The Executive Reporting and Review-Cadence Standard: Turning Dashboards into Decision Systems

An executive dashboard should help a leader spot an important change, understand the evidence and decide what to do. Agree the thresholds, review schedule and action record so the team can also check whether its decisions worked.

1. The Market Connection Context

Customers experience the consequences of reporting design even when they never see the report. A late service warning, an unexplained fall in demand, an unresolved margin variance or a repeated handoff failure reaches leadership through reporting, or stays hidden until value has already leaked. The operating question is therefore not “Which dashboard should we build?” It is “Which recurring decisions must leadership make, from which governed evidence, under which conditions, and how will the organisation remember what it decided?”

This standard converts an existing set of decision and metric contracts into a governed review loop:

Decision demand → governed measures → contextual narrative → human-approved threshold
→ scheduled review → recorded disposition and action → outcome observation → learning

By [decision date], the CEO will classify the proposed reporting system OPERATE WITHIN LIMIT, VALIDATE, CHANGE, DEFER, RETIRE or STOP. The classification applies to a named forum, decision set, metric versions and review period. It is not certification of the underlying data estate, technical architecture or every management report.

The Executive Decision Inventory: Mapping Signals, Thresholds, Owners and Actions owns the inventory and definition of decisions. The Metric Contract and Semantic Model Standard: One Definition, Known Limits and Accountable Ownership owns metric meaning, lineage and threshold contracts. The Executive Reporting and Review-Cadence Standard: Turning Dashboards into Decision Systems owns the management interface, explanatory narrative, review cadence, action memory and learning loop. The Management Intelligence Architecture Playbook: From Source Systems to Decision Interface owns source-to-interface architecture. A report cannot repair an undefined decision, a disputed metric or an unsafe data pipeline.

Working files

• Reporting Charter and Forum Authority Register
• Decision Page and Narrative Specification
• Threshold, Review and Action Ledger
• Reporting Learning and Change Register

The UK Government Analysis Function guidance on communicating quality, uncertainty and change supports visible caveats and explanations of uncertainty. The page currently identifies itself as under review and says it has not been updated for accessibility legislation; it is not treated here as current accessibility authority. The ONS dashboard guidance addresses user need, contextual writing, maintenance and accessibility. The Government Analysis Function dashboard testing guidance covers representative user tasks, keyboard operation, zoom and manual accessibility testing. The W3C Web Content Accessibility Guidelines 2.2 provides the normative accessibility criteria used for the interface gate. These sources are VERIFIED primary standards or official guidance. Their relevance to this operating method is INFERRED; the Abrasim sequence, templates and dispositions are PROPOSED and their local effectiveness REQUIRES VALIDATION.

2. The Leadership vs. Delegation Boundary

The reader recognises that the interface, diagnostic layers, narrative, thresholds, meeting cadence and action log must be designed as one management system.

An internal team can run this standard when it has the authority to convene decision owners, inspect definitions and evidence, challenge narratives and close actions. Independent facilitation is useful when reports defend functions rather than decisions, thresholds threaten incentives, or the same unresolved issue recurs. Legal, privacy, security, accounting, regulatory, statistical and accessibility specialists retain their authority.

Stop if the report disguises missing evidence as zero, removes material segments, presents inaccessible colour-only meaning, changes definitions without version notice, preselects a conclusion, or assigns action to a forum rather than a person. A meeting held on schedule is not proof of a decision. A clicked dashboard is not proof of adoption. A threshold breach is a prompt for human review, not automatic authority to act.

When a narrower specialist or simpler intervention is sufficient

Use a reporting or accessibility specialist when the decision contract is accepted and the remaining issue is one bounded page, narrative, keyboard or export defect. Use a metric or data specialist when a definition, lineage or refresh defect is isolated; route governance, privacy, security, legal or accounting conclusions to the qualified authority. A simpler facilitated review is proportionate when owners already hold the evidence and authority.

Questions for evaluating a provider

Ask how the provider will preserve executive and metric-owner authority, test representative decision tasks and accessibility, expose uncertainty and failed evidence, leave editable specifications and action records, and return any material meaning, resource or risk decision to the named internal authority. Reject a proposal that equates views or dashboard delivery with decision use or adoption.

3. The Step-by-Step Execution Standard

Step 01: Charter the reporting decision system

Purpose and decision served Bound the reporting system around recurring executive decisions and the value at risk.

Accountable owner CEO or executive chair of the forum.

