Buying intent signals are the difference between guessing who might need you and knowing who is shopping this week. Most teams already collect them. Very few act on them fast enough to matter, because the signal lands in a dashboard nobody opens instead of in a rep queue with a deadline attached.
This guide covers the part other articles skip: how to score buying intent signals, how to route them with a real SLA, what to say in the first touch, and how to tell a genuine signal from a false positive. You will also get benchmark response times, a weighting model you can copy, and the compliance limits that apply.
What buying intent signals actually are
A buying signal is any observable behaviour that suggests an account is moving toward a purchase decision. It can be an action the buyer takes on your property, an action they take somewhere else, or a change in their circumstances that creates a need.
The reason this matters comes down to arithmetic. Research from the Ehrenberg-Bass Institute, widely known as the 95:5 rule, found that only about 5% of a B2B category is in market at any moment. Meanwhile Gartner reports that buying groups spend roughly 17% of the evaluation window with vendors at all. So the window is small, and it is shared. Buying intent signals tell you when it opens.
The four categories of buying intent signals
Most articles list two categories. There are four, and the last one is where the cheapest wins usually hide.
First-party signals
Behaviour on properties you own. Pricing page visits, repeated documentation reads, a switch from feature pages to security pages, demo requests, and trial activity all qualify. These are the strongest signals you will get, because there is no inference involved.
The strongest first-party pattern is not a single visit. It is multiple people from the same domain inside 72 hours. One person browsing is curiosity. Three people browsing is an evaluation.
Third-party research signals
Activity on properties you do not own. Review-site comparisons, topic surges across publisher networks, and competitor content consumption all sit here. They arrive earlier than first-party signals, but they carry more noise, because the account is identified by inference rather than by login.
Circumstantial signals
Changes in the account that create a need. Funding rounds, leadership hires, hiring surges in a specific function, office openings, acquisitions, and technology adoption or removal all count. Nobody is researching yet, so you arrive before the shortlist forms.
Relationship signals
This is the category most teams ignore. A former customer starts a new role. A champion from a closed-lost deal gets promoted. An investor in your existing account joins a new board. These buying intent signals convert far better than anything third-party, because trust already exists and only the logo changed.
Track job changes across your entire CRM, not just current customers. Closed-lost champions in a new seat are often the single highest-converting list a team can build.
How to score buying intent signals
A signal without a weight is just a notification. Score every signal on two axes: how strongly it predicts a purchase, and how quickly it decays. Then multiply.
| Signal | Strength (1-10) | Decays in | Response SLA |
|---|---|---|---|
| Demo or pricing form submitted | 10 | Minutes | Under 5 minutes |
| 3+ people from one domain in 72 hours | 9 | Days | Same business day |
| Champion job change at a past account | 9 | Weeks | Within 3 days |
| Repeat pricing page visit | 7 | Days | Within 24 hours |
| Security or compliance page visit | 7 | Days | Within 24 hours |
| Hiring surge in your buyer function | 6 | Weeks | Within 1 week |
| Funding round announced | 6 | ~90 days | Within 1 week |
| Third-party topic surge | 4 | Weeks | Add to nurture |
| Single blog post read | 1 | Immediately | No action |
Two rules make this model work. First, stack signals rather than treating them individually, because a funding round plus a hiring surge plus a pricing visit is a different account than any one of those alone. Second, decay scores automatically. A pricing page visit from six weeks ago should score close to zero today.
Routing: the step where most signal programs break
Detection is solved. Routing is not. A high-intent signal that lands in an unattended Slack channel on a Friday afternoon is worth nothing by Monday. So define three things before you turn anything on.
- An owner per signal type. Not a team, a named person, with a documented backup for holidays and time zones.
- An escalation path. If the owner does not act inside the SLA, the signal reassigns automatically rather than expiring quietly.
- A closed loop. Every routed signal ends in a logged outcome: contacted, disqualified, or no reply. Without this you cannot tell which signals are real.
Measure the median and the 90th percentile of your response time, never the average. The average hides exactly the failures you are trying to find. Our deeper look at why so few teams hit their own targets is in this piece on speed to lead.
What to say when a buying signal fires
Here is the mistake that ruins otherwise good signal programs: naming the signal. “I saw you visited our pricing page three times” is accurate, unsettling, and kills the conversation. Reference the context instead, never the surveillance.
- Funding round: lead with what teams typically do in the two quarters after a raise, then ask whether that is on their roadmap.
- Hiring surge: reference the public job posts, since those are meant to be seen, and connect them to the bottleneck the hiring implies.
- Champion job change: congratulate them, remind them of the outcome you delivered together, then offer to set it up again.
