Canonical SignalQL examples

Each example lists the product question, SignalQL, logical output shape, and a compiled Postgres SQL pattern summary. For exact generated SQL text + params, run the query through the CLI or MCP run_signalql.

Dataset assumptions match fixtures/seed.sql.

Funnel — signup to activation

Question: What share of signups reach activated within the window?

SignalQL:

FUNNEL "signup" THEN "activated" FROM events DURING LAST 28 DAYS

Shape: One row with step_1, step_2, … counts and conversion_rate.

SQL pattern: Filtered aggregate with COUNT(DISTINCT CASE WHEN event_name = $k …) (see compiler output).

---

Retention starter — active vs signup volume

Question: How many users fired signup vs activated recently?

Use two queries (v0.1 keeps retention algebra minimal):

COUNT users FROM events WHERE event_name = "signup" DURING LAST 28 DAYS

COUNT users FROM events WHERE event_name = "activated" DURING LAST 28 DAYS

Shape: Single-row count each.

SQL pattern: Two parameterized COUNT(DISTINCT user_id) queries over events filtered by event_name and time window.

---

Growth — daily signups

Question: How many signup events per day?

COUNT events FROM events WHERE event_name = "signup" DURING LAST 30 DAYS GROUP BY DAY

Shape: Rows (day, count) ordered by day.

SQL pattern: SELECT date_trunc('day', timestamp) AS day, COUNT(*) ... GROUP BY day ORDER BY day.

---

Onboarding dropoff — funnel slice

Question: Where do users drop between onboarding completion and activation?

FUNNEL "onboarding_complete" THEN "activated" FROM events DURING LAST 14 DAYS

Shape: Step counts + conversion_rate.

SQL pattern: Distinct-user conditional counts per funnel step with derived conversion ratio.

---

Feature usage — view items

Question: How often was view_item tracked?

COUNT events FROM events WHERE event_name = "view_item" DURING LAST 14 DAYS GROUP BY DAY

Shape: (day, count) rows.

SQL pattern: Day-bucketed aggregate over events filtered by event_name and bounded timestamp.

---

Segmentation — pro plan purchases

Question: Among purchases tagged plan=pro on the event, how many distinct users?

COUNT users FROM events WHERE properties.plan = "pro" AND event_name = "purchase" DURING LAST 28 DAYS

Shape: Single-row count.

SQL pattern: COUNT(DISTINCT user_id) with predicates on properties.plan, event_name, and timestamp window.

---

These examples are mirrored in fixtures/query-fixtures.json and the playground presets.

v0.4 Operational Graph

These use the aspectgraph-work-pack relationship vocab over v0.4 core operators, and run against the bundled graph fixture.

Unvalidated assumptions (structural absence — not "at risk"):

MATCH entity(assumption) WITHOUT validated_by RETURN entity_id

Reasoning lineage — why does this ticket exist?

TRACE entity(ticket) WHERE id = "123" VIA derived_from DEPTH <= 6

Shape: ordered lineage chain ticket → decision → experiment → metric → assumption → interview.

Reachability — decisions connected to an experiment:

FIND entity(decision) CONNECTED TO entity(experiment) WHERE name = "onboarding-v2" VIA supports RETURN entity_id

Context bundle for an LLM:

CONTEXT FOR entity(roadmap) WHERE quarter = "Q3" DEPTH <= 2 INCLUDE entities, relationships, evidence

Time-travel — reasoning as believed at a past date:

TRACE entity(ticket) WHERE id = "123" VIA derived_from AS OF "2026-03-01T00:00:00Z"