I’ve spent much of my career designing systems that make people want to do the right thing. Incentive programs, to me, are the invisible scaffolding that make parts of my business run more effectively than traditional 9-5, salary-based work.

But they can also be utterly disastrous when designed poorly. History is full of well-intentioned programs that produced exactly the wrong behavior simply because the side effects weren’t thought through.

If you’re not careful, you’ll find yourself swimming in game theory with incentive programs, and before you know it you’ll be rattling off things like “Nash equilibrium” and “rational actor models” like you’re designing a Cold War simulation instead of a bonus plan.

It can get pretty abstract, and while it certainly can help, I’ve found there’s some more down-to-earth ways to talk about the theories in a way that apply to incentive programs.

Over time, I’ve distilled it into three rules that keep me honest when I’m designing any incentive structure. Let’s get into it.

Reward the Real Outcome

The most-successful, hardest to game incentive programs reward the real outcome. They do not reward anecdotal evidence, or proxies, but instead directly reward when the essential, measurable result has been accomplished.

This can be trickier than it sounds, and if you’re not careful, you can pigeon-hole yourself into thinking you’re rewarding for the actual result, and yet you’re actually rewarding for a proxy. Doing so inevitably leads to people gaming the system by improving the proxy number, which doesn’t always lead to the actual result you want, or worse it actually contradicts it.

In the early 1900s, the French colonial government had a rat problem. Streets and sewers were overrun, so they created an incentive program to pay people a bounty for every rat tail they turned in.

It worked at first. Tails poured in, but officials eventually noticed something strange. The rat population wasn’t shrinking. In fact, it was getting worse.

Instead of killing rats, many just chopped off the tails and set the rats free to breed more. Even better, some started farming rats outright. The incentive, meant to solve the infestation, ended up multiplying it.

This story serves as a cautionary tale when designing incentive programs. Elegant, well-intended programs can have unintended consequences that create the opposite of the intended goal.

This went wrong because the incentive program did not reward the actual outcome. What they actually wanted out of this had little to do with rats at all. They actually wanted:

• Less damage to stored grain and crops
• Less damage to infrastructure/property
• Less visible filth in public spaces
• Less spread of plague and disease

The Hanoi story is extreme, but the principle is universal. If you reward the wrong thing, people will optimize for the wrong thing. Every incentive program lives or dies by whether it’s truly bound to the outcome that matters.

Make the Winning Strategy Sustainable

A healthy incentive program is one people can stay in for years without burning out. The program rewards performance, and pays well enough that they don’t have to overextend to satisfy their needs, but also provides opportunities for them to stretch when they want a little more.

Most programs do the opposite. Instead, their system inadvertently punishes their best people for taking a break. That drives burnout, and ends up making the incentive program less-effective in the long run. An incentive program is only as good as the people leveraging it, and if they aren’t able to stay healthy, and productive, the system will falter.

DoorDash gives their most-committed dashers “platinum status”. This status promises higher paying perks. Platinum status requires consistent activity or you lose it. It’s designed to incentivize people to dash consistently.

This ensures that DoorDash gets a much-needed consistent commitment from a subset of their dashers, but it does run the risk of burning out their VIPs, who fear losing status if they pause.

DoorDash solved this problem by adding something that makes keeping platinum status sustainable in the long-run. They periodically give their top dashers a “Platinum Pass”. This pass allows them to keep their status even if they don’t dash at all for a brief period of time. This gives dashers space to rest without losing what they’ve earned.

A similar problem manifests with salespeople. Traditional work environments have paid vacations, but for a salesperson that paid vacation only applies to their base salary which is a fraction of their real income. So a week off feels like a financial loss, and many simply don’t take it.

In that setup, the rational winning strategy is to never take a vacation.

The result is that the best salespeople burn themselves out, and leave. Worse yet, the ones that stick around are the ones that aren’t performing, because they don’t lose as much when they take a break.

The long-term consequence is that your system has cultivated a community of poor performers who are inadvertently doing the opposite of what you want the incentive program to do.

Imagine if incentive programs covered a salesperson’s typical commission during vacation. Rest would finally feel affordable, and they’d stick around. It would cost more money on the surface, but the long-term gains of keeping high-performing salespeople happy, and productive would outweigh the expense of incentivizing their vacation.

Ultimately, If rest is irrational, burnout is inevitable. Take care of your people, and ensure that the winning strategy empowers a healthy work/life harmony.

Only Reward What People Can Control

Profit sharing at very large companies is a great example of misaligned control. It works when ten owners all move the business together. It breaks when thousands of people are “measured” on a number they barely influence.

A team could be killing it on something and see a bad profit share payout because the rest of the company is struggling. Conversely, dead weight in the company can see a big profit share payout, all for things they had nothing to do with.

Best case, people ignore it because it’s not actionable. Worst case, they depend on it and carry stress for something they can’t move. If you want it to motivate behavior, bring the metric closer to the levers people actually pull, and reserve company-wide profit share as a true, disclosed bonus instead of a performance signal.

Ultimately, If a reasonable person can’t see how their daily work moves the metric, don’t pay them on it.

Conclusion

Incentive programs are powerful tools, but they’re also fragile systems. You can’t set them and forget them. Instead, you have to design, observe, and adjust as people inevitably find new ways to play the game, but also because the rules will subtly change as your business changes. It’s crucial to keep an eye out for these shifts and adjust your programs as those changes come.

If you continue to reward the real outcome, make the winning strategy sustainable, and only pay for what people can control, you’ll build something that will motivate people in the right ways, keep them happy, and empower them to make a difference in their corner of your business.

At their best, give people a reason to win in ways that make the whole system stronger.