Something changes. A container is stuck, a launch slips six weeks, a large account cuts their order in half. Someone asks what it means for the year.

The analysis comes back a fortnight later. The decision was made in week one.

01 / 04

The decision does not wait for you

Operational decisions have a window. A supplier will hold capacity for a few days. A media booking has a deadline. Air freight stops being an option once the ship has sailed.

Inside that window somebody has to choose. If the numbers are not there they choose on experience, and experienced people are often right, which is why this problem stays hidden for years. The cost sits in the cases where they were wrong and nobody could see it in time.

Scenario work is only worth anything if it arrives before the window closes.

02 / 04

Why it takes two weeks

The delay is almost never the modelling itself. It is the assembly.

Someone pulls sales history out of one system, the order book from another, costs from a third, and the plan from a spreadsheet. Then the assumptions get rebuilt, because the last version was built for a different question and nobody remembers what went into it.

Then it gets checked, because a number that surprises people always gets challenged. The checking takes about as long as the building.

Every one of those steps is a data problem.

03 / 04

Build the model once

When the numbers already sit in one place with agreed definitions, a scenario becomes a change of input.

Move the launch six weeks and look at what happens to the season, the cash phasing and the inventory position. Take out a container and see which products go short, which orders are at risk, and what air freight would cost against the margin it protects.

The answer comes back the same day. It uses the same numbers as the board pack, so the meeting is not spent arguing about where it came from.

04 / 04

Keep the common ones ready

Most businesses face the same handful of shocks. Those are worth building once and leaving in place.

A supplier delay at four, eight and twelve weeks. A major account cancelling. A launch moving. A currency move of ten percent. A market opening later than planned. A discount at three different depths.

When they are ready in advance, the meeting starts from what happens when rather than what happens if, and gets to the decision faster.

What changes is how many decisions get made with something behind them.

Most companies only model the large questions, because modelling is expensive and slow. When it takes an hour, the ordinary questions get modelled as well. Those are the ones that come up every week.