The Pre-MMM Data Readiness Audit
Will your data carry a media mix model?
Most MMMs fail on inputs, not math. Upload spend and KPI history, confirm the column mapping, answer a short questionnaire. You get a 17-check readiness score with the evidence behind every gap and the fix for each.
We sign you in with an emailed link, no passwords. Free, and your raw file auto-deletes after 30 days.
Upload to score in about five minutes
How it works
01
Upload your history
One xlsx or csv: a date column, channel spend columns, and your KPI. Daily, weekly, or monthly, up to 20 MB.
02
Confirm the mapping
The engine parses the file and suggests a role for every column. You confirm dates, KPI, channels, controls, and geo.
03
Answer what the file cannot show
A short questionnaire on promos, KPI definition, and experiments. Only relevant questions appear, under two minutes.
04
Get your score by email
17 checks run against published thresholds. The scored report lands in your inbox, typically within a few minutes.
Published thresholds, deterministic scoring
What the 17 checks cover
Four phases, ordered the way gaps actually bite. The method and every threshold come from the QuantVibe guide, so the same file always earns the same score.
Phase 1
Signal in the spend history
Does spend vary enough, and independently enough, for a model to read anything from it?
- 01Spend Variance
- 02Independent Variation Across Channels
- 03Channel Concentration
- 04Natural Experiments in the History
- 05Statistical Power and Minimum Detectable Effect
- 06Per-Channel Materiality Floor
Phase 2
Separating media from everything else
Can media effects be told apart from promos, pricing, seasonality, and shocks?
- 07Media x Non-Media Confounding
- 08Demand-Seasonality Confound
- 09Control-Variable Coverage
- 10Target Regime Shifts and Systemic Shocks
Phase 3
The data foundation
Is the dataset itself clean, deep, and at the right granularity?
- 11Data Quality and Completeness
- 12Historical Depth and Stability
- 13KPI Definition and Scope
- 14Execution Granularity vs. Data Granularity
- 15Disaggregated and Geo Data Availability
Phase 4
Ground truth and priors
Can the model be anchored to experiments and informed priors?
- 16Priors for Long-Carryover Channels
- 17Experiment and Incrementality Availability
0 to 17, three bands
How to read your score
15 to 17
Ready
Proceed. Hold the finished model to a separate diagnostics bar when results land.
11 to 14
Fixable gaps
Close the Phase 1 and Phase 2 items first. Most fixes are cheaper than a confused model.
10 or under
Not yet
Use incrementality tests for the questions that cannot wait. Fix collection gaps and revisit in a quarter.
Find the gaps before your model does.
17 checks, about five minutes, free. Fix the cheap problems before they become expensive ones.