Picture the intent data review that plays out at a lot of Series B companies. A long report from the vendor. Glowing graphs. Heat maps. A score for every account in the ICP. A five-figure annual contract behind it.
Then someone asks the question that matters: "How many meetings did this book last quarter?"
The honest answer is usually a handful, out of the thousand-odd accounts the platform flagged as high intent. And some of those accounts were already in pipeline before the intent score lit up.
That is the story of most B2B intent data in 2026. Big spend, beautiful dashboards, almost no impact on pipeline. The vendors sell signal. What they ship is noise dressed up as signal. And the SDR team that is supposed to act on it ends up calling the same accounts they would have called anyway.
I have been buying, building, and dismantling intent data programs for years, across more than 30 engagements as a RevOps consultant and now as Founding GTM Engineer at Peec AI. The pattern is the same. Buy a vendor. Get a list. Watch reps ignore it. Cancel the vendor. Buy a new one. Repeat.
This post is the one I wish someone had given me eight years ago. What B2B intent data actually is, what works, what is a waste of money, and the four signals I would build a program around if I were starting from zero next Monday.
What B2B intent data is, plainly
B2B intent data is anything that suggests a buyer at a target account is moving closer to a purchase decision. That is the only definition that matters.
Vendors will tell you it is third-party content consumption patterns, aggregated from publisher networks, weighted by topical relevance, scored against your ICP. That is not wrong. It is just one slice of the picture, and the worst slice.
I split intent data into three buckets:
First-party intent. Things happening on your own properties. Website visits, demo requests, pricing page views, content downloads, opens on a specific sequence, replies that contain a question, second-touch behavior in a 30-day window. You own this data. It is the most reliable signal you will ever have.
Second-party intent. Things happening in places where you have a structured relationship. G2 page views and category browses, a review left on Capterra, a comparison search that includes your brand, a search inside a partner's marketplace. You do not own the platform, but the data is intentional and self-identified.
Third-party intent. What most vendors sell. Content consumption signals from publisher networks. Account-level browsing on sites that participate in a bidstream or cookie pool. Aggregated topic spikes. This is the noisiest tier and the one most teams pay too much for.
The part nobody wants to talk about is the size of the gap between those three tiers. First-party converts to meetings at a rate that makes outbound feel easy. Second-party converts well enough to be worth working. Third-party, on its own, converts at a rate low enough that most teams cannot tell it apart from a cold list. That ordering is consistent across SaaS, fintech, devtools and security, and it matters far more than any single conversion rate you will be quoted.
Measure the three tiers in your own CRM before you accept any vendor's benchmark. The ordering is the point. The rates are yours to find.
That ordering is also why an expensive third-party intent contract can book fewer meetings than a free Google Search Console export of your own branded queries.
Why most third-party intent data fails
The vendor pitch is appealing. They scan 3 billion web pages and 200 million B2B identifiers. They tell you which accounts are "researching" your category. You get a list of in-market companies before your competitors do.
Here is what is actually happening behind the curtain.
A user on an account's network reads a Forrester report. The vendor's bidstream picks up the visit. They map the IP to an account using a reverse-DNS database that is right about 35 to 60 percent of the time on enterprise IPs and far worse on remote workers. They classify the page topic using NLP. They roll the visit into an account-level score. They flag the account as "high intent on data infrastructure."
But the person reading that report was a researcher. Or a procurement analyst doing diligence on a vendor you do not compete with. Or someone clicking through from a tweet. Or a marketing intern doing competitor research. Or someone whose home network shares an IP with the account.
The vendor does not know which. The vendor cannot know. And the account score they push to your CRM treats all of those visits the same.
There is a cheap way to find out whether your own vendor's score means anything. Take a few hundred accounts the platform calls "surging" on your category and a matched set it calls low intent. Same ICP filters, same sequence, same reps, same two-week window. Then count meetings.
Most teams never run that test, which is precisely why the renewal keeps getting signed. Run it before yours comes up. If the two cohorts land in the same place, you are paying for a score that is not telling you anything you could not get for free.
Third-party intent without a tie to behavior on your site is closer to astrology than science.
The best third-party signals work as ICP filters, not buying triggers. Use them to narrow the universe, not to predict who is ready to buy.
The four signals that actually work
If I were rebuilding a B2B intent program from scratch tomorrow, this is the stack. Four signals. In this priority order.
1. Second-touch behavior on your own site
The single most predictive signal in B2B is repeat behavior. One pricing page view in a six-month window is noise. Two visits from the same account in 30 days, especially when the second visit lands on a different page than the first, is signal.
