The Grandmother Test

Someone asked me for a dashboard and I can't work out what for

Ask one question before building anything: what will you do differently depending on the answer? Roughly 60% of analytics requests quietly disappear at that question, because there was no decision underneath. That question is the cheapest money anyone will ever save you.

Gergana Tyaneva · 21 September 2026 · 12 years in product and marketing analytics

The grandmother version

Someone asks you for a ladder. You could spend the afternoon finding a very good ladder.

Or you could ask what's on the roof.

Half the time it turns out the thing on the roof can be reached from the bedroom window, and nobody needed a ladder at all. The other half, you now know how long it has to be.

The question that does the work

"What will you do differently depending on the answer?" Write the answer down as a fork with two branches. If either branch is missing, or both branches lead to the same action, there is no decision and the work is not worth doing.

Expect a large share of requests to collapse here. That is the point. It is not obstruction; it is the cheapest possible result.

ProfitNew customersRepeat revenueCost to serveTrafficConversionFrequencyBasketSupportRefundsIf the thing they asked for doesn't hang off this tree, that is your answer.

Strip the deliverable language

A dashboard, a report, an update and a deck are deliverables, not goals. Nobody wants a deliverable. Restate the ask as the outcome: "you want to know whether to keep funding this channel next quarter."

Read your restatement back to them. If they correct you, you have just avoided building the wrong thing — for the price of one sentence.

Draw the tree

Sketch the metric tree: one output metric at the top — profit or revenue — branching into what actually moves it. Paper is fine.

Two things fall out. The request finds its place in the tree, or it visibly doesn't connect to anything, which answers the question. And everyone in the room gets a shared vocabulary, which is usually the real deliverable.

The short version

Goal, then decision, then metric, then a hypothesis you could be wrong about. In that order.

Most requests die at step two, and the ones that survive are worth doing properly.

This is step one of every engagement, and the 15-minute call that does it is free.

Book 15 minutes →

More answers

Attribution — Which channel actually pays?Plumbing — Our numbers live in eight tools and no two agreeTime — Month-end reporting eats three days and nobody reads itMoney — We know people churn. We don't know who, when, or what it's worth

One question before you read on. I'd like to count visits with Google Analytics, which sets cookies and sends your IP to Google. Nothing has loaded yet, and the site works exactly the same either way. What's collected.