How I turned engineers into delivery couriers

The first time I consciously manipulated motivation, I was paying couriers for speed.

Kuper, food delivery, 2021. I'm 23. Built a feature: "Deliver in 30 minutes, get 150₽. In 20 minutes, 220₽." Simple model. Do X, get Y.

It worked. Average delivery time dropped. Metrics green. Unit economics solid. I was proud of myself. (I didn't track traffic accidents at the time. Shame on me.)

The conclusion I drew: motivation = right incentives. Want results, pay for them. If it works for couriers, it works for everyone.

I carried that belief into technology products.

A few years later I was managing a team of PMs and engineers, smart people solving complex problems every day. Same playbook: "Hit your OKR, get a big bonus at review." "Ship the feature by date X, get an extra payment."

Didn't work. Worse: people started optimizing for the metric instead of the outcome. They stopped thinking. Started cutting corners. Gamed the system.

I thought I just hadn't made the carrot big enough. So I raised the stakes.

It got worse.

Then I found research that explained what I'd been doing wrong:

Sam Glucksberg's candle experiment. Participants need to attach a candle to a wall without dripping wax: non-obvious solution requiring creative thinking. One group gets promised $20 for a fast answer. The other: nothing, just solve it. The group with money took 3.5 minutes longer. The reward narrowed their focus. Instead of exploring, they tried to earn and got stuck. (When the same task was made algorithmic with an obvious solution, money helped. The rewarded group was faster.)

Dan Ariely replicated this in India across reward levels from daily wage to multiples of it. Result: the bigger the reward, the worse the performance on cognitive tasks. In 8 of 9 thinking-heavy tasks, more money meant worse outcomes.

Here's the actual distinction:

Mechanistic work (courier: pick up, deliver): no intrinsic motivation to destroy. Path is clear. Pay more, they go faster. The carrot works perfectly.

Creative work (PM, engineer, designer): solution is non-obvious. Requires genuine thinking. "Do X, get Y" compresses that thinking space and converts an interesting problem into a mechanical one. The brilliant people you spent months recruiting quietly turn into delivery workers.

TL;DR

I spent years trying to motivate engineers and PMs the same way I motivated couriers. It cost me burned-out people and failed projects. The carrot is a real tool — it just breaks what it wasn't designed for.

For routine work: pay for results. For creative work: give hard problems and get out of the way.

Stay tuned 🥷🥷🥷