Press release distribution key metrics framework

Press release distribution key metrics should answer one decision while keeping each delivery lane visible. Use separate records for delivery, publication and audience impact: a newsroom page, outreach dispatch, technical receipt, paid placement and independent article are not interchangeable.

This non-ranked taxonomy covers observable operational and evidence-quality states. It excludes invented industry targets, supplier rankings, advertising-value equivalence and unsupported causal or financial claims. Observable response and outcome measures are included; a causal impact claim needs a separate evaluation. No account or dataset was tested, and no conflicts or supplier funding are known.

What to take away

Start with one complete card

For every measure, record purpose; exact event; eligible population; numerator; denominator; unit; start and stop clock; timezone; source system, query and version; deduplication; missingness; exclusions; evidence owner; retention; uncertainty; buyer threshold; action; and review trigger. Leave the threshold empty until the organisation has an approved baseline and tolerance.

ONS defines dimensions of official-statistics quality that include relevance, accuracy and reliability, timeliness, accessibility, and coherence and comparability. These ideas can challenge a private measure's fitness for purpose, but ONS does not validate a commercial dashboard.

Use three evidence lanes: delivery covers approvals, dispatches and technical status; publication covers owned pages, paid placements and independently observed coverage; impact covers audience response and outcomes. Evidence in one lane does not establish evidence in another.

AMEC's Integrated Evaluation Framework distinguishes outputs, outtakes, outcomes and impact; the Barcelona Principles 3.0 call for measuring outcomes and impact. These are useful distinctions, not universal numeric targets.

1. Source-copy readiness

Purpose: decide whether the announcement is authorised for any route. Event: final approval of an identified source-copy version; eligible population: every required approval. Report evidenced approvals over required approvals in approval records, excluding superseded versions.

The editor owns the ledger, and unresolved factual, rights, advertising, privacy or accessibility gates keep the copy on hold; security or sector gates hold it.

2. Newsroom publication integrity

Purpose: verify the buyer-controlled page. The event is a rendered page matching the approved copy. Count matching pages over intended pages, recording the system, source ID, publication time, exclusions and corrections. Deduplicate by canonical page and source version.

A correct live page proves only owned publication. It says nothing about discovery, readership or response.

3. Outreach dispatch integrity

Purpose: reconcile authorised sends. The event is a recorded dispatch to an eligible, unsuppressed recipient under the approved contact-set version. Report evidenced dispatches over approved records after exclusions, by unique recipient-route pair. State how retries and aliases are handled.

The ICO's business-to-business marketing page explains that recipient type and communication method affect PECR treatment and that UK GDPR applies to personal-data processing. The guidance is under review after the Data (Use and Access) Act. A send count never establishes permission or attention.

4. Technical status reconciliation

Purpose: find failures between submission and a contract-defined endpoint. Retain the supplier's exact labels without translating them into publication. The population is authorised submissions under one edition and configuration, counted by submission-endpoint pair and deduplicated by buyer and supplier transaction IDs.

The buyer owns the mapping and stores supplier documentation, export date, missing endpoints and failure codes. The response is investigation or rework, not a claim of coverage.

5. Paid-placement conformance

Purpose: check that the controlled placement matches its order. Count placements with the specified location, asset version and approved label over known contracted placements. CAP recognition rules require marketing communications and advertorials to be identifiable within the Code's scope.

The advertising reviewer owns classification. A conforming placement is not an editorial endorsement or evidence of audience effect.

6. Monitoring observation quality

Purpose: describe what a defined search found. The event is one retained result matching the query and inclusion rule. State searched sources, language, period, geography, query/version, access limits, duplication key, exclusions and missingness. Unit: qualifying item, not person reached. Independent publication remains an editorial decision even when found by buyer monitoring.

7. Correction closure

Purpose: ensure known errors follow a controlled route. Report evidenced closures over eligible accepted issues, and preserve original and corrected states rather than overwriting history.

ICO accuracy guidance is relevant when personal data is involved and emphasises clear source and status. Qualified privacy review decides applicability.

8. Audience response and impact

Purpose: record observable audience response, outcomes and a defined audience change without treating publication as effect. For response, count sessions arriving through campaign-tagged links; for an outcome, report completed target actions, such as an enquiry, over those sessions.

For impact, define a change such as audience awareness or intended behaviour and compare it before and after among a specified audience, stating the survey question, population, timing, numerator and denominator. Keep the tracking method, exclusions and missingness with the evidence. A before-and-after change is not proof the release caused it; causal claims require a separate evaluation and appropriate comparison.

Reading the set

Report each card separately with freshness, evidence gaps and its next decision. Security logs, finance records and causal evaluations need their own owners. If a definition, denominator or retained query is missing, show unknown and stop the affected decision.

Before you act

Common questions

What does a correct live page actually prove?

It proves only owned publication, meaning the buyer-controlled page rendered and matched the approved copy. It says nothing about discovery, readership or response, so it cannot support claims about audience reach or impact. Treat it as publication evidence, not proof of audience effect.

Why should the buyer threshold stay empty at first?

The article says to leave the threshold empty until the organisation has an approved baseline and tolerance. Without those, any threshold is arbitrary and cannot support a decision. The card still records purpose, event, population, numerator, denominator, unit and review trigger.

What should happen when a definition or denominator is missing?

Show unknown and stop the affected decision. The article says if a definition, denominator or retained query is missing, report it as unknown rather than guessing. Each card should also carry freshness, evidence gaps and its next decision, with separate owners for security logs, finance records and causal evaluations.