Badly designed incentives really do backfire. The ones that work measure what each one returns, so you catch gaming early and cut what doesn't pay.
The worry about incentives is fair. Put a reward on the wrong number and people will chase it right past the result you actually wanted. There's real research behind that fear, and a few famous, expensive examples of it playing out.
So yes, incentives backfire. They backfire when nobody is watching what the reward buys. A reward you set and forget is a check written against a behavior you never measured, and you tend to find out it went wrong once the rework piles up or the numbers stop matching the profit.
The incentives that work aren't gentler or smaller. They're aimed at a specific behavior you can watch, so the day one starts pulling the wrong way, you catch it and shut it off. That's the difference between spending on incentives and hoping they help, and spending on them and measuring what each one returns.
When the reward pays for what you'd get anyway
The oldest objection has teeth. Pay people for something they already do, and you can dull the reason they did it in the first place. Psychologists call it the overjustification effect, and decades of experiments back it up: reward an activity someone already enjoys, and their own motivation to do it can drop.
That objection skips where the effect actually shows up. It's strongest on work people already find interesting, and for work they don't enjoy, a reward tends to raise effort rather than lower it. The Saturday night shift nobody wants isn't the labor of love the research is warning you about.
On the frontline the waste is quieter. You put an attendance bonus on the whole floor, and most of it goes to the reliable people who were already showing up. You spent the budget and moved nothing, because you paid full price for behavior you were getting for free.
When people chase the number instead of the result
Give people one number to hit and a real reason to hit it, and some will hit the number without doing the thing it stood for. Once a measure becomes a target, it stops being a good measure of much.
Wells Fargo is the case everyone remembers. Under sales targets tied to their pay, employees opened around two million accounts customers never asked for. The cross-sell metric looked incredible right up until the profit behind it turned out to be fiction.
It happens in smaller ways on any floor. An upsell rate climbs because associates ring up add-ons that come back as returns next week. A call-handling target gets hit by cutting customers off, so a second rep handles the callback and the work gets done twice. You pay out on a number that never reaches the bottom line.
When the target itself is wrong
Sometimes the incentive is honest and nobody games it, and it still costs you, because it was aimed at the wrong thing. The reward does its job, but the job was pointed at a number that doesn't carry the whole picture.
Reward speed on the line and you can get scrap and rework. Reward units shipped and quality escapes climb. The throughput gain lands on one report, and the cost of it lands on another that nobody set next to the first.
What the incentives that work have in common
Line the three failures up and they share one root. In every case nobody could see the incentive going wrong while it ran, because the spend was never tied to a result anyone checked.
The fix is three things about the target you aim it at.
- Specific. One behavior, one shift or role, not the whole floor at once. A narrow target is one you can actually watch, and one that doesn't quietly pay people for what they'd have done anyway.
- Visible. You see it move day by day against the data your operation already keeps, from the POS, the schedule, the systems finance trusts. Gaming shows up as a metric that climbs while the profit behind it sits still, and you only catch that if you're watching both.
- Measured. Every reward weighed against the outcome it was meant to buy, so the one that returns more than it costs gets widened and the one that doesn't gets cut. That's when you can put a number on what a reward actually returned instead of guessing.
This is also why dressing the floor up as a game rarely moves much on its own. Badges and leaderboards with no measured outcome under them are the version that wears off in a month. What makes an incentive work is the target you point it at and your ability to measure it, which is why treating incentives as something you engineer beats running them on a hunch.
An incentive that backfires and one that pays look identical the day you turn them on. What separates them is whether you can see what each one is doing while it runs. Aim it at one measurable behavior, watch it against the numbers you already keep, and the reward that pads a metric or pays for nothing shows itself early, while it's still cheap to kill.
What's left is the spend you can point to: the weekend shift that got covered, the upsell that survived to the next statement, the overtime you didn't pay, and the rework that never came back.
Frequently asked questions
Do employee incentives actually backfire?
Yes, when nobody measures what the reward buys. They pay for behavior you'd get anyway, or people chase the number without hitting the result, or the target was wrong to begin with. The ones that work aim at a specific behavior you can watch, so you catch the ones going wrong while it's still cheap.
What is the overjustification effect?
It's the finding that rewarding people for something they already enjoy can dull their own motivation to do it. The effect is strongest on work people find interesting, and for the shifts and tasks nobody enjoys, a reward tends to raise effort instead. On the frontline it's rarely the real risk.
How do you stop employees from gaming an incentive?
Keep the target narrow and watch the profit behind the metric, not just the metric. Gaming shows up as a number that climbs while the money it was supposed to stand for sits still. When you measure both against your real operational data, a padded result shows itself early enough to cut.
Why did Wells Fargo's incentive program fail?
Employees faced sales targets tied to their pay with no check on the profit behind the accounts they opened, so about two million were opened without customers asking. The metric looked strong while the value under it was fake. A measured target would have exposed the gap early.
How do you know if an incentive is working?
Measure what the reward returns. Weigh what it cost against the outcome it was meant to buy, shift by shift or role by role. If it returns more than it cost, widen it. If it doesn't, cut it. That's how you tell a reward that pays from one that just spends.
Sources
- Wikipedia, Overjustification effect: rewarding an activity someone already enjoys can reduce their intrinsic motivation, with the effect strongest on high-interest tasks and reversing to raise motivation on low-interest ones.
- Consumer Financial Protection Bureau, Wells Fargo fined $100 million for secretly opening unauthorized accounts: employees opened more than 2 million unauthorized deposit and credit card accounts (roughly 1.5 million deposit and 565,000 credit card) to hit sales targets tied to incentive compensation.
