Will my geo test work?

Answer what you know and this will tell you whether a geo holdout is available on your footprint, which constraint is the one that binds, and which design fits instead if geo does not. Nothing is stored and nothing is sent anywhere.

What this answers

Whether a test is possible, not what it would find. It will never predict a lift, because no honest instrument can do that before the test runs.

Most of what decides this is structural. The number of markets you can observe revenue in sets a ceiling on statistical significance that no amount of budget moves, and that is usually the first thing worth knowing.

If geo is not available to you, that is not the end of the conversation. There are other designs, and this will tell you which one your situation actually points at.

Written from the same feasibility floors we use on client work. If you want the reasoning rather than the verdict, the notes cover it at length.

Your footprint

The outcome

Can you see revenue broken out by geographic market, in your own system?

Wholesale and marketplace businesses usually cannot.

The footprint

How many markets can you observe separately? DMAs, metros, states, whichever you actually report on.

How often can you observe that revenue?

How many months of that history do you have?

How separable is your media by market?

Broadcast and loosely targeted CTV spill across market lines.

What significance convention does your finance function use? This sets how many markets you need. It is not a detail.

This tool needs JavaScript. The reasoning behind it is in the notes, which do not.