METHODS / GOALS, CADENCE, AND OPERATING SYSTEMS / KPIS AND SCORECARDS

PRACTICE

KPIs and Scorecards

A key performance indicator is a defined signal used to understand progress, operating health, or risk in relation to an important outcome. A scorecard brings selected indicators into one recurring review so leaders can identify a change, investigate it, and decide what to do.

REFERENCE NIST Baldrige, Thinking about Performance Measurement

WHAT IT HELPS WITH

A leadership team that has reports or dashboards but lacks consistent definitions, ownership, review logic, or a clear response when a number changes.

TYPICAL SCOPE Team
SETUP EFFORT Moderate
ONE MATERIAL LIMITATION

A scorecard simplifies reality. Measures can be delayed, incomplete, gamed, or improved at the expense of the outcome they are meant to represent.

WHAT IT IS

How a KPI becomes part of a working scorecard.

A KPI needs an important outcome, a precise definition, a reliable source, an owner, a review cadence, and a possible response. The number alone is a signal. The definition and review practice make it usable.

A scorecard brings a selected set of indicators into one review. It helps leaders see movement, ask what changed, investigate the cause, record the response, and return at the next review point.

A KPI is one number, fully defined. A scorecard is a few of them, read together on a fixed day.

FIT

When KPIs and scorecards are useful.

Reach for them when

  • Leadership needs a recurring view of an important outcome or risk.
  • Teams use different definitions for the same number.
  • Reviews spend time rebuilding data instead of deciding what to do.
  • A change in the signal can lead to a real response.

Use another approach when

  • The question requires a one-time analysis.
  • The outcome cannot yet be defined or measured responsibly.
  • The data is too unreliable for the consequence attached to it.
  • Leadership wants a dashboard but has no review or response process.

THE MEASURE RECORD

What each KPI should explain.

HOW TO SET IT UP

Start with the decision the measure should support.

01
Name the decision.
State what leadership may need to understand or change.
02
Define the outcome.
Describe the result, condition, or risk the measure is meant to represent.
03
Write the calculation.
Specify the source, exclusions, timing, and unit so two people can reproduce it.
04
Test the measure.
Check a small sample for reliability, delay, and ways the number could mislead.
05
Assign ownership.
Separate the outcome owner from the person maintaining the definition when appropriate.
06
Run the review.
Examine movement, context, possible causes, decisions, and the next review point.

EXAMPLE

A client-onboarding scorecard.

A leadership team wants clients to reach their first usable outcome sooner. It reviews onboarding cycle time, required inputs received on time, and handoffs blocked beyond the agreed threshold. Formulas calculate the official values.

Executive operations prepares the review and links exceptions to the underlying accounts. Leaders investigate the pattern before assigning cause and record the response they want to test.

KEEP IT USEFUL

Common measurement and behavior considerations.

Common measurement issues

  • Tracking so many indicators that leadership cannot review them meaningfully.
  • Using one label for measures with different calculations.
  • Interpreting movement before checking missing, delayed, or changed data.
  • Reviewing the number without the source, definition, or data date.
  • Keeping a measure after it no longer supports a decision.

Behavior and governance considerations

  • Pair speed or volume measures with a relevant quality guardrail.
  • Check how targets and incentives may change behavior.
  • Avoid using a team-level signal as an individual performance verdict.
  • Investigate causes before assigning blame or prescribing a fix.
  • Keep sensitive financial, workforce, client, and security data appropriately restricted.

AI AND JUDGMENT

Use AI to prepare the review. Keep the official number and the decision under human control.

AI can help

  • Check draft measure records for missing definitions, sources, owners, or dates.
  • Flag stale, missing, or unusual values for review.
  • Draft a source-linked narrative explaining what changed in the data.
  • Retrieve the accounts, projects, or events behind an exception when access is approved.

People remain responsible for

  • Choosing the outcome and deciding whether the indicator represents it responsibly.
  • Approving the formula, source, exclusions, target, and access level.
  • Investigating causes and validating material exceptions.
  • Making personnel, budget, policy, customer, and resource decisions.

Use deterministic formulas or queries for the official value whenever possible. AI may summarize or investigate the data, but it should not invent or replace the number.

SOURCES

What the guidance supports.

Independent evidence

A NIST Baldrige article and GAO criteria emphasize measures that are linked to important goals, clearly defined, reliable, balanced, timely, and useful for decisions. Independent research shows that measurement systems can shape behavior and performance in helpful or harmful ways depending on their design and use.

View sources (3)OpenClose
NIST Baldrige, Thinking about Performance Measurement. Source, accessed August 31, 2026.
U.S. Government Accountability Office, IRS tax filing season performance measures, Appendix I. Source, accessed August 31, 2026. The GAO criteria for a useful performance measure.
Franco-Santos, Lucianetti, and Bourne, review of performance-measurement system consequences, 2012. Source, checked August 31, 2026.

WHERE IT FITS

What this pairs with.

LAST REVIEWED: AUGUST 31, 2026 Every claim is checked against the source it names, and where the evidence does not support a claim we say so. Spotted a problem? Tell us and we correct it in the open.