i.
Peas & Honey
There’s an old rhyme about a man who eats his peas with honey — he’s done it all his life. It makes the peas taste funny, he admits, but it keeps them on the knife.
If you take his words at face value, the request is absurd: who wants honey-flavored peas? But he isn’t trying to enjoy peas. He’s trying to keep them on the knife. The honey is glue. Once you understand the real objective, the strange solution makes perfect sense — and you can finally offer him a better one. A spoon.
The lesson
Most requirements are misunderstood goals. Find the real objective and the strange solution makes sense.
ii.
The Pot Roast
A woman always cuts the ends off the pot roast before she puts it in the pan. Asked why, she says it’s how her mother did it. Her mother says it’s how her mother did it. They go and ask grandmother, who laughs: her roasting pan was simply too small for a full roast.
The reason vanished two generations ago. The ritual remained. Most processes are inherited constraints whose reason has long since disappeared — and nobody remembers the pan was too small.
The lesson
Question every inherited process. Ask whether the original reason still exists — usually it doesn’t.
iii.
The Horse & Carriage
It’s often said that if you’d asked people what they wanted before the automobile, they’d have asked for a faster horse. They couldn’t want the car — they had no language for it yet.
This is what asking customers “what do you want?” usually produces: a faster version of the thing they already know. Our work isn’t to deliver the faster horse. It’s to sit with people long enough to understand the journey they’re actually taking — and then help them imagine the car.
The lesson
Asked to imagine the future, most people describe a faster horse. The work is helping them see the car.
iv.
The Backup
An organization runs its backups religiously. Green checkmarks every night, dashboards full of success. The backup is flawless. Then the day comes when something is actually lost — and no one can restore it. The backups ran perfectly; they were never tested as recoveries.
They had optimized the convenient metric — “did the backup job complete?” — instead of the real objective: can we get our data back? The thing you measure is rarely the thing you want. Measure the recovery.
The lesson
Optimize the real objective, not the convenient metric. Perfect the recovery, not just the backup.
v.
Never Automate Waste
The fastest way to make a mess worse is to automate it. Take a tangled, redundant, half-understood process, wrap it in software, and you haven’t fixed anything — you’ve just made the wrong thing happen faster, at scale, with confidence.
So there’s an order to the work, and it never changes. First understand what’s really being done and why. Then simplify — remove the steps whose reason vanished. Then align the work to the actual objective. And only then, once the thing is worth doing, do you automate it.
The lesson
Understand → simplify → align → then automate. Never automate waste, bureaucracy, or confusion.