Last year I sold out in November. The year before I marked down 40% of my stock in January. There has to be a smarter way.
The situation
Marcus runs a boutique clothing store and places a single seasonal order for winter coats in August, before he knows what demand will look like. Too few units means stockouts by November and lost full-margin sales. Too many means January markdowns at 40% off, eating margin he can't recover.
The analysis
When the cost of running out exceeds the cost of excess, you should order above your expected demand — not at it.
His current approach: order the same quantity as last year, maybe plus 10% if he's feeling optimistic. This is a reasonable heuristic, but it ignores the asymmetry of the two failure modes. A stockout costs him $62/unit in lost gross profit. A January clearance dump costs him $18/unit below cost. Those are not symmetric — and that asymmetry should change the order quantity.
The Spreadsheet Sim lets him model demand as a distribution rather than a single number. He estimates demand is roughly normally distributed with a mean of 115 units and a standard deviation of 22 (based on the last three years: 98, 119, 128 units sold). He sets his order quantity as the decision variable and runs 5,000 simulations.
At 115 units (his old approach), he has a 50% chance of a stockout and a 50% chance of excess. But because stockouts hurt over 3× more than clearance losses, the profit-maximizing order quantity is 132 units — about 15% higher than his instinct. The dollar improvement in expected profit is modest, but the risk profile shifts: stockout probability drops from 50% to 22%.
The insight: when the cost of running out exceeds the cost of excess, you should order above your expected demand. The model makes the right number concrete.
The decision
Marcus orders 132 units and saves the fitted demand distribution to his library, with Monitoring turned on against his weekly POS feed. Three weeks into the season, Monitoring flags that actual sell-through is running above the top of his fitted range — a cold snap pulled demand forward. He re-fits before Black Friday instead of after, and places a small top-up order while there's still lead time to get it. No more finding out he's short in the last week of November.
Try it yourself
Don't have this data yet? Here's where to start.
Pull your sales history from your POS for the last 3 seasons. You only need two numbers per season: how many units you ordered and how many you sold before you started marking down. That's enough to estimate a mean and a rough range — which is all the model needs. Even rough estimates reveal surprising insights.
Try this in the app
Open Spreadsheet Sim with this scenario pre-loaded and ready to explore.