Collect the evidence before you write the coverage story.

Write the acceptance rule first

Decide what qualifies as a coverage item before collecting it. A workable rule names the publication window, campaign, markets, content types, and minimum evidence. It also says whether a syndicated copy is a separate item, how updates are handled, and which date controls inclusion. A reviewer should be able to apply the rule without knowing which total you hope to reach.

For example: count one unique editorial or creator item published from 1 September to 30 September, linked to the Solstice campaign, when a primary URL or dated stored copy shows the item and its source. Record reposts as distribution evidence under the original item. Keep unverified mentions in a review queue, and keep delivered products outside the coverage count unless content is observed.

Capture enough context to verify each item

A URL alone is fragile. Pages move, social posts change, and a link does not explain why the item belongs in the report. Give every candidate a stable ID and record the publication, author or creator when known, publication date, discovery date, campaign, product, format, market, primary URL, evidence type, evidence location, and the person who checked it.

Then describe only what the evidence shows. Product visible, product named, spokesperson quoted, and campaign message present are observations. Estimated audience, commercial value, and business effect are separate claims that require their own source and method. If they were not measured, leave them unknown instead of borrowing precision from the coverage file.

Deduplicate without erasing distribution

Start with exact identifiers: canonical URL, platform post ID, publication title, author, and timestamp. When two pages reproduce the same original article, choose one primary record and attach the other URLs as secondary distribution. When the reporting question is specifically about pickup across publications, report that distribution separately and disclose the rule.

Do not merge merely because titles look similar. An edited regional edition, an independently written follow-up, and a social post that cites an article may be distinct outputs. Record the relationship, compare the text and publication context, and leave the pair flagged when the decision is uncertain. The consequence of a rushed merge is lost evidence; the consequence of no review is an inflated count.

• Normalize tracking parameters before comparing URLs.
• Keep the earliest known original and list confirmed copies beneath it.
• Preserve separate items when authorship, format, or editorial treatment differs.
• Record who made the merge decision and why.
• Recalculate totals after every inclusion, exclusion, or merge.

Reconcile the fictional Solstice evidence queue

The fictional queue begins with six candidates. COV-101 is J-014's Northline Journal article with a primary URL and stored PDF. COV-102 is a creator post with a dated screenshot and profile URL. COV-103 is a newsletter mention forwarded without its header or send date. COV-104 is a syndicated copy of COV-101. COV-105 is C-008's delivered Dune tote with no observed post. COV-106 is an older unrelated article.

Four candidates match the campaign and period: COV-101 through COV-104. The syndication rule folds COV-104 under COV-101, leaving three unique coverage items. Two are evidence-ready, while COV-103 remains in the missing-proof queue. COV-105 remains one seeding activity and COV-106 is outside scope. The defensible headline is therefore two evidence-ready items, one in-scope evidence gap, one linked repost, one non-coverage send, and one out-of-scope candidate.

Prepare a reviewable, shareable evidence set

Open every primary link before the reporting cutoff and keep a dated local copy when your permissions allow it. Name files with the stable coverage ID rather than a person's private details. A screenshot should show enough context to identify the publication, account, date, and relevant passage. A note should state who observed what, when, and why no stronger artifact is available.

Before sharing, remove private contact details, internal comments, addresses, access tokens, and unrelated browser content. Keep the unredacted working file in the operator's private location and share only the evidence needed for the stated report. This is an operational privacy check, not legal advice. Record inaccessible or deleted pages as limits instead of implying they were reverified.

• Open links and files from the final evidence index.
• Confirm dates and source labels are visible.
• Redact personal and confidential context.
• Mark deleted, restricted, or unverified material explicitly.
• Freeze the evidence index used for the report.

Move from a clean register to the next decision

AMEC's Barcelona Principles 4.0 call for qualitative and quantitative analysis and transparent methods. Its evaluation taxonomy also warns against substituting an output measure for a result at another level. This checklist applies those principles narrowly: it improves the traceability of observed coverage, but it does not establish audience response, organizational impact, or causal ROI.

In cutePR, coverage and evidence are entered manually. An operator can link the reviewed records to a campaign and freeze a report snapshot, but there is no automatic monitoring, post capture, client workspace, or scheduled report delivery. Current report exports use English labels. Once the register reconciles, choose whether the next report answers a monthly operating question or a campaign decision, then apply a disclosed qualitative reading grid.

Sources & methodology

• AMEC: Barcelona Principles 4.0
• AMEC: Taxonomy of evaluation