Most outbound still runs on the wrong clock. You build a list, you work it, and you reach people on the day your sequence says rather than the day they care. Signal-based selling flips that: instead of deciding who to contact and hoping the timing works out, you let an event decide when. The logic is simple and well supported. Fit tells you who could buy. Timing tells you who might buy now. Lists only ever gave you the first half.
Signals and intent data are not the same thing
These two get used interchangeably and they behave very differently in practice.
| Trigger signals | Third-party intent data | |
|---|---|---|
| What it is | A discrete event: a hire, a funding round, a tech change | Aggregated research behaviour across a category |
| Precision | High. The event either happened or it did not | Lower. Probabilistic, often account-level only |
| Window | Clear. Usually days to weeks | Vague. "In market" can mean months |
| Who to contact | Often names the person | Usually just the company |
| Best use | Deciding when to reach out and what to say | Deciding which accounts to prioritise |
Use intent data to shape the target list and trigger signals to time the touch. Teams that confuse the two end up emailing a whole account because someone there read a blog post.
The signals that actually predict a conversation
Not all events are worth acting on. The useful ones share a trait: they create a problem or a budget that did not exist last month.
- A new person in a relevant role. The strongest common signal. New leaders change tools, review vendors and need early wins. The window is roughly the first 90 days.
- Funding. Not because they now have money, but because a raise comes with headcount and process commitments.
- Hiring for a role adjacent to your product. A company recruiting three support agents is telling you about volume it cannot handle.
- Tech stack changes. Adding or dropping a tool in your category, or adding one that pairs with yours.
- Expansion. A new office, market or product line breaks processes that used to work.
- Your own website. A known account visiting your pricing page twice in a week is the highest-quality signal you own. See website visitor identification.
- Dormant pipeline waking up. A closed-lost account whose blocker just left. See re-engaging old leads with AI.
Weak signals to be careful with: a single social like, generic category research, and anniversaries. They produce outreach that reads as pretextual, because it is.
Every signal has an expiry date
This is the part most teams get wrong. A signal is only valuable inside its window, and the window is shorter than your sequence.
- New hire: weeks two to twelve. Too early and they are still finding the canteen; too late and they have already chosen.
- Funding: the first two months, while plans are being made rather than executed.
- Job posting: while the role is open. Once it is filled, the pain is being solved another way.
- Website visit: hours to days. This is the one where speed genuinely changes the outcome. See speed to lead.
- Tech change: one to three months, while the new stack is still being shaped.
If your process cannot act inside the window, the signal is decoration. Measure how long it takes you to go from event to first touch, because that number decides whether any of this works.
Where AI helps, and where it quietly hurts
AI is what makes signal-based selling practical at more than ten accounts, because monitoring events manually does not scale. It is genuinely good at:
- Watching many sources at once and spotting the event.
- Enriching the event: who the new person is, who they report to, what the company does.
- Drafting a first touch that references the specific event rather than a generic hook.
- Ranking today's signals so a rep works the ten that matter.
Where it hurts: AI will find a signal for every account if you let it, because finding one is what it was asked to do. That produces outreach built on a pretext the buyer can feel. The discipline is to keep a short list of signals you actually believe in and ignore the rest, even when the tool surfaces them.
The second trap is volume. Signal-based selling should mean fewer, better-timed touches. If your send volume went up after adopting it, you implemented list-based outbound with extra steps, and the deliverability cost is real.
Writing the touch so the signal does not feel creepy
There is a line between "I noticed something relevant" and "I am watching you". Three rules keep you on the right side.
- Use public events only. A funding announcement or a job posting is public. Inferring from someone's browsing is not something to name out loud.
- Lead with the implication, not the observation. Not "I saw you hired a Head of Support." Rather, a short line about what usually breaks when a support team doubles.
- Be useful without the meeting. If the email only has value when they reply, it is a pitch wearing a trigger as a hat.
The honest test: would this message still make sense if they knew exactly how you found them? If not, pick a different signal.
A setup you can run in two weeks
- Pick three signals, no more. Start with new hires in your buyer role, open job postings that imply your problem, and known-account visits to high-intent pages.
- Define the window for each one and write it down.
- Set a daily feed of matching accounts, capped at a number a rep can actually work.
- Write one message template per signal. Not per persona, per signal, because the event is what makes the message land.
- Track time from event to first touch. Fix that before you add a fourth signal.
- After a month, drop the signal with the worst reply quality. There will be one.
What to measure
- Time from signal to first touch. The core operational metric.
- Reply rate by signal type, which tells you which events are real and which were convenient.
- Meetings held per 100 signals acted on, rather than per 100 emails sent.
- Coverage: what share of your target accounts produced any signal this quarter. If it is low, your signal set is too narrow for your market.
- Pipeline from signal-sourced outreach against your list-based baseline. See AI sales forecasting.
When list-based outbound is still the right answer
Signal-based selling assumes your market produces observable events. If you sell into small businesses that do not announce funding, post jobs publicly or change tooling visibly, you will find very few signals and most of them will be weak. In that case a well-researched list with a strong offer beats waiting for events that never surface.
It also assumes you can act fast. A team that reviews leads weekly cannot run on windows measured in days. Fix the response time first, then add the signals. For the wider picture, implementing AI in B2B sales covers the sequence, and the autonomous SDR covers how much of this you can safely hand over.
