Streams progressively — you don’t need to wait for the whole file. You can also download it to listen offline.
Ask a seller who just ran out of stock what went wrong and you’ll usually hear a version of the same answer: the forecast was off. So they go looking for a better forecast — a smarter formula, more history, a tighter model.
Then it happens again.
This episode argues that the search for a better formula is usually the wrong search. Forecasting error is real, but for most sellers it isn’t the binding constraint. The binding constraint is the decision the forecast feeds into: how much cash is available, how long the lead time is, how much a stockout actually costs relative to the cost of holding too much, and how tolerable being wrong in each direction really is.
A more accurate number changes nothing if the decision it feeds was never sensitive to accuracy in the first place. Find the constraint before you refine the arithmetic.
The written companion
This episode covers the same ground as one of the written pieces in this section, in conversational form. If you prefer to read the argument, or want it laid out step by step:
- Inventory forecasting isn’t about finding the perfect formula — the same thesis, written, with the reasoning set out in sequence.
Where to go next
Once you accept the decision framing, the mechanics still matter. These three deal with them:
- The reorder point formula — the baseline calculation and what each input is really doing.
- Why the classic formula is wrong for most sellers — where the standard version quietly breaks.
- The mistakes that quietly destroy margin — the failure patterns behind most stockouts.
- How AI and customer interviews predict market demand — the companion episode, on forecasting when no history exists.