Setting a Target for a Metric Without Inviting the Wrong Behavior

Why this matters

The moment you attach a target to a number, people start optimizing for the number instead of the thing the number was supposed to represent. A target on "jobs per day" gets you more jobs and shorter ones. A target on "average call time" gets you calls cut off before the customer's actual problem is solved. This is not a character flaw in your team, it is a predictable response to how targets work: whatever you measure and reward becomes the thing people protect, sometimes at the expense of the outcome you actually cared about. Setting targets well is a skill, not a formality, and a badly set one can make performance look better while quality gets worse underneath it.

The core trap: the metric is a proxy, not the goal

Every metric you track is a stand-in for something you cannot directly measure in real time. "On-time arrival rate" is a proxy for "customers feel respected and can plan their day." "Jobs completed per tech" is a proxy for "the crew is productive." The proxy is useful precisely because it is easy to measure, but it is never a perfect substitute for the real thing, and the gap between the two is exactly where gaming lives. The tighter you squeeze the proxy, the more that gap gets exploited, often without anyone intending to cheat. A tech chasing a jobs-per-day target is not being dishonest; they are responding rationally to what gets rewarded.

Set the target with the failure mode already in mind

Before you announce a target, spend two minutes asking: what is the laziest way someone could hit this number without actually improving the thing I care about? If you can answer that question easily, so can your team, and they will find it faster than you expect once the pressure is on.

• Target: "increase jobs completed per day." Laziest cheat: rush jobs, skip steps, undercount time. Fix: pair it with a quality or callback metric so speed without quality does not look like a win.
• Target: "reduce average time-to-invoice." Laziest cheat: send incomplete or estimated invoices just to hit the clock. Fix: pair it with an invoice-accuracy or dispute-rate metric.
• Target: "increase close rate on estimates." Laziest cheat: only send estimates for jobs you already know will close, and quietly avoid harder or more competitive quotes. Fix: track close rate alongside total estimates sent, so a rising rate on a shrinking pool is visible.
• Target: "reduce overtime hours." Laziest cheat: stop logging hours worked rather than actually reducing them. Fix: cross-check against completed job volume; a drop in logged overtime with flat output is a red flag, not a win.

The pattern in every example: a single-number target invites a shortcut, and a paired metric that would expose the shortcut closes most of the gap. This is the single most reliable technique in this whole discipline.

Prefer a range or a direction over a hard number when the metric is noisy

Some metrics naturally swing week to week for reasons that have nothing to do with effort, seasonality being the most common. Setting a rigid target on a noisy metric forces people to manufacture consistency that is not real, whether that means shifting which week a job gets logged in or padding a slow period with busywork. Where the underlying number is genuinely variable, set a target as a range or a trend direction ("improving over the quarter") rather than a single hard threshold hit every single week. This removes the incentive to fudge timing just to land inside an artificial line.

Involve the people who will be measured in setting the number

A target imposed from above, with no input from the people whose work it measures, gets treated as an obstacle to route around. A target the crew helped set, informed by what they know about real constraints, gets treated as a shared goal. This is not about lowering the bar to make everyone comfortable, it is about anchoring the number to reality: the people doing the work usually know exactly where a target is unrealistic or where it would push a shortcut, and they will tell you if you ask before you announce.

Revisit the target on a schedule, and watch for a number that stopped moving

A target that never gets revisited becomes stale: either it was hit long ago and now does nothing, or it was too aggressive from the start and everyone quietly stopped trying to hit it, which teaches the whole team that targets are not really enforced. Two signals worth checking each review cycle:

• A metric that improved fast and then plateaued exactly at the target line, never above it, is a strong tell that people are optimizing to the threshold rather than the underlying goal. Genuine improvement usually overshoots a target at least sometimes; a number that lands suspiciously close to the line every period deserves a closer look at the paired metric.
• A metric that has not moved in months despite attention either means the target was set wrong, the process cannot actually improve further without a different intervention, or the target quietly stopped being tracked. Any of those three is worth a conversation, not silent tolerance.

The mental model to keep

A target is a lever, and every lever moves more than the one thing you meant to move. Before you set a number, name its cheapest workaround out loud, and pair it with a second metric that would expose that workaround. Prefer a direction over a rigid line where the underlying number is naturally noisy, and build the target with the people it measures, not just for them.

References

• U.S. Small Business Administration (SBA), small business performance management practices
• Trade-standard practice for KPI design in field-service operations
• See related: A Metric Is Easy to Game and Someone Is Gaming It; A New Metric Is Worth Adding to the Dashboard or Not