Handbook

Part II · Strategy before tools · Chapter 5

Signals: the reason to write today

Events that make a company likely to buy now, how to tell a strong signal from noise, and how fast they go stale.

12 min read

A signal is a reason to contact a company now rather than some other month. Good signals turn a cold message into a timely one, because you write when the problem is fresh. Bad signals add cost and noise. The skill is telling the two apart, and acting fast enough that the moment hasn't passed.

What you'll take away

  1. 01What makes an event a buying signal rather than just news.
  2. 02The main sources of signals, inside and outside your own systems.
  3. 03Four tests for judging whether a signal is worth automating.
  4. 04Why every signal has a shelf life, and how that sets your response time.
  5. 05How to turn a signal into a complete play, and where privacy limits apply.

What a signal is

A signal is an observable event that makes a specific company more likely to need what you sell soon. The test is the link to your product. A funding round is a strong signal for a recruiting tool, because funded companies hire. For a tool that cuts cloud costs it is a weak one, because newly funded companies rarely worry about costs. The same event can be gold for one product and noise for another.

Where signals come from

SourceExamplesNotes
Your own dataPricing-page visits, trial sign-ups, product usage, replies to past campaignsStrongest and cheapest; check consent rules for website tracking
HiringJob ads for roles that use or need your productVery informative: a company advertises what it's about to do
People movesA new leader in the buying team; a past champion joins a new companyNew leaders often review tools in their first months
Company eventsFunding, acquisitions, new offices, product launchesWidely available, so widely used: competitors see them too
Technology changesAdopted or dropped a tool you integrate with or replaceDetected from websites, job ads and technology-lookup services
Public recordsRegulatory filings, planning permissions, procurement noticesNiche, often under-used, great for specific industries

Judging a signal

Before automating a signal, test it on four questions.

  1. 1

    Is it linked to buying?

    Look back: did companies that showed it in the past buy more often? If you can, compare win rates for accounts with and without it.
  2. 2

    Can you detect it reliably?

    Is there a source that catches most occurrences without many false alarms?
  3. 3

    Is it timely?

    Can you learn about it within days, not months?
  4. 4

    Does it suggest something to say?

    A good signal gives you an opening line. 'Saw you're hiring two SDRs' leads somewhere; 'you're on our list' doesn't.

The best signals are the ones that also tell you what to write.

Signals go stale

Every signal loses value over time, at very different speeds. Someone comparing pricing pages is deciding this week. A new department head is reviewing tools over the first few months. A funding round shapes priorities for the following quarters.

days since the event0306090120value of actingWebsite pricing-page visit (half-life ~3 days)New head of department hired (half-life ~30 days)Funding round announced (half-life ~60 days)
Fig 1. Illustrative shelf lives. The faster a signal decays, the more of the response has to be automated.

The shelf life sets the engineering requirement. A signal that is stale in days needs an automated pipeline from detection to a rep's queue within hours. A signal that stays useful for months can run as a weekly batch. Match your build to the decay, and don't over-build for slow signals.

From signal to play

A signal on its own is a row in a table. A play is everything that happens next, decided in advance.

  • Trigger: which signal, from which source, with which checks (is the company in the ICP? already a customer? contacted recently?).
  • Who: which persona to contact, and how to find them.
  • What: the message, built around the signal, plus the follow-ups.
  • When: how fast, given the signal's shelf life.
  • Who acts: automated send, a rep's queue, or both.
  • How you'll know: reply and meeting rates for this play compared with your baseline.

Privacy and good judgment

Website visitor identification and person-level tracking come with legal limits that vary by country, especially in Europe under GDPR. Check what your company's legal team allows before using them, and never write a message that reveals you watched someone browse your site. "I saw you were on our pricing page yesterday" reads as surveillance. Use the signal to decide when to reach out, and lead with something useful instead.

Try it

Choose and design a signal play

You sell a tool that helps companies manage employee equipment such as laptops and phones. Candidate signals: (a) the company raised a Series B, (b) the company posted a job for an IT manager, (c) the company opened an office in a new country, (d) someone from the company read your blog.

Rank them, pick one for a play, and outline the play.

Write your answer down first, then reveal the worked answer

Ranking: (c) and (b) first. A new country office means shipping, tracking and tax on equipment, a concrete pain. An IT manager hire means someone is about to own the problem. (a) is a weak proxy, since growth is coming but the link is indirect. (d) is too vague alone.

Play for (c): trigger on an office announcement or job ads in a new country, checked against the ICP and excluding customers. Contact the head of IT or operations, or HR if there is no IT lead. Message about getting laptops to new hires in a country with no local stock or supplier. Send within two weeks of detection. Measure reply and meeting rates against your general campaigns.

A strong answer explains each ranking by the link to the product, not by how impressive the event sounds.

Practise the skills