Evidence guide: KPIs, proxy metrics and benchmarks

The decision question

How can I tell whether a KPI or benchmark is still measuring the goal instead of becoming the goal?

Reviewed lenses

Surrogation – when a performance measure starts replacing the goal — Surrogation is a domain-specific management-accounting construct: a measure that was designed to represent a strategic objective can start to be treated as though it were the objective itself. It is closely related to metric fixation and proxy problems, but those broader labels should not be treated as exact synonyms without checking the setting and mechanism.
Anchoring Effect – when a starting number pulls later estimates toward it — Anchoring is a well-supported effect in which an initial numerical value can pull a later estimate toward it. A 2026 meta-analysis covering 2,601 effect sizes found a large overall effect, but also substantial variation across studies. The effect should not be treated as a rule that every number changes every judgment: incidental anchors, anchors from a different dimension, clearly random values, incentives, and some debiasing conditions were associated with smaller or null effects.
Decoy Effect – when an inferior option changes the comparison between better options — The decoy effect is a well-documented family of context effects in which adding an inferior or otherwise strategically positioned alternative can change choice shares among the original options. The classic attraction effect uses an asymmetrically dominated decoy that is worse than the target but not necessarily worse than the competitor. The effect is not guaranteed: its size depends on the geometry of the options, prior preferences, task design and other moderators, and preregistered replications have found both successful and weak or null results.
Dunning–Kruger Effect – when performance and self-assessment do not line up — The original studies found substantial overestimation among low performers in specific tasks, but the popular story that incompetent people are simply very confident while experts systematically underestimate themselves is too broad. Later work showed that common quartile plots can exaggerate the pattern through regression and better-than-average effects, while a 2023 reanalysis using more appropriate methods still found a small significant effect in intelligence data.
Systematic Bias – a measurement and statistical concept, not one cognitive bias — Systematic bias is a general statistical and measurement concept: a non-random tendency for measurements, estimates, or study results to deviate from a reference or target value in a consistent or predictable way. Standards in metrology define systematic measurement error separately from random error and define measurement bias as an estimate of that systematic error. This is not one specific psychological bias with one mechanism, so the page should not present it as a personal cognitive tendency alongside constructs such as confirmation bias or hindsight bias.
Backfire Effect – when a correction can strengthen a false belief — Corrections usually improve factual accuracy. A true backfire effect, where a correction makes the targeted false belief stronger, appears to be uncommon and sensitive to context and measurement. It should not be treated as the default response to being corrected.
Declinism – when the present is judged against an idealized past — Declinism is best treated here as an umbrella label for judging the present as worse than an idealized past, not as one standardized cognitive-bias construct with a single mechanism. Research directly supports several narrower ingredients: people can remember past experiences more positively than they experienced them, negative affect associated with autobiographical memories often fades faster than positive affect, and large multi-study work finds a pervasive illusion of moral decline. None of this means that every claim of social, technological, institutional, or personal decline is false; real decline must be tested against domain-specific evidence.
Value Selection Bias – choosing supplied numbers instead of constructing the Bayesian answer — Value Selection Bias is a narrow reasoning pattern reported in Bayesian diagnostic problems: when participants did not derive the correct solution, incorrect answers were strongly drawn toward numerical values already present in the problem. The evidence supports this behavior across multiple problem variations in the authors' research program, but it should not be generalized to any situation where a person uses an old price, stale KPI, or convenient number.
Availability Heuristic – when easy-to-recall examples shape frequency judgments — Availability is a judgment heuristic: people can use how easily examples or scenarios come to mind when estimating frequency or probability. That shortcut is not automatically an error because memorable or accessible examples can correlate with real frequency. Bias appears when accessibility is driven by factors that are not diagnostic of the quantity being judged.
Conservatism Bias – when new evidence doesn't change old beliefs — Here, conservatism means under-updating: new evidence changes a probability judgment less than a Bayesian benchmark would predict. It should not be confused with the separate memory phenomenon in this corpus that pulls remembered extreme values toward the middle.

Decision guides

How to use this page

Start with the decision, not with a label for a person. Open the evidence behind the lenses that fit, compare nearby concepts when needed, and keep uncertainty visible.