Delegated lead and contributors Reporting Lead with CFO, COO, decision owners, Metric Steward and Independent Challenger.

Preconditions, inputs and access Accepted decision inventory, current reports, forum terms, action records and access approvals.

Actions

1. Name the forum and executive chair.
2. Link each proposed page to one recurring decision.
3. State the population, period and economic consequence.
4. Record exclusions, conflicts and stop conditions.

Evidence to retain Signed charter, decision references, current packs, authority record and conflict declaration.

Decision or scoring rule Proceed only when each page serves a named decision; do not score aesthetics as decision fitness.

Output and template Charter rows in the Reporting Charter and Forum Authority Register.

Acceptance gate CHARTER ACCEPTED or RE-CHARTER.

Leadership checkpoint The chair confirms what the forum must decide and what it will not own.

Human control point Only the executive chair accepts scope and value-at-risk boundaries.

Failure modes and safeguards Safeguard against an omnibus dashboard by rejecting pages without a decision owner.

Change, stop or escalation condition Stop if authority, decision demand or permitted evidence cannot be established.

Handoff to the next step Transfer the accepted charter to the authority design.

Step 02: Establish forum authority and challenge

Purpose and decision served Make attendance, decision rights, evidence ownership and protected challenge explicit.

Accountable owner Executive chair.

Delegated lead and contributors Governance Lead with Secretariat, decision owners, control owners and Independent Challenger.

Preconditions, inputs and access Accepted charter, organisation authorities, delegations and conflict declarations.

Actions

1. Assign chair, decision owner, evidence owner and action owner roles.
2. Define quorum and substitute rules.
3. Protect dissent and specialist escalation.
4. Record access and confidentiality limits.

Evidence to retain Authority matrix, quorum rule, access approvals, conflicts and escalation map.

Decision or scoring rule No report enters operation without a human authorised to decide and another able to challenge.

Output and template Authority rows in the Reporting Charter and Forum Authority Register.

Acceptance gate AUTHORITY COMPLETE or HOLD.

Leadership checkpoint The chair tests whether absent roles would make a decision non-binding.

Human control point Delegation cannot transfer statutory or professional authority.

Failure modes and safeguards Prevent collective ambiguity by naming one accountable person per decision and action.

Change, stop or escalation condition Escalate unresolved conflicts or missing control authority.

Handoff to the next step Give the authorised decision set to the page-design team.

Step 03: Design one page around one decision

Purpose and decision served Create a reading order that moves from condition to explanation, options and action.

Accountable owner Decision Owner.

Delegated lead and contributors Report Designer with Reporting Lead, Metric Steward, operators and accessibility specialist.

Preconditions, inputs and access Decision definition, user needs, forum constraints and approved metric candidates.

Actions

1. Write the decision question as the page heading.
2. Place current condition and threshold before diagnostic detail.
3. Reserve space for uncertainty, segments and narrative.
4. Place disposition and action record in the reading flow.

Evidence to retain Page wireframe, reading-order annotation, user-needs record and exclusion rationale.

Decision or scoring rule Accept only if a leader can state the decision, condition and required response without navigating elsewhere.

Output and template Page row in the Decision Page and Narrative Specification.

Acceptance gate DECISION PAGE COHERENT or REDESIGN.

Leadership checkpoint The decision owner performs a cold read using only the proposed page.

Human control point Visual hierarchy cannot determine materiality without owner approval.

Failure modes and safeguards Avoid metric walls, decorative charts and hidden filters by documenting every element's decision purpose.

Change, stop or escalation condition Return to The Executive Decision Inventory: Mapping Signals, Thresholds, Owners and Actions when the decision itself remains ambiguous.

Handoff to the next step Pass accepted page purposes to metric-contract verification.

Step 04: Bind every displayed measure to a metric contract

Purpose and decision served Prevent familiar labels from concealing competing meanings or unsupported totals.

Accountable owner CFO or Metric Definition Owner.

Delegated lead and contributors Metric Steward with Finance, Analytics, Data Owner and Report Designer.

Preconditions, inputs and access Page specification and The Metric Contract and Semantic Model Standard: One Definition, Known Limits and Accountable Ownership contracts with versions, limitations and owners.

Actions

1. Link every figure to a contract version.
2. Surface population, period, unit and exclusions.
3. Reconcile displayed totals to authoritative sources.
4. mark unresolved meaning or quality gaps visibly.

