January 12, 2024

Automating drudgery

John Simmons looked down at the room full of young women working away at their adding machines, each passing the results of their calculation to the next person in the chain, and thought, ‘This is no job for a human being.’

The women were ‘computers’. And the way they worked was pin factory thinking applied to the collection and aggregation of data. The final result came not from a single person’s work, but from the entire chain.

The Lyons bakery and tearoom business was booming, and one of the secrets of their success was their ability to accurately predict demand, so that they never ran out yet wasted little. Tearoom managers phoned in the numbers every evening, those numbers were fed through the ‘computers’ into the bakery production schedules, then fed back to the managers – who always had the final say, since they knew their own shop and clientele best.

For years after this observation John Simmons and his team mapped out how the Lyons business worked at its best. Designing manual processes, testing them in real life and refining them until they worked like clockwork.

Then, when he learned about these new ‘computers’ – machines that could do the drudgery in minutes, he jumped at the chance to automate, persuaded Lyons to part fund EDSAC at Cambridge, and go on to build the world’s first business computer.

Lyons went on to become a producer not only of cakes, but also, for a while, mainframes.

This is a great story, and if you want to know more, you can read about it in ‘A computer called LEO’ by Georgina Ferry, but for me there are three key lessons for anyone looking at automation now:

1) Aim to reduce drudgery for humans, not create more of it.

2) Build and document an excellent manual system before you automate it.

3) Let the human have the last word.

Obvious, I know, but rarely done.

Discipline makes Daring possible.

Interestingly, the Post Office used a LEO computer until 1981. Perhaps they should have stuck with it.