Names and aliases
quality of fit, match fit
Formula
no single formula exists; the published output is a vector of measurement-model indicators — completion rate, refund rate, dispute rate, repeat engagement with the same counterparty, and deviation from the original request
Metric card
| Unit | vector of indicators; no composite score |
|---|---|
| Data source | The original demand request, attributes of the selected supply unit, completion events, disputes, refunds, repeat engagements, and explicit feedback where it exists. |
| Frequency | By realised-match cohort after every outcome window has closed. |
| Segments | micro-market, category, request type, matching channel, seller type, cohort |
Interpretation errors
- Publishing a composite match-quality score: aggregating incomparable indicators creates the appearance of measurement where weights were chosen rather than a quantity measured.
- Treating an explicit user rating as a measure of fit: it blends match quality, fulfilment quality, and the rater's mood.
- Reading the absence of a complaint as confirmed quality: silence more often reflects a low cost of walking away than satisfaction.
Example
Illustrative vector for one cohort: 90% completion, 3% refunds, 1.2% disputes, 11% repeat engagement with the same counterparty. The numbers describe no real platform and do not sum to a score.
Use
Declare the measurement model before analysis: list the indicators, their windows, and the direction in which each relates to fit. Publish the whole vector and do not collapse it into a score. Disagreement between indicators is a useful finding rather than a defect: rising completion alongside rising refunds means the market is closing transactions that do not satisfy the need.
The OECD handbook on composite indicators applies directly here: before aggregating, the construct must be justified, the correlation structure of the indicators examined, and the sensitivity of the result to the choice of weights demonstrated. None of these conditions is met on current data, so no composite score is published.
Anti-example
An average five-star rating is announced as matching quality. Ratings are given after fulfilment and reflect courtesy, delivery speed, and the absence of conflict; a participant who received something other than what they sought typically leaves without rating anything. The metric rises while quality of fit remains unmeasured.
Limitations
- The original need of a demand participant is not directly observable; every indicator is a proxy and they can disagree with one another.
- Any weighting that collapses the indicators into one score is an editorial decision and must be published as such, together with a sensitivity analysis.
- The indicators observe different horizons: completion is visible within days and repeat engagement within months, so recent cohorts look systematically worse.
- The construct has not been validated against data; until independent methodological review it remains a working measurement model rather than an established metric.
Related metrics
transaction-completion-rate, refund-rate, buyer-repeat-rate, qualified-response-rate
Context sources
These works provide two-sided-market context; the formula above remains an author-defined working definition.
- Kong-Pin Chen, Yen-Chi Huang. A Search-Matching Model of the Buyer-Seller Platforms (2012).
- Hanna Halaburda, Mikolaj Jan Piskorski, Pinar Yildirim. Competing by Restricting Choice: The Case of Matching Platforms (2017).
- OECD, European Union, European Commission Joint Research Centre. Handbook on Constructing Composite Indicators: Methodology and User Guide (2008).