A signal is the reason to write.
Last updated 24 September 2026
HeyLead does not only find people who match a title. It watches for the moment something changes, and the first message says why now. Signals add people and angles on top of the ICP search; they are not the main lead source yet.
What it collects
- On LinkedIn. Posts that mention your keywords or competitors; your prospects' own posts about a problem you solve; job changes, promotions and headline changes; hiring surges; who viewed your profile (Premium or Sales Navigator seats); who engaged with or followed your company page; who commented on your competitors' posts.
- Around the web. Funding, acquisitions, executive hires, expansions and layoffs from news; Reddit, Hacker News, G2 and public job boards.
- On your site and in your tools. A one-line script tag shows which companies visit and which pages say intent (company-level, never a named person); an ingest API and a batch endpoint for Clay, Zapier or anything that can POST; webhooks out above a score you choose.
- Inbound. Every connection request accepted and qualified; strangers' DMs triaged; vendor pitches and recruiters filtered out.
What it stores
Each signal with its type, source, person or company, content, intent, confidence, score, status and an expiry. Watchlists (keyword, competitor, company, person, industry) created from your ICP the moment a campaign is active. Provenance on every prospect: "Described a problem you solve", "Viewed your profile", "Changed jobs".
How it scores
A classifier assigns intent and confidence. Score = type weight × confidence × freshness. Every signal expires; the window depends on how long the moment lasts:
| Signal | Expires after |
|---|---|
| Profile view | 3 days |
| Keyword mention | 7 days |
| Prospect's post, competitor mention, hiring surge, news, competitor commenter | 14 days |
| Job change, promotion, company change, funding | 30 days |
Two signals on one person within 14 days form a named pattern: a new role plus hiring is "new leader building". Every signal is matched to your ICP before anyone is enrolled.
What it triggers
Above the threshold, the person joins the best matching campaign with the signal written into the opener. A company-level signal (hiring, funding) becomes people: HeyLead searches your ICP's title at that company. An existing contact gets a priority boost instead. Dismissed signals keep their reason: no identity, ICP mismatch, below threshold, over budget.
A worked example
You sell a webhook debugger for Stripe. Your ICP's watchlist holds "Stripe webhooks", "failed payments" and two competitors. On Tuesday a payments lead at a 40-person fintech posts: "our webhook retries have been failing silently for a week". The classifier reads it as a problem you solve, confidence high, freshness day one, and the person matches your ICP. Score above the threshold: she joins your campaign with the post as the reason, and the opening message cites it. On Friday the same company appears in a hiring surge for payments engineers: two signals on one company within 14 days, a pattern, and the priority goes up. Names and figures here are invented.
Limits, in numbers
- Profile reads are counted per seat against a ceiling of 150 a day, signals included.
- In the 14 days to 14 September 2026, 3.7% of collected signals became outreach. Signals are a minority lead source; the ICP search is the majority.
- The signal tools run in the cloud and on the dashboard; in chat they need
HEYLEAD_TOOLS=all. - Site visits are company-level. No named visitor is ever stored.
HeyLead is not affiliated with LinkedIn. It reaches LinkedIn through Unipile, which is also not affiliated. LinkedIn's User Agreement prohibits automated messaging and may restrict an account; HeyLead's pacing lowers that risk and does not remove it.
Related: ICP and enrichment · Tool reference · Pricing