Evidence to retain Contract links, reconciliation evidence, version map, limitation text and exception record.

Decision or scoring rule Do not display a material measure as decision-fit when its contract is absent or expired.

Output and template Metric rows in the Decision Page and Narrative Specification.

Acceptance gate MEASURES BOUND or VALIDATE.

Leadership checkpoint The CFO accepts the reconciled meanings and visible limitations.

Human control point Semantic acceptance remains with the metric authority, not the interface builder.

Failure modes and safeguards Control silent substitutions by pinning contract versions and recording impact.

Change, stop or escalation condition Return disputed or unreproducible measures to The Metric Contract and Semantic Model Standard: One Definition, Known Limits and Accountable Ownership.

Handoff to the next step Give governed measures to narrative design.

Step 05: Specify the explanatory narrative

Purpose and decision served Explain what changed, why it may have changed, what remains uncertain and what decision follows.

Accountable owner Decision Owner.

Delegated lead and contributors Narrative Editor with analysts, operators, Finance and Independent Challenger.

Preconditions, inputs and access Governed measures, prior-period context, known events, segment evidence and claim classifications.

Actions

1. State the observed condition separately from its explanation.
2. Identify contributing evidence and contradictory evidence.
3. label hypotheses and validation gaps.
4. Name the requested disposition and consequence of delay.

Evidence to retain Narrative draft, claim-evidence map, contradiction log, author and timestamp.

Decision or scoring rule Every causal statement needs evidence or an explicit INFERRED or REQUIRES VALIDATION status.

Output and template Narrative rows in the Decision Page and Narrative Specification.

Acceptance gate NARRATIVE DECISION-FIT or REVISE.

Leadership checkpoint The decision owner distinguishes observation, interpretation and recommendation aloud.

Human control point Generated summaries require named human verification before circulation.

Failure modes and safeguards Prevent confident fiction by retaining contradictions, missing evidence and source limits near the claim.

Change, stop or escalation condition Stop circulation if a material explanation cannot be traced.

Handoff to the next step Transfer the page and narrative to threshold design.

Step 06: Define thresholds as review triggers

Purpose and decision served Translate material conditions into explicit prompts for human attention and response.

Accountable owner Executive Decision Owner.

Delegated lead and contributors Metric Steward with Finance, Operations, Risk and Analytics.

Preconditions, inputs and access Metric contracts, historical distributions, service obligations, risk limits and action capacity.

Actions

1. Define threshold population, period and direction.
2. Attach a named review action and owner.
3. Add segment and guardrail conditions.
4. Record expiry, override and recalibration rules.

Evidence to retain Threshold rationale, supporting data, approved response, guardrails and expiry.

Decision or scoring rule A threshold is incomplete without an owner, action, evidence boundary and human override route.

Output and template Threshold rows in the Threshold, Review and Action Ledger.

Acceptance gate THRESHOLDS AUTHORISED or VALIDATE.

Leadership checkpoint Leadership tests false-alarm and missed-condition consequences.

Human control point No colour, alert or model may authorise a material action by itself.

Failure modes and safeguards Avoid arbitrary targets and averages that hide mandatory segment failure.

Change, stop or escalation condition Escalate where a response exceeds delegated authority or capacity.

Handoff to the next step Pass authorised triggers to cadence design.

Step 07: Design the review cadence and agenda

Purpose and decision served Match review frequency and preparation to the speed, consequence and reversibility of decisions.

Accountable owner Executive chair.

Delegated lead and contributors Secretariat with decision owners, Reporting Lead and control owners.

Preconditions, inputs and access Thresholds, refresh timings, evidence latency, action lead times and governance calendar.

Actions

1. Set cadence from decision need rather than data availability alone.
2. Establish cut-off, pre-read and exception rules.
3. Order agenda by material decision, not function.
4. Reserve time for prior-action learning and dissent.

Evidence to retain Cadence rationale, agenda, cut-off rule, attendance rule and escalation path.

Decision or scoring rule Accept a cadence only when evidence can arrive before action becomes ineffective.

Output and template Review rows in the Threshold, Review and Action Ledger.

Acceptance gate CADENCE FIT or REDESIGN.

Leadership checkpoint The chair confirms which decisions happen in-cycle and out-of-cycle.

Human control point An automated alert may convene review but cannot replace authorised quorum.

Failure modes and safeguards Prevent meeting theatre with a decision-led agenda and cancellation rule when no decision is ready.

