How should KPIs be designed to avoid encouraging the wrong behavior?
Goals can improve performance while narrowing attention and distorting behavior. Use goal-setting research and the balanced scorecard to pair outcome and process indicators and design rewards.
What principles should guide KPI design?
Performance indicators change where people actually direct their effort. Concentrating evaluation and rewards on a single outcome indicator can therefore neglect work that the indicator does not capture. Alongside financial outcomes, include measures of customers, work processes and learning, and pair each outcome measure with an indicator that reveals side effects. Review quarterly whether the measures are serving their original purpose. Starting with a few indicators and improving them is more practical than trying to find perfect measures immediately.
Why can goals and indicators encourage the wrong behavior?
Goal setting is one of the most repeatedly tested theories in management. Hundreds of studies have found that specific, difficult goals guide behavior and improve performance. However, a paper reviewing the side effects of goal setting (Ordóñez and colleagues, 2009) argues that the benefits have been overstated while systematic harm has largely been ignored. This helps explain why raising a sales target can increase sales together with returns and unpaid receivables.
The side effects identified include narrow attention that neglects areas outside the goal, distorted risk preferences, increased unethical behavior, inhibited learning, erosion of organizational culture and reduced intrinsic motivation. The researchers recommend treating goal setting more like a prescription medicine requiring careful dosage, consideration of side effects and close monitoring than a general remedy anyone can use.
Side effects of goal setting and safeguards
| Side effect | How it appears at work | Safeguard |
|---|---|---|
| Narrow attention | Sales targets take priority over collections and quality | Pair each outcome indicator with a side-effect indicator |
| Distorted risk preferences | Excessively risky transactions in months below target | Establish a review process when targets are missed |
| Unethical behavior | Pulling results forward or inflating figures | Check supporting evidence alongside the numbers |
| Inhibited learning | Avoiding attempts to use new methods | Evaluate experimental tasks separately |
| Cultural erosion | Withholding information from colleagues | Include team-level indicators |
| Reduced motivation | Ignoring work not covered by a measure | Combine indicators with qualitative evaluation and discussions |
There is no need to introduce every safeguard at once. Start with signs already visible in your organization and add safeguards addressing them.
What is missed when only financial measures are used?
The paper that first proposed the balanced scorecard (Kaplan and Norton, 1992) was based on a year of research with 12 companies at the forefront of performance measurement. The researchers argued that financial indicators show the consequences of actions already taken. They therefore added operational measures of customer satisfaction, internal business processes, and organizational learning and improvement to reveal activities driving future financial performance. Financial results are finalized only after a quarter or year; operational measures enable course corrections in the meantime.
The paper recommends translating the company's strategy and mission into concrete objectives and measures. A customer-oriented strategy to bring products to market faster, for example, can become a measure of time to launch. Indicators should start from strategy and change with it.
Which indicators can be used for the four perspectives?
The four balanced-scorecard perspectives can be translated into SME indicators as follows.
Examples of indicators and paired side-effect measures by perspective
| Perspective | Question | Example indicators | Paired indicator |
|---|---|---|---|
| Financial | How do shareholders and management see us? | Operating margin and cash collection period | Share of discounted sales |
| Customer | How do customers see us? | Repeat-purchase rate and on-time delivery | Number of complaints |
| Internal processes | What must we do well? | Defect rate and processing time | Rework rate |
| Learning and improvement | Can we keep improving? | Share of new products and post-training application assignments | Turnover among key employees |
Use paired indicators as warning lights rather than additional targets. If they cross a certain level, examine the causes together.
How many indicators should there be, and who should decide?
Turning every activity into an indicator can obscure what matters. Reduce the number to a set whose trends managers can view on one screen, and assign an owner, formula and data source to each measure. Involving operational staff in their selection can reveal where people might find shortcuts to meet the numbers. The effects of generative AI adoption provides examples of effectiveness measures and accompanying risks. Testing new business ideas explains how to establish continuation and stopping criteria in advance. If an indicator is measured by a survey, such as customer satisfaction, first check the margin of error implied by the respondent count. New indicators can be piloted for one or two quarters before they are included in evaluation. Keep a record of changes to calculation formulas, since results based on a revised formula cannot be compared directly with earlier periods.
How should indicators be linked to rewards?
The more strongly rewards depend on one measure, the greater the risk of side effects. Consider outcome indicators together with their paired measures, and use discussions and qualitative assessment to recognize collaboration and improvement efforts that numbers miss. Set targets by balancing feasibility and challenge. Giving people an opportunity to explain shortfalls can reduce excessive risk taking and inflated figures. Explaining which indicators affected reward decisions and how also helps employees see them as judgment criteria rather than figures to game. The Korea Research Institute of Business & Branding treats sustainability as a core value: considering short-term performance alongside long-term competitiveness and trust.
How can we check whether indicators are doing their job?
Each quarter, check whether improvements in the numbers have been accompanied by improvements in the original objective. If on-time delivery rises while complaints increase, that may signal a problem with how the measure is being achieved. In review meetings, ask what behavior produced the number rather than focusing on the number alone. The Korea Research Institute of Business & Branding analyzes organizational structures, decision-making systems, productivity, quality and operational efficiency through organizational diagnosis and operational improvement. For the wider range of related areas, see business research and advisory services.
Frequently asked questions about KPIs
How do KPIs and OKRs differ?
Both connect objectives and measures. Before focusing on the name, check whether the indicators start from strategy and are paired with measures that reveal side effects.
Should team or individual indicators take priority?
Where collaboration matters, giving greater weight to team measures supports information sharing. Individual measures alone can encourage competition among colleagues.
Is it wrong to change indicators frequently?
Changing too often prevents meaningful trend analysis. Change a measure when strategy changes or there is evidence of side effects, and record why it was changed.
How should work that is hard to quantify be evaluated?
Agree on output standards in advance and combine them with qualitative assessment through regular discussions. Forcing the work into a number can encourage people to focus only on meeting that number.
References
- Academy of Management Perspectives, “Goals Gone Wild: The Systematic Side Effects of Overprescribing Goal Setting” (Ordóñez, Schweitzer, Galinsky and Bazerman), 2009. https://doi.org/10.5465/amp.2009.37007999
- Harvard Business Review, “The Balanced Scorecard: Measures That Drive Performance” (Kaplan and Norton), 1992. https://pubmed.ncbi.nlm.nih.gov/10119714/