Why technology bets go wrong: 30 years in the R&D chair

Christophe opens with the tension every R&D head knows: you are judged on the future while the present eats your calendar. When he started, a problem had a handful of possible routes. Today it can have hundreds, coming from startups, academia and suppliers, and each funding decision is a bet on which one keeps improving.

The middle of the conversation is about the bets that went wrong and the one that nearly didn’t happen. A health-claims programme ran for around ten years and tens of millions of euros before it was stopped. At the same time, a small startup brought stevia to Danone. It tasted of licorice, it didn’t fit the zero-sugar strategy, and company policy ruled out sweeteners in children’s products. Every reason to say no was defensible. Christophe’s team said yes, and a first product at 30% sugar reduction grew, generation by generation, into a category standard.

The last third turns to method. Christophe explains how measuring improvement rates took his teams from a hundred candidate technologies to the four or five worth examining, why he welcomed it when the data contradicted his own experts, and how to bring AI into an R&D department by starting with the people willing to try it. He closes with the one question he’d ask of any technology-intelligence process.