Jeremy Stevens has more than two decades in retail, with senior roles across Very, Holland & Barrett and Matalan Group. The episode is about what actually happens in a Monday morning trading meeting, and why it so often produces twenty actions and no decisions.
He starts with the mechanics. Monday morning exists to work out what happened last week so a senior group can make short-term calls against this period's numbers, and getting there means somebody triangulating internal data, external data and several providers into one coherent picture. He describes that as pretty stressful for the people pulling it together.
Then the line the episode is named after: all that triangulation gives you the what, not the why. The why has to be worked out from data, experience and logic, and in fashion it is never one variable — season, a social trend, a festival driving exposure to one item, and weather, which he calls insanely hard to predict.
His method for handling it is the most practical thing in the series so far. Bridge the movement out to three or four variables and weight them: which one explains half of last week, which explains a quarter, which explains the rest. Get to 80% of the answer and act, because 80% with experience on top is good enough for an in-week trading decision.
It is also his argument against generic AI. Point an off-the-shelf model at a specific brand in a specific season and you get a different answer every prompt, because the weighting comes from years inside that business. What would work is a trained structure of rules and causal links, built by people who have actually traded the category.