Change, stop or escalation condition Change cadence when evidence latency, risk or action lead time materially changes.

Handoff to the next step Transfer cadence requirements to accessible prototype testing.

Step 08: Test usability and accessibility

Purpose and decision served Ensure intended users can perceive, navigate, interpret and operate the reporting interface.

Accountable owner Reporting Product Owner.

Delegated lead and contributors Accessibility Specialist with Report Designer, representative users and Quality Lead.

Preconditions, inputs and access Prototype, representative devices, assistive technology, user tasks and WCAG 2.2 AA criteria.

Actions

1. Test keyboard order, focus, headings and labels.
2. Verify colour-independent meaning and text alternatives.
3. Test zoom, reflow, table reading and export.
4. Observe users completing decision tasks and record failures.

Evidence to retain Test protocol, issue log, screenshots, assistive-technology results and remediation evidence.

Decision or scoring rule Mandatory accessibility or decision-task failure cannot be averaged away.

Output and template Accessibility and usability rows in the Decision Page and Narrative Specification.

Acceptance gate USABLE WITHIN LIMIT or REMEDIATE.

Leadership checkpoint The owner accepts unresolved limitations or holds release.

Human control point Qualified human inspection complements automated accessibility tests.

Failure modes and safeguards Prevent mouse-only navigation, colour-only status and inaccessible exported packs.

Change, stop or escalation condition Stop operational release for a material access barrier.

Handoff to the next step Provide the remediated interface and limitations to rehearsal.

Step 09: Rehearse the complete review

Purpose and decision served Test the interface, narrative, thresholds, quorum, agenda and action capture as one system.

Accountable owner Executive chair.

Delegated lead and contributors Facilitator with Secretariat, decision owners, Metric Steward and Independent Challenger.

Preconditions, inputs and access Approved prototype, realistic controlled scenario, role assignments and stop route.

Actions

1. Run a representative decision scenario.
2. Introduce uncertainty, contradiction and a threshold breach.
3. record disposition, dissent, action and due gate.
4. Debrief delays, ambiguity and unsafe automation assumptions.

Evidence to retain Scenario, attendance, observations, decision record, defect log and remediation owner.

Decision or scoring rule Pass only when a named decision and action can be reconstructed from evidence.

Output and template Rehearsal rows in the Threshold, Review and Action Ledger.

Acceptance gate REHEARSAL PASSED or REMEDIATE AND REPEAT.

Leadership checkpoint The chair confirms the forum produced a decision rather than discussion alone.

Human control point Only authorised participants make or simulate material dispositions.

Failure modes and safeguards Do not treat a synthetic scenario as evidence of live business outcomes.

Change, stop or escalation condition Hold release when quorum, evidence or action ownership fails.

Handoff to the next step Transfer resolved defects to the executive operating decision.

Step 10: Make the operating disposition

Purpose and decision served Authorise bounded operation, further validation, change, deferral, retirement or stop.

Accountable owner CEO or named executive authority.

Delegated lead and contributors Reporting Lead prepares; CFO, COO, Risk and Independent Challenger advise.

Preconditions, inputs and access Complete evidence pack, rehearsal result, limitations, dissent and resource implications.

Actions

1. Review scope and material limitations.
2. Select one disposition.
3. Record conditions, resources and expiry separately.
4. Sign dissent, appeal and stop routes.

Evidence to retain Signed disposition, evidence references, limitations, dissent and resource decision.

Decision or scoring rule Use OPERATE WITHIN LIMIT, VALIDATE, CHANGE, DEFER, RETIRE or STOP; no composite score substitutes for judgement.

Output and template Decision row in the Reporting Learning and Change Register.

Acceptance gate DISPOSITION RECORDED.

Leadership checkpoint The executive states what is authorised, excluded and still unknown.

Human control point Final disposition and resource commitment remain separate human decisions.

Failure modes and safeguards Prevent conditional approval from being represented as universal endorsement.

Change, stop or escalation condition Stop if evidence, authority or material dissent is concealed.

Handoff to the next step Transfer the bounded decision to operational handover.

Step 11: Operate the action and decision memory

Purpose and decision served Make every forum decision, action, owner, gate and closure reconstructable.

Accountable owner Executive chair.

Delegated lead and contributors Secretariat with action owners, Reporting Lead and evidence owners.

Preconditions, inputs and access Authorised reporting system, action taxonomy, controlled repository and notification routes.