- Website activity: send the specific resource that answers what they were reading about. Provide the answer without narrating how you knew.
Keep the first touch under 90 words and end with one question. Long emails signal a template, and templates get deleted.
False positives: the buying intent signals that lie
Nobody writes about this, and it costs teams real money. Several common signals look strong and convert badly.
- Competitor research. Your competitors read your pricing page constantly. Filter their IP ranges and domains before anything routes to a rep.
- Job seekers. Careers page traffic often bleeds into product pages. Exclude sessions that entered through careers.
- Existing customers. A support search is not a buying signal. Suppress current accounts unless the visit hits expansion pages.
- Shared IP inference. Coworking spaces, ISPs, and VPNs produce confident but wrong account matches. Weight IP-based identification lower than authenticated activity.
- Topic surges without a role match. An intent spike from an account means nothing if the researcher was an intern. Confirm the function before you act.
Run a monthly audit of routed signals against outcomes. If a signal type produces conversations below your baseline cold rate, it is noise, so drop its weight to zero and move on.
Signals are useless without a clean data layer
A signal identifies an account. It does not hand you a person to contact. If your contact data is stale, the workflow stops one step before it produces anything, which is the most common way these programs quietly fail.
Three requirements sit underneath every signal program:
- Coverage. You need the right contacts at the account, not just whoever is in the CRM. See how enrichment waterfalls and match rates work.
- Deliverability. A perfectly timed email that bounces is worse than no email. Our email verification guide covers keeping bounces under 2%.
- Field integrity. Scoring runs on CRM fields, so blank fields mean broken scores. The field-level hygiene framework handles this.
We made the broader argument in why AI fails without a data layer. The same logic applies here, only faster, because signals expire.
Compliance limits you cannot ignore
Signal-based outreach touches behavioural data, so the rules are stricter than for a static list. Under GDPR, intent inference on identifiable individuals needs a lawful basis, and legitimate interest requires a documented balancing test. CCPA gives California residents deletion and opt-out rights that apply to enriched records too.
Practically, that means three habits. Keep a record of where each signal originated. Honour suppression lists across every channel rather than per-tool. Never reference behavioural tracking in your messaging, which is both better outreach and safer ground. The full picture is in our guide to GDPR, CCPA, and DNC rules for outbound teams.
A 14-day plan to launch buying intent signals
- Days 1 to 3. Pick three signals only: one first-party, one circumstantial, one relationship. Resist the urge to turn everything on.
- Days 4 to 6. Build your suppression lists. Competitors, customers, job seekers, and shared IP ranges all come out before launch.
- Days 7 to 9. Assign owners and SLAs per signal type, then set up automatic escalation.
- Days 10 to 12. Write one short message per signal type. Context only, never surveillance.
- Days 13 to 14. Launch, then log every outcome. Review after 30 days and re-weight based on what actually converted.
How to measure whether the program works
Track four numbers and ignore the rest.
| Metric | What it tells you |
|---|---|
| Signal-to-conversation rate by type | Which signals are real and which are noise |
| Median time from signal to first touch | Whether routing works in practice |
| Win rate: signal-sourced vs cold | Whether timing is worth the tooling cost |
| Sales cycle length by source | Whether you are arriving earlier or just louder |
Set a baseline before you launch, otherwise you will have no way to prove the lift. Then review by signal type monthly and prune anything below baseline.
Buying intent signals FAQ
How many buying intent signals should we track at first?
Three. One from each of first-party, circumstantial, and relationship. Teams that switch on twenty signals at launch cannot tell which ones work, so they end up trusting none of them.
Are third-party intent signals worth paying for?
Only after your first-party and relationship signals are fully worked. Third-party data is the most expensive category and the noisiest. Exhaust the free, high-converting signals first.
Should reps see raw signals or a scored queue?
A scored queue, always. Raw signal feeds get ignored within two weeks. Give reps a ranked list with a suggested action and a deadline attached to each entry.
What if we have no website traffic to work with?
Start with circumstantial and relationship signals instead. Funding rounds, hiring surges, and job changes are all public, so they require no traffic and no tracking pixel.
Act on three signals well
The teams that win with buying intent signals are rarely the ones tracking the most. They are the ones who picked three signals, routed them to a named owner with a deadline, and wrote a message that sounds human. Everything else is instrumentation.
For adjacent plays, see our guide to twelve growth plays that build pipeline, or browse more on B2B data and sales intelligence.
Ready to act on signals in real time? ZenBee surfaces buying, hiring, and job-change signals across 35M+ companies, then connects them to 700M+ verified contacts and multichannel outreach in one workflow. Request a demo, or start for free.