A four-panel social scheduling dashboard that keeps the platform boundary visible

A social media scheduling reporting dashboard should show what decision a measure supports and whether the evidence is fit for that use. Begin with a metric register. Do not begin with chart colours.

What to take away

Create the metric register

Give each measure a stable identifier and version. The record contains the exact event, eligible population, numerator, denominator, unit, date range, timezone, exclusions, source system, query or export, owner and uncertainty. It also says whether the value is an operational observation, supplier-reported platform value, internal outcome, financial measure or attribution output.

Define OPS-RELEASE-MATCH as the live item matching the approved revision. The population is all items due in the report period. The numerator is exact matches; the denominator is all due items. The unit is a percentage with counts.

Dates use Europe/London. Formal pre-release withdrawals are excluded and counted separately. Sources are approval, schedule and live-capture records. The operations analyst owns the query; missing captures are uncertainty.

Blank formula: OPS-RELEASE-MATCH = exact live matches / eligible due items x 100.

Use four panels

The first panel covers delivery. Show eligible due items, match status, monitoring completion and unresolved defects. The second holds platform observations, each labelled with product, surface, export date and supplier definition. The third contains destination outcomes from the organisation's governed systems. The fourth is an evidence panel for missingness, revisions, definition breaks and attribution strength.

Avoid a blended score. A failed approval match is an operational defect even if platform response looks favourable. A high supplier count cannot cancel a privacy or evidence failure.

ONS defines statistical quality through fitness for purpose, including relevance, accuracy, timeliness, clarity, coherence and comparability. An internal dashboard is not official statistics, but these dimensions expose weak joins and misleading comparisons.

Preserve the platform boundary

If a supplier reports impressions, reach or interactions, retain its wording. Note organic or paid scope, account, audience, timezone, processing delay and export version. Never use one platform's denominator to calculate another platform's rate.

The LinkedIn content analytics page, checked on 6 September 2026, calls impressions and members reached estimates and defines its engagement-rate components. That is a dated product record, not a general definition and not an England benchmark.

Expose missing data and privacy choices

The dashboard should show unavailable records, rejected joins, untagged traffic and data intentionally not collected. The ICO's storage and access technologies guidance was finalised in April 2026 and includes analytics and advertising-measurement questions. A qualified reviewer must assess the actual tags, identifiers, purposes and controls. Measurement convenience does not override that gate.

Make change readable

Put the current value beside its numerator, denominator and prior comparable period. Mark any change in event rule, platform surface, query, coverage or attribution window as a series break. Do not present a percentage movement where either period has incompatible definitions.

The Government Analysis Function's uncertainty guidance recommends explaining data sources, bias, definitions, adjustments and what users can and cannot conclude. Put the material limitation next to the chart, not behind a tooltip nobody opens.

Attach decisions and review triggers

Each panel names an owner, permitted action and trigger for review. A release mismatch can pause scheduling. A supplier-definition change can suspend comparison. A broken outcome join can prevent value reporting. Weak attribution can block causal wording.

At close, archive the metric version, source extracts, query, access controls, adjustments and sign-off. A dashboard screenshot is not enough to reproduce a figure. Re-run quality assurance after a platform change, source migration, privacy decision or unexplained discontinuity.

Before you act

Common questions

What should a metric register contain?

Each measure gets a stable identifier and version. The record holds the exact event, eligible population, numerator, denominator, unit, date range, timezone, exclusions, source system, query or export, owner and uncertainty. It also states whether the value is an operational observation, supplier-reported platform value, internal outcome, financial measure or attribution output.

Why should a dashboard avoid a blended score?

A failed approval match is an operational defect even if platform response looks favourable. A high supplier count cannot cancel a privacy or evidence failure. Keeping delivery, platform observations, destination outcomes and evidence in separate panels stops one strong number hiding a weak join or a definition break.

What should happen when a platform changes?

Re-run quality assurance after a platform change, source migration, privacy decision or unexplained discontinuity. Mark any change in event rule, platform surface, query, coverage or attribution window as a series break. A supplier-definition change can suspend comparison, and a broken outcome join can prevent value reporting.