Skip to Content
CoursesSignals & ABM02 Signal types

Signal Types in Sync GTM

Outcome: three signals chosen for your motion, each with a detection action, a cost per row, a window and a check cadence written down.

Surface
App and MCP server
Level
Intermediate
Uses
The signal enrichments — see the catalogue below
Credits
~5 for a five-row test of three signals
Prerequisite
Lesson 01

The catalogue

Grouped by what the event tells you.

Money

SignalDetectsWindowReference
Recent fundingA round closed90 daysRaised funding recently
Total fundingCumulative capital— (attribute)Company total funding
M&AAcquired or acquiring180 daysMergers and acquisitions

People

SignalDetectsWindowReference
Job changeA person moved companies90 daysJob change
PromotionNew authority, same company90 daysRecently promoted
Leadership changeA new exec in seat90 daysLeadership change
Headcount growthThe org is expanding90 daysHeadcount growth

Building

SignalDetectsWindowReference
Is hiringOpen roles exist30 daysIs hiring
Job listingsWhich roles, with the text30 daysCompany job listings
Tech-stack changeA tool added or removed180 daysTechstack change
New product launchedA launch60 daysNew product launched
New locationGeographic expansion180 daysNew location

Attention

SignalDetectsWindowReference
LinkedIn engagementInteraction with relevant content14 daysLinkedIn engagement
New blog postedContent activity30 daysNew blog posted
Mentioned in newsPress coverage60 daysMentioned recently on news
Website traffic growthDemand rising90 daysWebsite traffic growth

Latency matters as much as the signal

A signal is only useful if you find out inside its window. Each source has its own lag between the event and the data being available.

SignalTypical detection lag
Job postsDays — fastest of the useful signals
LinkedIn engagementDays
Job change and promotion1–4 weeks, depending on when the person updates their profile
Funding1–6 weeks — announcement lags the close
Tech-stack changeDetected on your re-run cadence, so up to a quarter
Headcount growthMonthly at best

Latency plus your check cadence is your real time-to-action. A funding signal with a four-week lag, checked monthly, means you reach the account on average six weeks after the round closed — most of the way through the window. Match the check cadence to the signal, not to your calendar.


Cost shapes the design

Cost bandSignalsImplication
CheapJob listings, is hiring, headcount, fundingRun across the whole watchlist weekly
ModerateJob change, promotion, leadership changeRun across a tracked list, gated
HigherStack change, traffic, newsQuarterly, or on high-value accounts only

The cheap band is where a signal pipeline lives. If your primary signal is in the higher band, restrict the watchlist rather than the cadence.


Choosing your first three

One that fires often

Job posts, for most B2B motions. It gives the pipeline enough volume to be worth running.

One that is high-conviction

Funding, exec hire, or job change. Fires rarely, converts well.

One specific to your product

The stack change, the launch, the expansion — whatever means “they now need what we sell”.

Three is the right number to start. More than that and you cannot tell which one is producing the meetings.


Do this now

Write your three signals

With the action that detects each.

Write a window for each

In days. Without a window it is not a signal.

Write a check cadence

Matched to window and latency, not to convenience.

Estimate the weekly cost

Rows in the watchlist × cost per signal × frequency.

Test each on five known accounts

Companies where you know the event happened. Did the signal fire? How stale was it?

Record the detection lag you observed

It changes what cadence you need.


Check your work

  • Three signals, each with an action, window and cadence
  • The cadence accounts for detection lag, not just the window
  • You have a weekly cost estimate
  • Each signal was tested against an account where you know the answer

Where this breaks

Running every available signal at once produces a pipeline nobody can debug. When ten triggers all feed one queue and reply rates are poor, there is no way to tell which signal is worthless and which is carrying the whole thing. Start with three, measure each separately, and add a fourth only when you can state the conversion rate of the first three.


Further automation

Store the signal date, not just a true/false flag. Everything in this course — windows, decay, dedupe, prioritization — depends on knowing when the event happened, and a boolean throws that away permanently.


Next lesson

03 — Funding trigger campaigns, the first complete play.

Reference for this lesson: Signals, How enrichment works, Credits, GTM Engineering — attributes versus signals.