facts. Intent is a probability. Build workflows that assume it’s wrong some of the time.
- No decay. Old signals keep stale accounts at the top of the list and burn rep time.
- Quoting intent to prospects. It’s invasive, and it’s often inaccurate.
- No feedback loop. If you never compare reply and meeting rates between signal-driven and non-signal campaigns, you never learn which signals actually work for you.
When third-party intent makes sense
Third-party intent is worth paying for when most of these are true:
- Your TAM is large (thousands of fit-qualified accounts) and you need to prioritize within it.
- Your category has clear research topics that buyers look up before purchasing.
- You already act on first-party intent and public signals, quickly and consistently.
- You can combine it with fit scoring and route it to someone who will act within days.
- You have the volume to measure whether it works.
If you’re a seed-stage startup with a TAM of a few hundred accounts, you’ll usually get more from monitoring those accounts directly, tracking your own outbound engagement and acting fast on public signals. If you want to go deeper on the signal side, start with signal-based outbound.
Frequently asked questions
What is B2B intent data?
Data indicating that a company or person is researching topics related to your product, such as content consumption, review-site comparisons or repeat visits to your pricing page. It suggests, but does not prove, that an account is in a buying cycle.
What is the difference between intent data and a buying signal?
A buying signal is any observable event that gives you a reason to reach out, like a funding round, a new VP hire or a job post. Intent data is one category of buying signal that measures inferred research interest. Many signals, like funding or hiring, are confirmed events; most third-party intent is a probabilistic estimate.
What is the difference between first-party and third-party intent data?
First-party intent comes from your own properties: website visits, email clicks, product sign-ups and replies. Third-party intent is collected by vendors from activity on other websites and publisher networks. First-party is more accurate; third-party covers accounts that have never interacted with you.
Is intent data accurate?
First-party intent is usually reliable because it's tied to your own systems. Third-party intent is directional: it depends on how the vendor identifies the company behind anonymous traffic and how it defines a surge. Treat it as a prioritization input, never as proof of a buying cycle.
Is intent data worth it for a small team?
Usually not as a first purchase. With a small TAM you can monitor accounts directly, and first-party intent plus public signals (hiring, funding, leadership changes) cover most of the value. Third-party intent starts to pay off when your TAM is large enough that you need help choosing which accounts to work this month.
Should I mention intent data in my cold email?
No. Saying 'I saw you were researching X' feels invasive and is often wrong. Use intent to decide who to contact and which angle to lead with, then write the email around a problem the account is likely to have.
How long does an intent signal stay relevant?
It depends on the signal. A pricing-page visit or a demo-page visit is hot for days. A third-party topic surge is relevant for a few weeks. Hiring and funding signals stay useful for one to three months. Build a decay into your scoring so old signals stop pushing accounts to the top.