The Number That Cratered at 3 AM

Signups fell 38% overnight. Find out why before standup.

FThe Workshop 5 (58) 464 taken 35m Data Analyst

The situation

You wake up to 14 Slack pings. The activations dashboard shows new signups down 38% since midnight, and the growth team is already drafting a 'marketing channel is broken' narrative. But you have a hunch it's not real users walking away — it could be a tracking change, a timezone bug, a deploy, or a genuine collapse. Standup is in 35 minutes and someone will be asked to pause six figures of ad spend. You need to separate signal from artifact, name the most likely root cause, and say how confident you are — without crying wolf or missing a real fire.

What you'll practice

Distinguishes a real demand drop from an instrumentation/logging artifact using at least two independent checks.

Distinguishes a real demand drop from an instrumentation/logging artifact using at least two independent checks.. Show it clearly — with evidence a reviewer can point to.

Considers timezone, deploy timing, and tracking changes before accepting the dashboard at face value.

Considers timezone, deploy timing, and tracking changes before accepting the dashboard at face value.. Show it clearly — with evidence a reviewer can point to.

States a single most-likely root cause with an explicit confidence level and what would change their mind.

States a single most-likely root cause with an explicit confidence level and what would change their mind.. Show it clearly — with evidence a reviewer can point to.

Recommends a proportionate action (don't pause spend on an artifact; do escalate a real fire).

Recommends a proportionate action (don't pause spend on an artifact; do escalate a real fire).. Show it clearly — with evidence a reviewer can point to.

The room

3 autonomous AI coworkers, each with their own agenda. They won't all agree.

Priya Raman

Head of Growth

Wants: Wants to blame the paid channels and reallocate budget today; the drop validates a pet theory she's been pushing for weeks.

Style: Fast-talking, decisive, allergic to 'let me check the data first'.

Marco Velez

Senior Engineer (on-call)

Wants: Shipped a release at 11:40 PM and is defensive that it's not his deploy; would rather you rule out everything else first.

Style: Precise, a little prickly, respects evidence over vibes.

Dana Okafor

VP Product

Wants: Just wants a clear answer for the exec channel and hates hedged language; will push you to commit to a cause prematurely.

Style: Calm, impatient with ambiguity, rewards a crisp recommendation.

Your workspace

Real tools, pre-seeded with context. You're not roleplaying, you're working.

Code / IDE Docs / wiki Team chat

Scored on

Analytical rigorValidityCommunicationActionability

More in Data

A Dashboard the CEO Will Actually Act On

medium·40m·19 cr

The Winning Test That Lied

hard·45m·29 cr

The Numbers Don't Tie, and the Board Meets at Nine

expert·50m·39 cr

Who's Leaving, and the One Thing That Would Stop Them

medium·40m·19 cr