Unilever bred nozzles after calculation failed

Unilever's detergent manufacturing process forced boiling, high-pressure chemical slurry through a spray nozzle to produce detergent granules, but the nozzles kept clogging and produced inconsistent granule sizes, a persistent production problem. The company's first approach was the obvious one: bring in fluid-dynamics mathematicians to calculate the theoretically optimal nozzle geometry from physical first principles.

That expert-calculation approach failed — the mathematically derived optimal design still clogged in practice, because the real fluid dynamics inside the nozzle were too complex to fully capture in a tractable calculation. Rather than continue refining the theoretical model, Unilever handed the same problem to a different kind of expert: biologists, who approached it with no equations at all.

The biologists treated nozzle design as an evolutionary process rather than a calculation: starting from an existing nozzle, they generated ten random variations, tested all ten against the real clogging and granule-consistency problem, kept whichever variant performed best, generated ten new random variations of that winner, and repeated the cycle for 45 generations.

Bred nozzles instead: generate ten random variants, test against real clogging, keep the best, repeat 45 generations, reaching a solution no one could explain.

Mathematical models couldn't capture real fluid dynamics, so calculated designs still clogged.

Testing random variants against the actual problem gave direct empirical feedback.

Iterative selection and mutation converged on a design no calculation could derive.

The process didn't require understanding why the winning design worked, only measuring its performance.

When first-principles calculation fails on a complex problem, switch to evolutionary search: generate variants, test against reality, select the best, repeat.

After 45 generations and 449 failed variants, the process converged on a nozzle design over 100 times more efficient than the original — reliably avoiding clogging and producing consistent granules — even though the resulting shape was strange enough that no one on the team could fully explain in physical terms why it worked better than the mathematically 'optimal' design the fluid dynamicists had calculated.