The short answer
A buying signal is an observable change at a company that makes your offer more relevant now: a new leader, a funding round, a job post, a launch. The signals that work are the ones tied to a problem you solve, checked against your ICP, and acted on before they go stale. Everything else is noise with a timestamp.
Three rules carry most of the result:
- Fit before timing. A signal never makes a bad-fit account a good one. Score fit first, then let signals move good-fit accounts up the queue.
- Signals decay. A funding round is news for weeks, not a year. Without decay, your active list fills with accounts whose moment passed months ago.
- The value test. Remove the call to action from your message. If what is left still helps the reader, send it. If it reads as "I saw you raised, want a call?", rewrite it.
The signal table
Decay windows below are starting points. Test them against your own reply and meeting rates, and adjust after three or four campaigns.
| Signal | What it suggests | Where to find it | Decay (starting point) | What to say | What it does not mean |
|---|---|---|---|---|---|
| New marketing, growth or revenue leader | A review of plan, stack and partners in the first 90 days | LinkedIn job changes, Sales Navigator alerts, company announcements | Full weight to 30 days, half by 90, gone by 180 | One observation about what new leaders at their stage usually inherit | That they have budget, or will change anything you sell |
| Former customer or champion in a new role | A buyer who already knows your work, with a fresh mandate | Your CRM contacts tracked against LinkedIn job changes | Strongest in the first 90 days | Congratulations and one specific note on the new company. No pitch | That the new company has the same problem as the old one |
| Funding round | Growth targets went up and some budget exists | Crunchbase, funding newsletters, press releases | Full weight to 30 days, half by 90 | Use it as context only. Pair it with a second signal before writing | That budget is set aside for your category. Everyone sends the funding email |
| Hiring a RevOps, marketing ops or GTM engineer role | They know the systems are not built and have named the gap | LinkedIn Jobs, careers pages, ATS boards (Greenhouse, Lever, Ashby) | While the post is open. After the hire, the new person is the signal | The problem the role is meant to fix, and what usually breaks before it is filled | That they want outside help. Many want to build in-house |
| Hiring a paid media or performance role | Acquisition is a priority and capacity is short | Job boards, careers pages | While the post is open, about 30 to 60 days | Something specific about their channel mix or measurement | That their current spend is failing |
| Public request for a partner or vendor | Stated need, stated budget, a shortlist forming | LinkedIn posts asking for recommendations, tender portals, communities | Days, often under two weeks | A direct answer to the request they wrote | That they will read anything generic |
| New market, product or region launch | The current motion will not stretch to the new market | Press releases, new site sections, leader posts | Full weight to 30 days, useful to 90 | What changes in pipeline and targeting when a team enters that market | That the launch is funded for outside work |
| CRM or core stack change | Routing, lifecycle and reporting are being decided now | Technographic tools (BuiltWith, Wappalyzer), CRM admin job posts | 90 days, longer for a full migration | What to settle before the migration locks in: stages, definitions, ownership | That they are unhappy with the new tool |
| A leader posts about a problem you solve | Self-declared pain, in their own words | LinkedIn, podcasts, community threads | Steep: about 14 to 30 days | A reply to the point they made, with something useful attached | That they want to be sold to in the comments |
| Event exhibitor or speaker | Demand will arrive in a burst and needs follow-up | Exhibitor and speaker lists | From 30 days before to 14 days after | Their follow-up plan for the leads the event creates | That the event is a priority beyond the booth |
| Repeat visits to pricing or comparison pages | Active evaluation, possibly against a competitor | Your own site analytics and visitor identification | Days: act within a week | Help with the decision they are making, not a reminder they visited | That the visitor is the buyer, or that the company is identified correctly |
| Third-party intent topic surge | Research activity on a topic across the web | Intent data providers | One to two weeks, weak on its own | Nothing on its own. Use it to rank accounts that already show another signal | That anyone at the account is in a buying cycle |
For what the hiring signal looks like in real data, including which GTM titles appear in which markets, see our Q4 2026 hiring data.
How decay works in a score
A simple model that works: score fit out of 70 (firmographic, technographic, organizational) and signals out of 30, and multiply each signal's points by its age.
| Signal age | Weight |
|---|---|
| 0 to 30 days | 100% |
| 31 to 60 days | 75% |
| 61 to 90 days | 50% |
| 91 to 180 days | 25% |
| Over 180 days | Expired |
Posts, event signals and requests for a partner use a steeper curve and expire after about 30 days, or sooner. A funding round worth 30 points on the day it is announced is worth 15 at 90 days and nothing at 180.
Combinations beat single signals
Two signals in the same window say more than either alone. Funding plus a RevOps hire in the same month says budget exists and someone is rebuilding. A new marketing leader plus a launch says a fresh mandate in a market the current motion does not reach. Give combinations a bonus in your score, and route them to a person, not a sequence.
Fit before timing
Every signal check starts with three questions, in order:
- Does the account fit? Size, B2B model, market, stack and your exclusions. If not, the signal is logged and nothing else happens.
- Who owns it? Customers, open deals and accounts in cooldown go to their owner. A signal does not override ownership.
- Is the signal still live? Check the event date, not the date you found it.
Only then does the signal decide what to say and how fast to act.
Write about the problem, not the signal
The weakest signal-based email leads with the signal: "Congrats on the raise." The prospect already knows they raised. A stronger message names the problem the signal usually points to and offers one useful thing about it: a benchmark, a common failure, a question worth asking before the next quarter.
Then run the value test. Delete the call to action. If the email still gives the reader something they can use, it is ready.
Keep the library honest
A signal library starts as a set of hypotheses. Each signal earns its place with results: track sends, replies, meetings and pipeline per signal, and review the points and windows after every few campaigns. Signals that never produce meetings get cut, however good they sound. The signals that hold up are often specific to your own closed-won deals, so look there first.
The step from a verified signal to a reviewed campaign is in Signal to sequence, and choosing between signal-led and other outbound motions is in Choose an outbound motion.
Questions buyers ask
Which buying signals work best for B2B outbound?
The ones tied to a problem you solve and checked against your ICP. For most B2B tech teams, a new marketing or revenue leader, a former champion in a new role, and a direct request for a partner come first. Funding and intent data work best as a second signal, not a reason to write.
How fast do buying signals decay?
A useful starting point: full weight for 30 days, 75% to 60, 50% to 90, 25% to 180, then expired. Posts, events and partner requests decay faster, inside about 30 days. Test the windows against your own reply and meeting rates.
Do we need intent data to run signal-based outbound?
No. Job changes, job posts, funding news, launches and your own site visits are public or already yours. Intent data can help rank accounts, but on its own it rarely says enough to write a relevant message.
How many signals should we track?
Start with five to eight that map to problems you solve, with points and decay windows for each. Add a signal only when you can say what you would write to an account that shows it. A long list nobody acts on is worse than a short one with results attached.
Can AI find and act on buying signals for us?
AI can watch sources, check dates, match accounts to your ICP and draft a first message from your own positioning. A person should still review the account, the reason and the first touch before anything is sent, because a wrong signal sent with confidence costs more than a missed one.
Publication and source record
Published on GTMhub: . Last reviewed: .