Actions

1. Record each disposition and rationale during the forum.
2. Assign one owner and evidence-based closure gate.
3. Link action to metric version and narrative.
4. Escalate overdue or invalidated commitments.

Evidence to retain Minutes, action ledger, closure evidence, exceptions and escalations.

Decision or scoring rule An action closes only on accepted evidence, not a verbal status or elapsed date.

Output and template Action rows in the Threshold, Review and Action Ledger.

Acceptance gate ACTION MEMORY ACTIVE or CONTROL FAILURE.

Leadership checkpoint The chair reviews open and invalidated commitments before new decisions.

Human control point Only the decision owner accepts closure or changes the commitment.

Failure modes and safeguards Prevent orphan actions by prohibiting group ownership and undocumented closure.

Change, stop or escalation condition Escalate repeated overdue actions or evidence that invalidates the original decision.

Handoff to the next step Feed decisions and actions into outcome learning.

Step 12: Evaluate outcomes and attribution

Purpose and decision served Learn whether the action plausibly contributed to the intended result without inventing causality.

Accountable owner Decision Owner.

Delegated lead and contributors Measurement Lead with Finance, Operations, Analytics and Independent Challenger.

Preconditions, inputs and access Baseline, action record, outcome measure, guardrails, comparison logic and known confounders.

Actions

1. Compare observed evidence with the agreed baseline.
2. Examine guardrails and affected segments.
3. Record competing explanations and evidence limits.
4. classify contribution and next learning need.

Evidence to retain Outcome analysis, guardrail results, confounder log, attribution statement and challenge notes.

Decision or scoring rule Use bounded contribution language unless the design supports a stronger causal conclusion.

Output and template Learning rows in the Reporting Learning and Change Register.

Acceptance gate LEARNING RECORDED or VALIDATE.

Leadership checkpoint Leadership considers whether to continue, change, reverse or test again.

Human control point No model assigns causal credit without qualified review of assumptions.

Failure modes and safeguards Avoid outcome cherry-picking by preserving baseline, guardrails and contrary evidence.

Change, stop or escalation condition Escalate harm, displaced effects or invalid attribution assumptions.

Handoff to the next step Transfer accepted learning to system review and change control.

Step 13: Review, change or retire the system

Purpose and decision served Keep the reporting system aligned to current decisions, definitions, users and control conditions.

Accountable owner Executive chair.

Delegated lead and contributors Reporting Product Owner with decision owners, Metric Steward, Data Owner and Accessibility Specialist.

Preconditions, inputs and access Usage evidence, decision records, action outcomes, incidents, metric changes and user feedback.

Actions

1. Review whether pages produced decisions and learning.
2. inspect stale content, changed contracts and recurring overrides.
3. Test continuing accessibility and user need.
4. Choose renew, change, validate, defer, retire or stop.

Evidence to retain Review pack, change impact, version record, retirement evidence and new charter triggers.

Decision or scoring rule Retain only elements with a current decision purpose, accepted evidence and accountable owner.

Output and template Change rows in the Reporting Learning and Change Register.

Acceptance gate RENEW, CHANGE, VALIDATE, DEFER, RETIRE or STOP.

Leadership checkpoint The chair accepts the next version and its resource boundary.

Human control point Material changes to decision, meaning, threshold or authority require renewed approval.

Failure modes and safeguards Prevent dashboard accumulation through explicit retirement and archive controls.

Change, stop or escalation condition Re-charter when the forum, decision set or operating context materially changes.

Handoff to the next step Return material changes to Step 01; archive the controlled evidence for continuity.

4. The Evidence & Measurement Checklist

Preflight and authority

• The forum and executive chair are named. Evidence: signed charter. Owner: CEO.
• Every page links to a recurring decision. Evidence: decision-page map. Owner: Reporting Lead.
• Decision and challenge rights are explicit. Evidence: authority matrix. Owner: Governance Lead.
• Quorum and substitute rules are accepted. Evidence: forum terms. Owner: Secretariat.
• Scope exclusions and conflicts are visible. Evidence: boundary register. Owner: Executive Chair.

Evidence access and quality

• Every displayed measure links to a current contract version. Evidence: metric-version map. Owner: Metric Steward.
• Material totals reconcile to authoritative sources. Evidence: reconciliation record. Owner: Finance Lead.
• Population, period, unit and exclusions are visible. Evidence: page specification. Owner: Report Designer.
• Evidence latency is compatible with the decision. Evidence: refresh assessment. Owner: Data Owner.
• Missing or uncertain evidence is labelled. Evidence: limitation register. Owner: Narrative Editor.

