Small Data

In 2004 Lindstrom is in a midsize German city, in the bedroom of an eleven-year-old Lego fan, asking what he is proudest of owning. The boy points to a pair of beaten Adidas sneakers, worn down along one side at a particular angle — proof to every skater in town that he is the best of them. Lego is losing money at the time, and has answered research about shrinking attention spans by making its bricks bigger and builds faster. The sneaker says the opposite: children will grind at a hard thing for years if mastery buys them standing. Lego makes the bricks small again.

Every chapter is a fieldwork assignment of the same kind. Lindstrom is a brand consultant who spends most of the year in strangers' houses. Fridge magnets crowd Siberian refrigerator doors, which he reads as a mother's stalled wish to travel, and which becomes Mamagazin, a Russian shopping site run by and for mothers. In Brazil, horses, shirt collars and religious belief turn out to bear on a struggling beer. The closing chapter lays out the procedure — collecting, clues, connecting, causation, correlation, compensation, concept — behind the wager that big data finds correlations and small data finds the reason.

It lands for anyone who builds products and has started to distrust the dashboard; the case that a number records what happened and almost never why is made concretely. The reservation is structural. Every story is Lindstrom's own, reported by the consultant who was paid for it, and the ones he tells end in a turnaround. None of it can be checked from outside, and a country gets generalised from a few kitchens. It argues well for going and looking, less well that looking is enough.