ea

“Can you show me what the TV actually did?”

You, on the call, for the third quarter running.

“We can pull the campaign report and walk you through it.”

Them, also for the third quarter running.

You did not ask for a report. You asked a question with a number in it. The number is in your own data — your sales, your spend, your prices — and getting it out takes an afternoon and a CSV, not a supplier.

Request an invitation

In beta, invitation-only. Places go out a few at a time — tell us who you are and we get in touch.

Five questions you have already asked

On the left, roughly what came back. On the right, what the same question looks like once it has been answered with a model instead of a deck.

“How much did the campaign actually add?”

What came back

“Reach was up 14% year on year and engagement is well above benchmark.”

What an answer is

€22.33M of a €119.18M year came from the campaigns — 19%. The other €96.85M would have arrived with the budget at zero, and the model draws that year out week by week so you can see where the gap opens.

base vs incremental

“What if I move money out of TV?”

What came back

“We’d have to rebuild the plan and come back to you with scenarios.”

What an answer is

The same €13.95M, split differently, projects +€637k — a 2.9% uplift. Named channel, named cut: the line returning 0.85× on its next euro loses about a third of its budget to the two returning 2.67×.

budget optimization

“Which of the things we did actually worked?”

What came back

“It all works together — you can’t really separate the channels.”

What an answer is

Media 11.6% of outcome, promo 3.1%, seasonality 0.9% — and 89.2% that is simply your business, arriving without help. Week by week, with the detail behind every bar.

contribution decomposition

“Is my price wrong?”

What came back

“That’s a pricing question — not really our side of the business.”

What an answer is

Demand is elastic at −1.62: every 1% on the price costs about 1.6% of volume. The revenue curve says which direction to move, and the ±10 / 20 / 30% scenarios are written out in euros before you commit to any of them.

price elasticity

“They discounted. Do I follow?”

What came back

“Let’s monitor it for a few weeks and regroup.”

What an answer is

You are substitutes, and it is measured: εcross +0.820 at p < 0.001. A 10% move on their price is worth roughly 8.2% of your volume, which makes their discount a volume risk to defend against — not a price to match.

cross-elasticity

Every figure on the right is a real result from the demo dataset that ships with the product. Yours will say something else — that is the point.

You already own everything it needs

No tags to implement, no platform to connect, no data to request from anyone. Four things you can export this afternoon, in one spreadsheet.

What you sold

Revenue or volume, by week or by month. Two years is plenty; three is better.

from your own sales reporting

What you spent

Spend or GRPs per channel, over the same weeks. The plan you already sign off.

from the media plan, post-evaluation

What you charged

Average price or price index. If you know your competitor’s, bring that too.

from pricing or category reporting

When you promoted

Depth and timing of promotions, so their lift stops being credited to the advertising.

from the trade calendar

What you bring to the next review

Not a request for analysis. The analysis — as a document, or as an editable deck you can put straight into the meeting.

Base vs Incremental: 22.33 million euro came from marketing, with actual revenue against the reconstructed baseline
The year without the campaigns, drawn underneath the year you had.
Cross-elasticity result: competitor elasticity 0.820, with the rational move and what to discard
The competitor question, closed: the number, the move it implies, and the thing to stop doing.

This is not a way to go around your agency. Most of them would run this work if the budget and the weeks existed, and the good ones will read the output and argue with it, which is exactly what you want them to do.

What changes is who arrives with the number. When you open the meeting with the baseline, the incremental share and the reallocation already on the table, the conversation stops being about whether it worked and starts being about what to do next.

An afternoon, start to finish

You do not need to know what a regression is. You need to know what you are asking, and the product handles the rest — including telling you when your data is not good enough to answer it.

  1. Put the file in

    One spreadsheet, or start from a sample of your own industry to see the shape first. Columns are proposed and you confirm them.

  2. Pick the question

    Twelve of them, in plain language. Where there is a choice of method, each is explained with what it costs you — nothing is decided behind your back.

  3. Read the answer

    Minutes, not weeks. Written three ways: the summary for the room, the method for whoever checks it, and what to do about it.

  • €0Free. 300 tokens on signup, no card — enough to see the thing work.
  • €19.99/moPro. Forecasting, seasonality, elasticities, export to PDF.
  • €59.99/moPlus. Incrementality, decomposition, competitor elasticity, editable deck.

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.

Ask the question once more — with the answer in your hand

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