Set up the tracking with HubSpot or RB2B for de-anonymized companies. Then build a rule. Two visits within 30 days, at least one of them on a high-intent page (pricing, demo, integrations, security), triggers an SDR task with the account context.
This is the most dependable signal most teams already own and do not act on. The reply rates are not just higher. The conversations are completely different. The rep is not introducing the company. They are following up on observed interest.
2. Job change signals at target accounts
When a champion at your customer base changes jobs, three things happen. Their old account loses a champion. Their new account just got one. And there is a 90-day window where they have political capital and budget freedom that closes fast.
Track this with Champify, UserGems, or a Clay table feeding off LinkedIn job change data. Set the rule: customer contact or known champion changes jobs, new company is in ICP, alert the AE within 24 hours.
This is the highest converting signal I have seen in B2B, and it is not close. You are reaching someone who already knows the product, already trusts you, and has just walked into a budget they control. Deals sourced this way tend to move through the pipeline faster than anything else you run.
3. Hiring signals tied to your problem space
When a company posts a job that mentions your category, your competitor, or your problem space, they have just told you what they are buying. A "RevOps Manager" job that asks for HubSpot, Clay, and Apollo experience tells you they have or want that stack. A "Director of Demand Generation" post that mentions account-based marketing tells you they are building an ABM motion.
I pull this with Theirstack, built it manually for years with a Clay table off LinkedIn job APIs, and now use a custom n8n workflow for clients with specific stacks.
The signal is not "they will buy in 30 days." The signal is "they have committed budget and headcount to this problem space." That is the window when they are open to a conversation.
It converts best when the opener references the role and the stack mention directly, and worst when it is used as a generic congratulations note. The job post is the research. Use it.
4. Funding announcements paired with hiring
A Series B does not mean a company is ready to buy your tool. It means they have capital. But a Series B paired with hiring signals in the function you sell to is a different signal.
I run this as a two-step filter. Pull funding announcements from Crunchbase or Harmonic. Cross-reference against open roles posted in the last 60 days that touch the buyer persona. The intersection is small, qualified, and active.
This is the setup I run for my own outbound at Peec AI. A Clay table builds the intersection each month, a Smartlead sequence works it, and one person can hold the whole motion because the list stays small on purpose.
How to actually deliver intent data to reps
The second failure mode after buying bad intent data is delivering good intent data badly. A weekly CSV in someone's inbox is not delivery. A dashboard in HubSpot that nobody opens is not delivery. A Slack channel with 80 alerts a day is not delivery.
Here is the system that works, regardless of which signals you pick.
The signal goes into the CRM as a record, not a notification. Create a custom object in HubSpot or Salesforce called "Intent Event" or "Signal." Each signal becomes a record tied to the account and contact. The rep can see the full history on the account page. No CSV. No Slack scroll.
The signal triggers a task with context. Not "this account has intent." Bad. Try "Sarah at Acme viewed pricing twice in 4 days, last visit 2 hours ago, also visited the integrations page." Good. The task includes the page URLs, time stamps, and what to say in the opener.
The signal has a freshness window. A pricing page visit is hot for 48 hours. After that it is cold. Tasks expire. Closed tasks get logged for forecasting. Open tasks get reassigned if not actioned in 24 hours.
The signal has an outcome tracker. Did the rep reach out? Did they book a meeting? Did the meeting turn into pipeline? Every signal closes the loop. After 90 days you know which signals convert and you double down on those.
We build this as a custom workflow inside HubSpot for clients running on Sales Hub. For teams with bigger stacks we wire it through n8n with a custom data layer that handles deduplication, scoring, and routing. The pattern is the same. The CRM is the system of record. Reps work the queue. Marketing tunes the rules.
Which tool sends the outreach once the signal fires is a separate decision from which signals you trust, and the two get bundled together far too often. If you are picking that layer now, which account-based selling platform runs the queue is the question that post answers, split by signal, sequencing, and advertising rather than sold as one stack.
How to score intent without overengineering it
Most scoring models are built by RevOps people who love spreadsheets and lose the trust of sales in three weeks. The reps see a score of 87 and a score of 91 and ask why one is higher. RevOps cannot answer because the model has 14 weighted inputs nobody can hold in their head.
Keep it simple. Three tiers. Hot, warm, cold. Each tier has a clear rule a rep can repeat.
Hot. First-party second-touch within 30 days, or job change of a known champion within 14 days, or two signals from different sources within 7 days. Action: SDR or AE outreach within 24 hours.
Warm. Hiring signal matching ICP role, or funding plus hiring intersection, or single first-party visit to a high-intent page. Action: added to a personalized sequence within 5 days.
Cold. Third-party content consumption only. Action: included in a quarterly nurture campaign. Not worked as outbound.
