Campaign Funnel & ROI Analysis
This challenge runs as a locked, multi-stage workflow simulation. You submit one stage at a time. Once submitted, a stage becomes read-only and later stages build on your earlier decisions.
Choose what to investigate first, what metric you trust least, and what workflow path you’ll use.
Required inputs
- •What channel/issue do you investigate first?
- •What metric do you trust least (and why)?
- •Tool/workflow path you will use
Ingest data and submit your first budget leaning, anomaly hypothesis, and proof that the workflow ran.
Required inputs
- •First budget leaning (increase/hold/decrease and where)
- •Anomaly hypothesis
- •Proof note (what would prove your workflow ran?)
New constraints arrive. Revise recommendation and explain what changes due to the constraint.
Required inputs
- •Revised budget recommendation
- •Constraint impact explanation
- •One-action decision you would take now
Reconcile contradictions, manually recompute one KPI, and describe what AI output you’d verify.
Required inputs
- •Reconcile one attribution contradiction
- •Manual recomputation of one KPI
- •What AI output would be risky to trust without validation?
Submit your weekly recommendation, one-action decision, and a mini experiment/validation plan with a publishable summary.
Required inputs
- •Weekly budget recommendation
- •One-action decision
- •Mini experiment plan or validation plan
- •Published summary or structured tracker output
What gets graded (later)
The backend grader will score workflow quality across stages: decision quality, verification, consistency, and the clarity of your final artifact.
- 1.Initial Read
- 2.First Update Workflow
- 3.Constraint Injection
- 4.Verification Loop
- 5.Final Artifact
- 6.KPI tracker
- 7.Anomaly tracker
- 8.Weekly recommendation output
- 9.Proof-of-work callback note (conceptual for now)
- 10.Session token field (conceptual for now)
- •Choose what to investigate before all constraints are known (Stage 1)
- •Submit a first budget leaning + anomaly hypothesis (Stage 2)
- •Revise recommendation after constraints change (Stage 3)
- •Manual KPI check + AI-risk checkpoint (Stage 4)
- •Final weekly recommendation + validation plan artifact (Stage 5)
Dataset access stays locked until you start the workflow session (timer start).
Open Stage 1 to start the timer and unlock the first dataset.
Available Before Stage 1
0 filesOnly the initial dataset is shown here. Additional files unlock inside the workflow as you move through stages.
AI Tools
Analysis Tools
Correct ROAS and CVR calculations per channel
Finds Display Black Friday spike and explains it
Specific reallocation % justified with data
CMO briefing is professional and data-driven