Signals — what changed for this customer, and does it matter?

Signals answers one question: What changed for this customer, and does it matter? A payment slipping, a plan being cut, usage dropping, a deal going quiet, a customer growing — each is a signal. Beacon looks for the same signals in the same way across all your tools, so every other module works from one shared reading of what is happening to each customer. It belongs to the Shared foundation — see Modules.

​

What it does

Beacon keeps a signal list. Each signal on it has a rule for what counts, a cut-off where it needs one, a direction — a warning sign or a good sign — and a fade setting of 14, 30 or 90 days, which says how quickly it stops mattering. When something in your tools matches a signal, Beacon records it against the right customer, with the date it happened. The signals are sorted into kinds — churn risk, customer health, expansion, buying intent, engagement, deal speed and others — and the modules that need them read them from here: Churn risk reads the churn-risk signals, Deal likelihood reads deal speed and buying intent, and so on. A signal switches on when two things are true: the tool it comes from is connected, and Beacon reads it. Beacon never changes anything in your tools, and it publishes no overall risk rating from signals.

​

What it produces

Its headline is two figures: Each name links to its definition, which also says how the Data API and MCP will carry it — neither serves these two figures yet. The signals themselves can already be read one by one. Behind them, the page shows today:

​

How signal coverage is worked out

For each customer, every signal Beacon expects counts as covered when its tool is connected and up to date, the customer is matched in that tool, and the data behind it is reliable enough. The nine kinds of signal are then combined at fixed weights into a score from 0 to 100. Churn risk weighs most, at 18%; customer financial health least, at 2%. A customer is High at 80 or more, with reliable data behind it. How old a signal is does not affect coverage: a quiet customer on connected, up-to-date tools is fully covered. Age matters only when Beacon weighs what a signal means. Signal coverage sets out the full rule.

​

What you see on its page

In Beacon, open The Brain and choose Signals. The page has six tabs:

• Overview — how many signals are switched on, how many were noticed in the last 30 days, what switches on next, and the tools it reads.
• Dashboard — signals noticed over 13 weeks, warning signs against good signs, and how many kinds of signal are covered today.
• Signals & alerts — the signal list in groups by tool, with which are on, and the alert rules.
• Scorecard — the fade settings and cut-offs. When each signal stops counting reads Being settled today.
• Latest signals — every signal noticed, with search, a filter for warning signs or good signs, and a panel that opens the customer’s page.
• Setup — the questions the module uses, and the answers in use.

​

What it needs

Signals start as soon as your billing system or your CRM is connected. Every other tool adds more of the list. A signal whose tool is connected but which Beacon does not read yet says so on the page: it starts in a later update, and there is nothing for you to do. See Connect your tools for how each connection is made.

​

Reads from

​

Which modules read it

The scores these modules take from Signals are not calculated yet, so none of them is weighting signals today. The Modules page in The Brain shows which modules are running for you.

​

Alerts

Most of what Signals finds goes to other modules rather than to people. No alert is sent yet. The three score alerts wait for the scores, and the other two wait for the parts they read.

​

Setup questions

• Which signals do your tools actually provide? Beacon proposes the list from what you have connected.
• Do you have signals of your own — a tool or a spreadsheet you rely on — that should be added to the list?
• Who owns the signal list? They approve any change beyond the small adjustments Beacon may make on its own.
• When should signals be weighted by segment? Recommended: once 90 days of outcomes have built up.

Which tools you use and who looks after your data are asked once and used here too.

​

What is not calculated yet

• Signal coverage, and the signal coverage score for each customer.
• When each signal stops counting, and its fade curve.
• Scores for each kind of signal, for each customer and by segment.
• Company growth, the one kind measured for the whole company.
• How reliable the signals are — how many fire with no predictive weight, how many are held back, and how often tools disagree.
• Most of the signal list — a first set of billing and CRM signals is read today; the rest start in later updates.
• Alerts — none is sent yet.

​

Related