CEDX Product Analytics · Data & Analytics
Product Analytics turns the event stream into answers: activation and retention against the funnel, session replay ranked by friction score, and a live view of who is doing what, about four seconds behind reality.
The friction leaderboard doesn't average the pain away: the invite modal at 68, with 25 rage clicks and 19 dead clicks across four sessions. That's a fix list, not a metric.
Runs on demo data — the header reads “ingest lag ~4s · demo seed” on the screen itself.
What it is
Goal mode shows conversion per step: signup completed at 100%, workspace created at 78%, project opened at 56%, task created at 39%. A free-plan D7 retention dip of −3.2pp gets an alert that names the day-2 empty project state as the cause.
Sessions carry rage clicks, dead clicks and errors rolled into a friction score — 98 at the top of the queue. The leaderboard aggregates by screen: invite modal 68, upload dialog 63, timeline view 51.
The Live screen streams the last 30 minutes: 1,284 active users, 9,000 events a minute, page_viewed leading at 412 users. The event feed reads like a log — and says demo stream on itself.
Product tour
Not a mockup and not a concept deck. This is what opens at /app/product-analytics.
01 — Overview
DAU 18.4K, MAU 42.1K, 2.1M events a day, D7 retention 34.6%, activation 28.4%, core feature use 41%. Product health reads 88 with schema drift, dead events, stale flags and SDK lag itemized.
02 — Session replay
28 queued sessions with friction signals, 9 scoring 60 or more. Each row shows duration, events, rage and dead clicks beside the score — 37 minutes in the invite modal with 13 rage clicks explains itself.
03 — Live
task_created, page_viewed, flag_evaluated, invite_sent, plan_upgraded — twelve recent events with offsets in seconds, one of them an upgrade worth +$24 MRR. The footer admits the plan filter is cosmetic in the demo.
Who runs it
Roles, not references. We have no named customers yet, so nobody in these photographs is quoted, credited or claimed as one.
Owns activation: watches the 28.4% signup-to-project rate, the step where it leaks, and the alert when a rollout changes the shape.
activation · 28.4%
Reads replay before roadmap: the invite modal's 68 friction score and 25 rage clicks outrank opinion in the planning meeting.
friction · invite modal 68
Keeps the event schema honest — 3 drift items, 5 dead events, 4 stale flags — so downstream numbers mean what they say.
schema drift · 3 openThe shape of it
Every figure below is legible in the captures above. Nothing here is a benchmark for your product — it is the state of the demo data.
How it runs
Events land with a governed schema — drift, dead events and stale flags are counted, not assumed away.
Activation, retention and feature use compute against the funnel; goal mode prints conversion per step.
Replay sessions rank by friction; rage and dead clicks roll up to a per-screen leaderboard.
Live shows the last 30 minutes with the lag on the card, so a launch's first hour is observable.
One record
Product Analytics is not an isolated event bucket. Its stream feeds the same governed record the metric layer certifies and the warehouse stores.
Product events feed the governed metric layer; the activation rate here is a certified metric there.
CEDX EmbeddedUsage on embedded analytics is itself product telemetry — the two consoles read each other's numbers.
CEDX SyncEvents reach the warehouse through the same pipelines Sync meters — ingest lag here, freshness SLA there.
CEDX NotifyAlert-worthy movements — retention dips, friction spikes — route to the people who own the screen through the estate's notification spine.
Limits
Finding this out on the third call is worse for you than reading it here, and worse for us.
Start
Open it right now
Your events, your funnel
Product Analytics with the rest of it
Questions
Yes. Every screenshot is a capture of the running build, and you can open the same build at /app/product-analytics. It runs on demo data, which the header states on screen.
Rage clicks, dead clicks and errors rolled up per session, then aggregated per screen. The invite modal leads the leaderboard at 68 — 25 rage clicks and 19 dead clicks across four sessions — which makes it a fix list, not a metric.
About four seconds behind, and the header card says so. Users per minute covers the last 30 minutes; the event stream shows the twelve most recent events with offsets in seconds.
Signup to project — 28.4% in the demo workspace. Goal mode prints the funnel step by step: signup 100%, workspace created 78%, project opened 56%, task created 39%, so the leak has an address.
No certification has been issued. What we can evidence about hosting, encryption, tenant isolation and production access is written up on the security page.
Live build, demo data, no card. Then check what your own invite flow's friction score would be.