For marketing leads, agencies and analytics teams

Econometrics, without the six weeks and the supplier.

Marketing mix modelling, incrementality, price and promo elasticity, budget allocation. Fifteen analyses that fit on a CSV of your own sales, spend and prices, and you run them yourself, in an afternoon.

Request an invitation

In beta, invitation-only. Places go out a few at a time, and the first tokens are on us.

15analyses, twelve of them out of beta
1file: dates, a number, and whatever drove it
3readings of every result: executive, technical, marketing
0lines of code, and a number you can defend

What lands on your screen

The product, not a picture of one

Six screens from the running platform, on demo data anyone can download. Nothing here is drawn by hand.

The catalogue. Choose the question, upload a CSV, and the module runs the econometrics. The result comes back with its diagnostics, not just its headline.
Geo / Segment MMM
Contribution Decomposition
Budget Optimization
Competitive Pressure
Scenario Simulation

What it answers

Five questions, and the model that closes each one

Every dashboard gives you a different answer. This gives you one number, from your own data.

These are not features. They are the things somebody asks you in a review, and the analysis that answers them with a number instead of a deck.

“How much of the year did marketing actually add?”

Base vs IncrementalThe revenue the campaigns created, separated from the demand that was arriving anyway, with the counterfactual year reconstructed underneath it.

“Which of these actually moved it?”

Contribution DecompositionMedia, promo, price, seasonality and controls, each given its share of the KPI across the window, with the drill-down behind every bar.

“What happens if I move the money?”

Budget OptimizationSaturation curves per channel become a reallocation: the same budget, the projected uplift, and the best split at 20% less and 20% more budget.

“Are we leaving margin on the shelf?”

Price ElasticityHow demand answers a price move, with the confidence band, the revenue curve, and the scenarios at ten, twenty and thirty per cent spelled out.

“They discounted. Do we follow?”

Cross-ElasticityWhether their move takes your volume or leaves it alone, stated as a decision: what to do, and what to stop doing.

The catalogue

Fifteen analyses, in six families

Every one of them runs on a file you already have. Two run free on any account; the others need the compute of a higher plan.

Brand & Performance

Brand equity measurement and competitive pressure analysis.

How it works

An afternoon, start to finish

No integration, no taxonomy project, no data warehouse. The whole thing begins with a file and ends with a number you can defend.

Step one

Bring the file

A CSV or an Excel sheet: a date column, the number you care about, and the columns that drove it. Spend, prices, promotions, whatever you have. A template and six sample markets are there if you want to see it work first.

Step two

Pick the question

The module tells you which models fit your data and argues both sides of the choice. It fits them on the server, checks the assumptions, and says so out loud when a series is too short to carry the claim.

Step three

Take the answer out

Every result is written three ways: executive, technical, marketing. The same numbers, three registers, so the slide and the appendix say the same thing. Out as a PDF or a deck when you need one.

What the file has to contain

  • A date column. Daily, weekly or monthly, as long as it is regular.
  • The number you are explaining. Revenue, volume, sessions, subscriptions.
  • The columns that might explain it. Spend by channel, price, promotions, distribution, weather, whatever you suspect.
  • Enough rows. Around a hundred points for anything with a seasonal claim in it, and the product says so before you pay for the run.

What happens to it afterwards

  • It is deleted three hours after upload, whether you finished the analysis or not.
  • It is never used to train anything, and it never leaves the analysis that asked for it.
  • The demos do not touch your data, and running one costs you nothing.
  • Every model is written down: the specification, the coefficients, the residuals and the assumptions checked, for whoever wants to read them before trusting the number.

Read the specification sheet

What it costs

You pay for compute, not for seats

An analysis costs what it takes to compute, counted in CT and measured by the server. No subscription to a shelf of features you are not using, and the demos never count.

Free1 CT300 ms of compute
Pro1 CT450 ms of compute
Plus1 CT750 ms of compute
Business1 CT1500 ms of compute

Same token, more compute as the plan goes up, and every movement of the balance says where it came from. How billing works

The number is already in your data.

Getting it out takes an afternoon and a CSV. tea is in beta and invitation-only: tell us who you are and what you measure, and we get in touch.

Learning

A report says what happened. An analysis says why.

Most of what lands on a marketing desk is reporting. Useful, and a different job. The blog tells one case a day where the difference decides the answer.

Reporting

Sales were up 11% in April, TV ran six weeks. What happened, by channel, side by side.

Analysis

4.2 of those points came from TV, 5.1 from Easter moving. What caused it, and what the next euro does.

Read the blog