Your KPIs are creating the behaviour you're trying to measure.
The Founder Test: If this KPI hasn't forced a decision in the last 30 days that wouldn't have happened otherwise, it's burning attention to manufacture false comfort.
The utilisation trap
A professional services firm (£10m revenue, 45 consultants) set a utilisation target of 75% billable hours. Within six months, consultants were logging internal meetings as "client development", stretching 4-hour engagements across 2-day site visits to inflate hours, and declining complex, lower-margin projects that required more internal coordination.
Revenue grew 11%, but client satisfaction scores dropped 19 points and two major accounts didn't renew. Gross margin fell an estimated £450k from project overruns and rework the utilisation metric never captured. The CEO only saw the pattern when a departing consultant mentioned the team had a weekly Slack thread for working out how to hit the numbers.
The metric was technically accurate. The behavioural response improved the number whilst destroying the strategic intent behind it. The question isn't whether your team will game a KPI. It's how long until they find the best gaming strategy, and whether you'll notice before it becomes systemic.
The same mechanism, sharper teeth
A manufacturing business tied production-line bonuses to a <2% defect rate.
Within a quarter, supervisors were reclassifying functional defects as "cosmetic" and holding problematic batches back until after the inspection window. The CEO only saw the pattern when a £1.2m customer contract was put on review over quality issues that did not appear in any internal report. The metric was green.
Gaming detection signals
If you see these signals, your metrics are being gamed:
- Threshold hugging: performance clusters tightly around the minimum target (e.g., 80% of team hits exactly 100-102% of quota). Natural performance distributes normally; gaming concentrates at thresholds.
- Early greens: metrics report success in months 1-3, then plateau or decline. Gaming is easiest early when scrutiny is low and baseline expectations unclear.
- Language tells: the team starts using phrases like "technically we hit it" or "the number looks good but..." Linguistic hedging precedes metric manipulation.
- Counter-metric collapse: when one metric improves, logically connected metrics deteriorate (sales up, NPS down). Gaming inflates the measured whilst eroding the unmeasured.
- Informal shadow metrics: the team creates unofficial tracking that contradicts official KPIs. When your sales team tracks "real pipeline" separate from CRM pipeline, they're signalling the official metric is compromised.
Where the gaming concentrates: thresholds
The more precisely you measure something, the more fragile the behaviour around it becomes. Metrics concentrate risk at thresholds.
Consider the standard commission structure: no bonus below 80% of target, accelerating above 100%. A rep at 78% in December has two options: abandon strategic deals that won't close before year-end and chase transactional wins to cross the threshold, or give up entirely because the gap is unbridgeable. The metric doesn't measure sales performance. It manufactures binary decisions.
The same fragility appears in operational KPIs. A logistics business sets a 98% on-time delivery target. Drivers refuse loads that might jeopardise the metric, even when the customer would accept a slight delay for a consolidated shipment. The metric chases the number, not the customer. In both cases, the threshold creates the binary; the binary destroys the strategy.
Four diagnostic questions
You can't eliminate gaming. You can make it visible rather than hidden.
1. Outcome or activity? Are you measuring the strategic outcome you want, or the activity you assume produces it? "Revenue growth" is an outcome. "Pipeline coverage ratio" is an activity. Activities can be gamed without moving the outcome. If your KPI is an activity metric, what's the falsifiable link to the outcome?
2. What's the gaming handbook? If you publicly announced the exact formula for this KPI and the incentive structure six months in advance, what would your team do by month three? If you can't articulate the gaming strategy, you haven't stress-tested the metric.
3. Where's the threshold concentration? What percentage of team compensation or status depends on this single number? In the £5-15m owner-managed businesses I work with, once one number carries roughly a third or more of total pay, the binary-decision pattern shows up reliably: when the threshold is hit, effort to improve stops; when it's missed, effort collapses. Spread the weight across at least three metrics so no single number carries the incentive on its own.
4. What does this metric cost you personally? Apply the Founder Test from the top of this piece. Then apply its mirror: has this KPI forced you into a recurring approval loop in the last 30 days? The KPI that requires your review every week is the one that keeps you in the building. If retiring or redesigning a metric would remove you from a recurring approval loop, that's a founder freedom signal, and probably the strongest argument for changing it.
Chaperone pairs
Every primary metric needs a chaperone metric that exposes gaming. You're watching for when the primary improves and the chaperone deteriorates.
Sales revenue / Net revenue retention: revenue can be gamed through discounting or contract stuffing; NRR reveals whether revenue was real or borrowed from future quarters.
Operations efficiency / Utilisation percentage: efficiency can be gamed by declining difficult work; utilisation reveals whether efficiency came from selectivity rather than capability.
Customer acquisition / Customer acquisition cost: acquisition can be gamed through aggressive spend; CAC reveals whether growth was bought or earned.
Product velocity / Technical debt ratio: velocity can be gamed by shipping fast without testing; debt ratio reveals whether speed came from discipline or shortcuts.
Employee headcount / Revenue per employee: headcount can be gamed through hiring sprees; RPE reveals whether scale was productive or dilutive.
Any KPI that doesn't make you stand up and act, is just taking up dashboard space.
One-screen founder action
Open a blank spreadsheet. Five columns:
- Your top 5 KPIs: the metrics that drive board discussion and pay decisions.
- Pay/status %: what percentage of team compensation or status rests on each metric? Flag any >30%.
- Outcome vs activity: label each as "outcome" (measures strategic result) or "activity" (measures proxy behaviour). Flag all activities.
- Public-formula test: for each KPI, write what your team would do in 90 days if you announced the exact formula and incentive structure today. If you can articulate a gaming strategy, the metric needs redesign.
- Chaperone assignment: for each primary KPI, identify the chaperone metric that would reveal gaming (revenue / NRR, efficiency / utilisation, acquisition / CAC). If you don't have the chaperone, you're flying blind.
Pick one metric to retire this quarter. Not revise: retire. Choose the one that's been unchanged longest, or where the pay concentration is highest. Replace it with an outcome-based metric that has an explicit 18-month review date.
The moment you publish a KPI, you stop measuring and start manufacturing.
If this audit surfaces metrics that need retiring rather than redesigning, the KPI Audit Framework gives you the quarterly discipline to manage the transition without losing trend comparability.
This article is a strategic compass, not financial, legal, or tax advice. It highlights patterns common in UK owner-led businesses. Consult qualified advisers for decisions specific to your situation.