The conventional narrative about Barnes & Noble's trajectory was settled by 2019: Amazon had won, physical bookstores were doomed, and the chain's bankruptcy was a matter of timing, not likelihood.

The narrative turned out to be wrong. Under the leadership of James Daunt — who had previously reversed the decline of the UK chain Waterstones using unconventional methods — Barnes & Noble has not just survived but returned to growth. The method deserves examination.

The problem with Barnes & Noble's previous incarnation, in Daunt's analysis, was that it had stopped functioning as a bookshop and had become a retail machine optimised for publisher relationships. The bestseller tables at the front were populated not by the books staff would recommend but by books whose publishers had paid for the placement. The result was a shopping experience that told customers what the publishing industry wanted to sell them, not what the store's readers thought was worth buying.

Amazon does this better. When your distinction from Amazon is eliminated, you are competing on a dimension (price, range, convenience) where you cannot win.

Daunt's solution was radical decentralisation. Each Barnes & Noble store was given the authority to select its own inventory, based on the tastes and knowledge of the staff who worked there and the readers who visited.

The large publishers' promotional payments — which had shaped the store experience for decades — were eliminated. The space previously allocated to paid-for prominence was reallocated to staff-selected recommendations.

The result: a different store in each location, shaped by local tastes, with a character that a purely algorithmic recommendation engine cannot replicate.

The irony of the turnaround is that the more Barnes & Noble resembles a good independent bookshop — with genuine staff expertise, local character, and idiosyncratic recommendations — the less it resembles Amazon, and the more it offers something Amazon fundamentally cannot.

Amazon's recommendation algorithm is better at predicting what you might like based on what you have bought before. A bookshop staffed by genuine readers is better at recommending what you would love based on a conversation about what moves you.

The Barnes & Noble case is not specifically about retail. It is about the strategic error of competing against a dominant player on the dimension where that player has an absolute advantage.

The correct response to Amazon's dominance in price, range, and algorithmic recommendation is not to improve on those dimensions. It is to identify the dimension on which Amazon is structurally weak — genuine human expertise and local character — and to dominate it.

Barnes & Noble survived by becoming less like what it had tried to be and more like what an independent bookshop is. The lesson is applicable to any business displaced by a dominant platform: compete on the dimensions the platform cannot replicate, not on the dimensions it does better.