ingestlayer/recipes

Catch churn-risk signals in Postgres

Turn the small signals — a downgrade, a usage drop, a support spike — into one scored alert that reaches the account owner while there's still time to act.

01source

sourcesdk.eventTypeScript SDK
matchchurn.signal

02pipeline · 2 steps

  • 01ENRenrich.entitycustomer → CSM · renewal · MRR
  • 02CTLfilter.matchscore ≥ 0.7 only

03destinations · 1

  • towarehouse.pgPostgres
    tableevents.signups

the event

You emit churn.signal with this shape. The TypeScript SDK keeps the call type-safe, and the event is stored whole — so every field below is available to the pipeline by name.

  • customer_idstring
  • signalstringdowngrade | usage-drop | support-spike
  • scorenumber0–1 risk
  • mrrnumberminor units at risk

emit it

From your code with the TypeScript SDK — or any language over the REST endpoint and signed webhook ingress.

emit churn.signal
import { ingest } from "@ingestlayer/sdk";

await ingest("churn.signal", {
  customer_id: account.id,
  signal:      "usage-drop",
  score:       risk,
  mrr:         account.mrr,
});

route it to Postgres

Insert each event as a row into a table in your own Postgres.

  1. 01

    add the connection

    Paste a Postgres connection string. Connections originate from our EU region — allowlist those egress IPs on your database.

  2. 02

    point at a table

    Name the target table. Top-level event fields map to columns, and the full payload is also available as a jsonb column.

  3. 03

    map columns

    Match event fields to columns with $event.* references, or accept the default mapping into a typed events table.

in postgresdelivered
INSERT INTO events.signups
  (user_id, email, plan, source, payload)
VALUES
  ('u_018f', 'ada@acme.com', 'pro',
   'marketing-site', '{ … }'::jsonb);

notes

questions

Can I tune how sensitive the alert is?
Adjust the score threshold in the filter. Raise it for only the clearest risks, lower it to catch more early.
Who gets the alert?
enrich.entity attaches the CSM and renewal date, so the warning reaches the person who can actually save the account.
Can I combine several signals?
Emit each signal as it happens and dedupe by customer in a window, so one at-risk account is one alert, not three.
build this pipelineor read the quickstart →

churn-risk signals, routed elsewhere

more, into Postgres