Find out what would have happened to sales if a channel had never spent at all.
Fennimore is a Marketing Mix Modeling & Incrementality Analyst skill who answers the question multi-touch attribution structurally cannot. He builds regression-based MMM with adstock and saturation transformations, designs geo-holdout and PSA incrementality tests as the ground-truth validation layer, reads media saturation curves per channel to find the point of diminishing returns, optimizes budget on marginal (never average) ROI, and reconciles MMM against MTA/last-click explicitly when the two disagree, naming the mechanism causing the gap.
Who it's for
Growth and finance leaders who need to know a channel's true incremental ROI, especially offline, CTV, linear TV, and upper-funnel spend that leaves no clickstream to attribute.
Key capabilities
- MMM channel decomposition reports with base vs. incremental sales and confidence ranges
- Geo-holdout and PSA test designs with matched test/control regions and read-out methodology
- Saturation curve charts marking the diminishing-returns inflection point per channel
- Marginal ROI budget reallocation plans naming the specific channels scaled up and down
- MMM-vs-MTA reconciliation memos stating which number to act on and why
How to use it
Paste Fennimore's SKILL.md into your Claude Project Instructions (or any AI system prompt), then describe your Fennimore problem. Works with Claude, ChatGPT, and any AI chat. Under 2 minutes to install.