I was working with a head of supply chain planning. Their company had been running S&OP the same way for years.
Single point forecast. Defend the number. Argue with finance about inventory. Repeat.
Then they changed how they showed up.
Instead of a point forecast to defend, they brought a range of scenarios into the next S&OP.
Three demand outcomes: with the inventory position, cost to serve, and cash exposure mapped across each one.
And not at the top line. At SKU level — which products were exposed in each scenario, and by how much.
Including what happened to inventory if demand came in 20% below plan.
The CFO had not seen it framed that way before. They went quiet for a moment. Then asked: “If demand misses by that much, what are we looking at in write-downs? And how quickly can we adjust?”
The team could answer it. Not with a rough estimate. They could show exactly which product lines were sitting on excess in the downside scenario, where the write-down risk was concentrated, and what the cash exposure looked like before the quarter closed.
That question had never come up in that room before. Not because the CFO was incapable of asking it. Because the planning team had never put the downside inventory exposure on the table at that level — only the aggregate forecast they were defending.
The argument about inventory did not happen that quarter.
What changed was not the data. It was that the planning team stopped defending a single number and started showing the range — including what it costs the business, by product, when the forecast is wrong on the low side.
That is what probabilistic modelling actually does in practice. Not a better forecast. A different conversation.
When the supply chain walks into S&OP with inventory and cash implications mapped by scenario and by SKU, the question shifts.
Finance stops asking “why do you need more stock?” and starts asking “what’s our exposure if we’re wrong, and what are we going to do about it?”
That is a harder question.
And it is the first question that leads somewhere useful.