Signals
A signal — a fit criterion, shown in the app as a lead tag — is one thing that makes a customer a better or worse match for a product. Each has a name, a description, a type, and a weight whose sign says which way it pushes: positive for a good sign, negative for a red flag.| Type | Shown as | Scored from |
|---|---|---|
profile | Who they are | The customer and company profile — industry, size, location, language |
conversation | What they say | The content of their messages |
engagement | How they engage | Opens, clicks and replies |
Building a score
Signals average within their type, and the three type scores combine into the overall. A type where nothing has been assessed is left out of the overall instead of being counted as zero. When too little weight has been assessed to say anything, the customer reads Not scored rather than 0%.Manage signals
Open a product, go to its lead-tags section, and add signals there. See Product Catalog. The type is fixed once a signal is created; its name, description, weight and active flag stay editable.When scores are recomputed
- A profile signal re-runs when the customer or company fields change.
- A conversation signal re-runs when a new message arrives.
- Submitting feedback in the Human Review modal re-runs everything on that product.
- Customer drawer → ”…” → Re-score now re-runs everything for one customer.
- The
adjust-scorenode nudges one signal from inside a campaign. See Contact Nodes.
Getting useful scores
Start with two or three signals
Start with two or three signals
A handful of clear signals sorts a list better than twenty vague ones. Add more once you can see which ones move.
Write signals a stranger could apply
Write signals a stranger could apply
The description is what the scorer reads. “Engineering team of 10+” is checkable; “good technical fit” is not.
Retune with won deals
Retune with won deals
Every quarter, compare which signals the customers you actually closed scored well on, and move the weights toward those.