EPI: Performance Assessment and Measurement

Performance Assessment and Measurement tests how to turn strategy and process objectives into a balanced, trustworthy measurement system. Questions target KPI design, operational definitions, baselines and targets, leading versus lagging indicators, data quality, review cadence, and benchmarking.

Core rules and concepts

• Build a KPI hierarchy: enterprise outcomes, value-stream or business-unit outcomes, process measures, and activity measures. Each level should explain the level above it.
• Write an operational definition for every KPI: numerator, denominator, inclusions, exclusions, data source, owner, frequency, and calculation rule. Ambiguity here produces disputes later.
• Pair lagging outcome measures with leading drivers. A lagging measure such as customer retention is useful for accountability; a leading measure such as first-contact resolution is useful for action.
• Establish a baseline before setting a target. A target without a baseline is an aspiration, not a measurement plan.
• Check data quality dimensions: completeness, accuracy, timeliness, consistency, validity, and uniqueness. A precise chart from bad data is false confidence.
• Use a review cadence matched to the measure: daily for operational flow, monthly for service performance, quarterly for strategic outcomes.

Worked example

A service unit proposes a complaint-rate KPI. The operational definition is complaints received in the month divided by eligible orders shipped in the month, excluding complaints later confirmed as unrelated to the unit. In January there are 48 complaints and 2,000 eligible orders, so the rate is 48 / 2,000 = 2.4%. The target is 1.5%, which means 30 complaints at the same volume, a reduction of 18 complaints or 37.5%. Because the target depends on volume, the team also tracks complaint count and order mix. The leading indicator chosen is first-contact resolution, since faster resolution should reduce escalations. The reasoning is that the rate gives comparability, while the count and mix prevent the team from missing a rising absolute problem.

Traps that cost marks

• Choosing vanity metrics that always rise and never drive a decision.
• Leaving the denominator undefined, so trends reflect mix changes rather than performance.
• Setting targets without baselines, or using arbitrary stretch numbers with no causal logic.
• Gaming a measure by suppressing complaints or reclassifying cases.
• Treating correlation as proof that one KPI caused another.