Method execution

• The reading order moves from condition to decision. Evidence: annotated wireframe. Owner: Report Designer.
• Observation and interpretation are separate. Evidence: narrative specification. Owner: Decision Owner.
• Each threshold names an action and owner. Evidence: threshold ledger. Owner: Metric Steward.
• Cadence follows decision need and evidence latency. Evidence: cadence rationale. Owner: Executive Chair.
• A complete review rehearsal has been recorded. Evidence: rehearsal log. Owner: Facilitator.

Governance, privacy, security and risk

• Access follows approved role and purpose. Evidence: access register. Owner: Security Owner.
• Sensitive exports retain equivalent controls. Evidence: export control test. Owner: Privacy Owner.
• Material professional mandates are preserved. Evidence: specialist sign-off. Owner: Governance Lead.
• Threshold override requires a named human. Evidence: override procedure. Owner: Risk Owner.
• Incident and stop routes have been tested. Evidence: control rehearsal. Owner: Operations Owner.

Deliverables and acceptance

• All four working files are complete and internally consistent. Evidence: controlled working pack. Owner: Reporting Lead.
• Each page has accepted purpose and limitation text. Evidence: page acceptance record. Owner: Decision Owner.
• WCAG 2.2 AA tests cover representative tasks. Evidence: accessibility test pack. Owner: Accessibility Specialist.
• The executive disposition is signed and bounded. Evidence: disposition record. Owner: CEO.
• Open defects have owners and release effects. Evidence: defect register. Owner: Reporting Product Owner.

Measurement

• Baselines use defined populations and periods. Evidence: measurement contract. Owner: Measurement Lead.
• Decision-cycle completion has a reproducible denominator. Evidence: review log. Owner: Secretariat.
• Action closure requires accepted evidence. Evidence: action ledger. Owner: Decision Owner.
• Guardrails remain separate from primary outcomes. Evidence: outcome review. Owner: Risk Owner.
• Attribution language matches the evaluation design. Evidence: attribution statement. Owner: Independent Challenger.

Decision and handover

• Every decision records rationale and evidence references. Evidence: decision log. Owner: Secretariat.
• Every action has one owner and acceptance gate. Evidence: action ledger. Owner: Executive Chair.
• Metric and page versions are linked to the decision. Evidence: version record. Owner: Reporting Lead.
• Operating support and escalation are accepted. Evidence: handover record. Owner: Operations Owner.
• Dissent and appeal routes remain available. Evidence: dissent log. Owner: Governance Lead.

Review, change and stop controls

• Review expiry and owner are recorded. Evidence: change register. Owner: Reporting Product Owner.
• Metric changes trigger consumer impact review. Evidence: change-impact record. Owner: Metric Steward.
• Accessibility is retested after material interface change. Evidence: regression test. Owner: Accessibility Specialist.
• Stale pages have an explicit retirement route. Evidence: content inventory. Owner: Reporting Lead.
• Harm, invalid evidence or lost authority can stop operation. Evidence: stop-control test. Owner: Executive Chair.

Measurement contracts

Measure A: Decision-cycle completion

Measure B: Evidence-linked action closure

Measure C: Narrative evidence fitness

Measure D: Decision-to-outcome learning coverage

Measure E: Accessibility task success

Measure F: Threshold usefulness

Evidence classifications used in this guide

• VERIFIED: directly supported by an inspected controlled record or cited primary source.
• INFERRED: a bounded interpretation from evidence with assumptions and alternatives visible.
• PROPOSED: a method, control, threshold or option awaiting authorised acceptance.
• REQUIRES VALIDATION: an important claim or condition that cannot yet be relied upon.

Source notes

1. UK Government Analysis Function: communicating quality, uncertainty and change: official guidance inspected for uncertainty and change communication; the page is currently marked under review and not updated for accessibility legislation, so it is not current accessibility authority.
2. ONS dashboard guidance: official guidance; inspected for user need, explanatory context, maintenance and accessibility limits.
3. Government Analysis Function: testing dashboards for design and accessibility: official guidance; inspected for representative tasks, manual testing, keyboard access and responsive use.
4. W3C WCAG 2.2: normative accessibility recommendation; local conformance still requires human and technical verification.
5. W3C WAI guidance on complex images: official accessibility guidance for text alternatives and equivalent information.