I spent a quarter building a bonus plan for two departments, and the hardest part had nothing to do with the money.

The mechanics were straightforward: quarterly, points-based, distributed against a small set of departmental metrics. Setting the payout pool took an afternoon. Deciding the metrics took three weeks, because every metric I wrote down, I could immediately imagine someone hitting it in a way that would make the business worse.

Reward speed and quality drops. Reward volume, and people stop turning away the customer who is going to be a nightmare. Reward a satisfaction score, and you get staff coaching guests on how to fill out a survey, which is a real thing that happens in real businesses run by good people every day.

The metric is not a measurement instrument. It is an instruction, and it will be followed literally by people who are tired and who need the money.

A bonus plan is a sentence about what the company wants

Whatever you write in the plan document, the version that matters is the one that gets said out loud in a break room. That version is short, it is concrete, and it is often not what you meant. Ocean Stride Advisory | Framer production copy 9: You wrote drive incremental attachment revenue across the guest experience. What gets said at four o'clock on a Saturday is: don't let anyone leave without asking twice.

Both of those are the same plan. Only one of them is the one your guests will actually experience, and you did not write it. Your staff did, because they had to translate your sentence into something a human being can do while standing up.

The Behavior Sentence

Before a metric goes in a plan, write the sentence it causes. Not the outcome you want. The instruction a tired person on a busy shift derives from it, in their own words, and hands to a new coworker.

Three questions get you there.

1. What is the cheapest way to hit this number? Not the best way. The cheapest. Assume good people under real pressure who have decided the number matters, because that is what you have engineered. If the cheapest path is something you would be unhappy to watch, you have not written a metric; you have written a trap.

2. Who can hit this by doing nothing different? If a department can clear the bar on a normal quarter with no change in behavior, you are not running an incentive. You are running a raise with extra paperwork, and everyone will figure that out by quarter two.

3. Who cannot hit this no matter what they do? Any metric heavily driven by weather, seasonality, or another department's performance will be read as unfair, and it will be right. People disengage from a game they can see is rigged much faster than they disengage from a game they are losing fairly.

Points beat percentages

Percentages invite negotiation and create cliffs. Points create partial credit, and partial credit is what keeps people in the game in a bad quarter.

Under a percentage plan, a team that knows by week six it will miss the threshold has nothing left to play for and will spend the remaining seven weeks accordingly. Under a points plan, they can still bank something on the metrics they can still move, which means the last third of a bad quarter is not simply forfeited.

Points also let you weight without arguing about weighting, which is the conversation that eats the most calendar time and produces the least value. Nobody fights about whether safety is worth thirty points the way they fight about whether it is worth thirty percent of their pay. Same math, less blood.

And keep a small discretionary block, published as discretionary. Five to ten points, awarded by a manager, with the reason written down. It is your only mechanism for rewarding the thing you did not think to measure, and the thing you did not think to measure is usually the behavior you most want more of.

The failure mode

The plan pays out in full, and nothing changed.

That is not a success. That is the plan telling you the metrics tracked something already happening, and you converted a variable cost into an expectation. You will never get it back. Bonus money, once received twice, is salary, and removing it reads as a pay cut regardless of what the document says.

The warning sign arrives about a week after the first payout, and it is quiet: nobody asks how the number is calculated. Curiosity means people are trying to move it. Silence means they have concluded it moves on its own.

One thing to try this week: Take one metric you currently pay against or report on. Write the sentence it causes, in the words a tired person would use on a Saturday. Then decide whether you want more of that. — Michael J Washington, Ocean Stride Advisory