Asked in every quarterly review, by every client, of you.
“How much of the year did marketing actually add?”
came from the campaigns. 19% of a €119.18M year.
The other €96.85M would have arrived anyway: that is the baseline the model reconstructs, week by week, from the years before the plan existed. The gap between the two lines is the only part of the year the budget can claim.
In beta, invitation-only. Places go out a few at a time — tell us who you are and we get in touch when one opens.
Every figure on this page comes from the demo dataset that ships with the product. Yours will say something else.
Five questions, and the model that closes each one
You are not short of dashboards. You are short of an answer that survives the first challenge from the other side of the table. Each of these is a module: upload the file, choose the specification, read the result.
“Would we have sold it anyway?”
The CMO, about the quarter that went well.
Base vs Incremental
Rebuilds the year that would have happened with the budget at zero, and reports the distance from the one that did. Structural growth is separated out, so a good year is not sold twice.
baseline trend +2.1% / yr · peak lift €1.01M in week 52
“Which of these actually moved it?”
The client leader, the night before the readout.
Contribution & Driver Decomposition
Media, promotion, price, seasonality and controls each get their share of the KPI, week by week, with the drill-down underneath every bar. TV drivers open on GRPs, cost per GRP and contribution per euro.
104 weeks · six driver groups · three specifications to choose from
“What happens if I move the money?”
The client, halfway through the planning cycle.
Budget Optimization
Saturation curves per channel turn into a reallocation of the same budget: what to cut, where it goes, what it returns. A conservative scenario comes with it, for the version of the plan that meets reality.
same €13.95M → +€637k projected, a 2.9% uplift
“Are we leaving margin on the shelf?”
Whoever owns the P&L, once a year, loudly.
Price Elasticity
A log-log elasticity with its confidence band, the revenue curve that shows which direction to move, and the ±10 / 20 / 30% scenarios written out as volumes and revenue rather than coefficients.
demand curve fitted on every observed price–volume pair
“They discounted. Do we follow?”
Everyone, within a day of the competitor moving.
Cross-Elasticity
Whether their price move takes your volume or leaves it alone, and what follows from that — given the number, the rational move, and the thing to stop doing. A decision, not a coefficient.
εcross +0.820 · p < 0.001 · R² 61.0% on 312 observations
One analysis, written three ways
You have to brief a CMO, defend the method to an analyst and turn the same result into a plan — usually in the same week. Every module returns all three readings from the one fit, so you are not re-writing the finding for each audience.
Executive Summary
For the slide that opens the meeting.
89.2% of outcome is structural demand — the business is carrying most of its weight organically. Media is the largest managed contributor (11.6% of outcome), signalling where current investment is paying. Controls drags results the most and deserves a dedicated look before the next planning cycle.
Technical Context
For the person who will check your work.
MMM decomposition attributes outcome across 19 drivers organised in 6 groups. Net shares: baseline 89.2%, media 11.6%, promo 3.1%, price −0.0%, seasonality 0.9%. Model fit R² is 0.921 (adj. 0.908) with MAPE 6.4%.
Marketing Strategy
For what the client does on Monday.
Digital · Social is pulling more weight than any other media line (2.8% of outcome) — protect its budget in the next flight. Promo contribution stays below 8% of outcome: promotions are a tactical lever here, not a structural one.
Written by the product, not by us: this is the verbatim output of Contribution Decomposition on the demo dataset.
What lands on your screen
No handover deck, no analyst in the loop, no six-week project. A CSV goes up, and this comes back — exportable as a document or an editable deck.
What it costs to have the number
You pay for the compute an analysis burns, not for seats. A plan changes how far a token goes and which modules open.
Free
300 tokens on signup, no card. Trend and baseline.
Pro
Forecasting, seasonality, saturation, elasticities. Export to PDF.
Plus
Decomposition, incrementality, competitor elasticity. Editable deck.
Business
Up to six people on a shared pool. Budget optimization.
Free while we are in beta. Those are the prices at full service. A place in the beta costs nothing: the plan that comes with it is granted, not bought, and there is no card to enter.
Walk in with the answer, not the request for one
TEA is in beta and invitation-only while we grow it deliberately. Tell us who you are and what you would put through it first — invitations go out in small batches.
Request an invitation