That is it. No 100-point scoring model. No machine learning layer. Three tiers, three rules, one action per tier. We get pushback every time we suggest this and we are right every time. The team that runs three rules consistently outperforms the team that runs 30 rules badly.
What it costs to build this
This is where I expect to lose half of you. The honest answer is that the right intent data stack usually costs less than what teams already spend, and puts the money behind signals the reps will actually work.
Here is a realistic budget for a 20-person sales team.
Champify or UserGems for job change tracking: $1,500 to $4,000 a month depending on contact volume. Clay for enrichment and signal orchestration: $800 to $2,500 a month depending on credit usage. RB2B or HubSpot website tracking: bundled with the CRM most teams already have. Theirstack or a custom Clay table for hiring signals: $400 to $1,200 a month. n8n for orchestration: $50 to $500 a month self-hosted on a small VPS.
Total: roughly $3,000 to $8,000 a month for a high-quality, owned, four-signal stack. Compare that to $7,000 to $15,000 a month for a single third-party intent vendor whose output your reps have to be persuaded to work.
The cost argument is the weaker half of the case anyway. The stronger half is ownership. Every signal in that stack comes from data you can inspect, tune and keep, rather than a score you have to take on trust and lose access to the day you stop paying.
The three mistakes I see every quarter
I get a version of this call once a week. A head of sales or CMO has a vendor contract coming up for renewal. They want a second opinion. Three mistakes come up every time.
Mistake one: confusing topic relevance with buyer intent. A topic spike on "data integration" does not mean the account is buying. It might mean they hired someone who blogs about it. Topic relevance is an ICP filter, not a trigger.
Mistake two: routing intent signals to SDRs without an opener. A rep who gets an alert with no context will revert to their default outreach. The signal is wasted. Always pair the signal with the message.
Mistake three: treating intent as a marketing project. Intent data is a sales delivery problem. If marketing builds the system, sales will not work the queue. The system has to be designed with the reps who will use it, scored on the meetings it produces, and owned by RevOps in the middle.
A short note on AI agents and intent data
Every vendor is putting an "AI agent" on top of their intent product right now. Some of these are real. Most are a wrapper that summarizes signals into a Slack message.
The actual use cases that work in 2026 are narrower than the marketing suggests. AI can take a job change signal and draft a personalized opener referencing the new role and the company's stated priorities. AI can take a website visit pattern and write a short outreach that mentions what pages were viewed. AI can pull funding context and competitor mentions into a research brief before a meeting.
What AI cannot do well yet is decide which signals to act on. That judgment still needs a person with context on the account and the rep's calendar. We are building these systems for clients now as part of our AI automation work, and the pattern is always the same. AI does the prep. Humans do the prioritization and the calls. If you want the wider version of that argument, where intent sits inside the plan rather than beside it, we wrote up how to build a GTM strategy using AI phase by phase.
Paying for intent data that does not book meetings?
We will audit your current stack and show you which signals are worth keeping, what to cut, and what to add. 30 minutes, no slide deck.
Book a free intent audit →FAQ
What is the difference between first-party and third-party intent data?
First-party intent is behavior on your own properties, like website visits, content downloads, or form submissions. You own the data and it is the most reliable signal. Third-party intent is content consumption tracked across publisher networks by a vendor, then mapped to accounts using IP databases. Third-party is much noisier and converts to meetings at a far lower rate.
How accurate is B2B intent data from vendors like Bombora or 6sense?
Account-level identification ranges from 35 to 65 percent depending on the source, with enterprise accounts on shared IPs being the most accurate and remote-heavy companies the least. The topic classification layer adds another 20 to 30 percent error margin. The reliability is good enough for ICP segmentation, not good enough as a standalone buying trigger.
Is intent data worth it for SMB sales teams?
For teams under 10 sellers, third-party intent vendors are usually not worth the spend. The list sizes they produce are too large to action and the signal quality is too low. SMB teams get more value from first-party tracking, job change signals, and a tight Clay-driven hiring signal workflow.
What is the best B2B intent data provider in 2026?
There is no single best. For first-party de-anonymization, RB2B and HubSpot built-in tracking work well. For job change signals, Champify and UserGems lead. For hiring and stack data, Theirstack is fast. For third-party topic intent, Bombora has the broadest publisher network but the same conversion limits as everyone else.
How do you measure ROI on intent data?
Track three numbers per signal source. Meetings booked, opportunities created, and closed-won revenue, attributed back to the signal that triggered the outreach. Most teams stop at meetings booked, which inflates the perceived value. The real test is closed-won pipeline. If a signal source is not producing closed revenue at 8 to 12 times its cost in 90 days, kill it.