Procurement analytics: useful metrics and common traps

Antoine SimonPublished 2026-09-13Updated 2026-09-133 min readv1.0.0

Commercial leaders need procurement analytics that support a market or resource decision. A rising notice count is not enough to conclude that spending grew or that more addressable work exists.

Start each report with the decision, unit of analysis, buyer and classification scope, date basis, source set and missing-data limitations.

Choose metrics that answer a question

Decision Useful metric Required qualification
Is our review queue useful? Relevant reviewed notices / all reviewed notices Measures precision, not market recall
Is demand expanding in our segment? Distinct comparable procedures or lots over time Stable sources, stages and date basis
Who buys the category? Buyers and observed comparable awards Identity matching and award coverage
Is supply concentrated? Supplier share of observed awards or value State count versus value denominator
Are we preparing earlier? Working time remaining at qualification Official deadline and review timestamp
Is the bid process improving? Wins / decided bids in a defined cohort Separate unresolved, withdrawn and cancelled cases

These are proposed operating metrics. Define them before comparing periods so that a change in data collection does not masquerade as commercial growth.

Keep notices, procedures and lots separate

A competition, amendment and award can describe stages of the same procurement. Summing them as three opportunities inflates the market. Conversely, counting only procedures can conceal distinct lots your team could pursue.

State which unit is being counted and how duplicates and versions are handled. The Open Contracting Data Standard provides a useful distinction between tender, award, contract and implementation data.

Do not label every published amount as spend

Distinguish estimated value, awarded value, framework ceiling and actual payments. Keep original currency and tax basis. Document conversion dates and rates if converting currencies. Do not add values with incompatible scopes or count the same total once for every supplier.

Missing amounts are not zero. Publish the share of records with usable values beside totals. A total based on 30 of 100 awards describes observed valued awards, not the entire market.

Illustrative comparison: one quarter contains 100 awards and 30 usable values; the next contains 110 awards and 90 usable values. An increase in summed value may reflect improved disclosure. Inspect a comparable cohort before claiming demand growth.

Make cross-country comparisons defensible

Compare similar sources, procedures, sectors, periods and publication obligations. Check whether classification changes or newly added sources explain the trend. Use publication date for publication trends; use award date for award cohorts when available. Do not silently mix the two.

The European Commission's public procurement scoreboard publishes indicators with methodological context. Consult that context before using its measures as a benchmark for a private sales funnel.

A single-bid indicator is a signal to investigate. It does not independently establish corruption, a supplier's win probability or the accessibility of your next tender.

Put uncertainty on the same page as the chart

Add source names, retrieval date, included and excluded stages, record count, missing-value share, uncertain identities and known coverage gaps. If a denominator is unavailable, label the metric unavailable. Do not replace it with the number of records that happened to be returned.

For leadership, end with a concrete decision: investigate a buyer cohort, expand a source check, change qualification capacity or retain the current market focus. Avoid presenting a dashboard as a conclusion.

Your next action

Audit a report using the source coverage checklist. Use buyer intelligence for account decisions and award-to-renewal analysis for forward-looking research.

Reviewed 13 September 2026. Examples are illustrative and contain no measured Duke customer outcomes.