|Home||Back to Index|
Carrots & Sticks: What Incentives Really Work?
The stick causes compliance - people pretend to do the new thing when they think they’re being watched.
Hypothesis - controlled experiment - result
We don’t do that in software development. A pilot or a trial isn’t an experiment.
Pavlov - animals could be trained to anticipate rewards on the basis of cues.
Carrots & sticks elicit desired behavior.
Taylor - Scientific Management
Most people did algorithmic work - followed a specific process (assembly line)
Consider the context.
Incentivize & monitor people.
We no longer work that way. Sometimes the incentives will backfire.
There may not be a “best” way to perform a task.
Try doing nothing. Daydream. Don’t feel guilty for it.
That part of your brain is used for creativity and innovation.
We even incentivize ourselves. “I’ll finish this and then get that latte.”
We’re still buying in to Taylorism.
The data shows you’re more efficient if you take a break.
Harry Harlow - Intrinsic Motivation
The performance of the task provided its own reward.
Giving rewards actually decreased motivation.
Do it just for the joy of it. For the fun of it. Because we love it.
Incentives can get in the way of this - if we’re getting rewards, it must be something we don’t want to do.
Kahneman & Tversky - Prospect Theory, Behavioral Economics
Classical Economists are based around rational decision-making.
K&T say that we’re not rational. (See Ultimatum Game - people will prefer nothing to something if they don’t consider the rewards “fair.")
Pizza Hut promised free pizza to children to get them to read. Children were likely to pick books that are short and simple. You’re rewarding the number of books that children can get through in a short time. You’re getting the behavior that you incentivize. You end up with fat kids who don’t like to read.
Cobra Effect - bounty on cobra skins caused people to breed cobras.
Freakonomics - someone more clever than you will figure out a way to beat your incentive scheme.
Do you want to reward developers for finding bugs? ;)
Dan Ariely - experiments with small, medium, & large bonuses showed negative correlation with performance of complex tasks. The largest bonus was the least effective.
Simple, mechanical tasks show positive results with incentives.
They also work if the incentive and the result are close.
Loss Aversion also works.
Give the bonus beforehand; if you don’t meet the target, we’ll take it away.
The altruistic center (which also likes to work on teams/cooperate) cannot be activated at the same time as the pleasure center. Pleasure is much stronger, so when rewards are on the line, people are less helpful toward others.
Good management involves giving people what they need to do their jobs.
See “Pygmalion in Management” - people behave as managers expect them to.
Michael Feathers: “Catch them doing something right.”
Katzenbach & Khan - studies of elite military units
These people love the other people on their team. That’s their incentive.
Even animals need a sense of purpose.
Book - The Progress Principle
Making progress in meaningful work is the most important thing.