Reviews
Part of Reviewing corporate communications agencies and software on equivalent work
How we review corporate communications suppliers, and what we refuse to publish
Corporate communications review methodology for England: define the assignment, evidence rules, reviewer disclosures and limits before assessing a supplier.
A corporate communications review methodology should specify which decision the evidence can support before anyone scores a supplier. For an organisation in England, start with a defined communication assignment, its audience and the work being purchased. A tool for staff discussion and an agency handling a sensitive announcement require different tests.
What to take away
- The protocol is ready for methodological review, but all provider and agency verdicts are withheld.
- Define the communication assignment, audience and work being purchased before scoring any supplier.
- Separate essential requirements from preferences and use pass, fail or unverified with written reasons.
- Record a delivered message separately from evidence that its audience understood or acted on it.
- Label supplier materials first-party and seek a client-controlled record for independent corroboration.
Disclosure for this methodology
This is an AI-assisted desk-research draft for commsleader.co.uk, prepared on 6 September 2026. No reviewing body has been appointed and no named human reviewer has signed it off.
The exact subject is the proposed evaluation protocol below, not a completed assessment of a particular service. Research consisted of reading the two linked UK primary sources and developing editorial assessment steps.
No agency was commissioned, product account exercised or customer interviewed. There are no first-party supplier submissions, contracts, delivery logs or test recordings in this review file. Independent guidance supports the method's boundaries; independent evidence of supplier performance is absent. Hands-on coverage is therefore nil. Procurement outcomes, usability and service quality remain untested.
Publisher ownership interests, sponsorship, affiliate arrangements, free access and other conflicts have not been verified. No recommendation can rely on an assumed absence of those relationships. The protocol covers method design only; it does not evaluate any supplier's delivery record. Named accountability, conflict checks and observed evidence remain publication conditions.
Freeze the question and the assessment rules
- Write the proposed purchase as a deliverablewho should receive what information, who approves it and what the supplier must hand over.
- Record the audience's location, working arrangements and access needs.
- Agree what is outside the assignment, especially support outside normal hours, media outreach and software administration.
Freeze the assessment rules
- Write the purchase as a deliverable
- Record audience location and access needs
- Agree what is outside the assignment
- Separate essential requirements from preferences
- Use pass, fail and unverified
- Choose the evidence window before submissions
- Separate essential requirements from preferences.
- Use pass, fail and unverified for essential checks, with a written reason beside each result.
- Do not turn an unanswered question into a failure or let attractive extras compensate for a failed essential requirement.
- Choose the evidence window before collecting submissions.
- Record the service edition, contract period and named delivery team where relevant.
- If the brief changes, issue the same revision to every participant and retain both versions so the review can explain what changed.
These are proposed editorial rules, not an externally validated scoring system.
Worked example: a 12-week internal announcement for 400 staff. The frozen brief covers one all-staff email and two manager briefings. The evidence window is the four weeks after the first send. Deliverables are the email draft, the briefing slide pack and a read-receipt log.
Examine delivery and effects separately
The UK government's communications functional standard distinguishes outputs, outtakes and outcomes in monitoring. Its requirements concern government departments and arm's length bodies; private purchasers may borrow the distinction without claiming government accreditation.
Outputs, outtakes, outcomes
Outputs
- What is recorded
- Delivered message
- Evidence source
- Delivery record
- Causal claim
- Not causal
- Keep separate
- Yes
Outtakes
- What is recorded
- Audience understanding
- Evidence source
- Feedback gathered
- Causal claim
- Not causal
- Keep separate
- Yes
Outcomes
- What is recorded
- Audience action
- Evidence source
- Observed change
- Causal claim
- Needs control
- Keep separate
- Yes
- Record a delivered message separately from evidence that its audience understood or acted on it.
- Ask who gathered any feedback, which people were eligible to respond and what else changed during the observation period.
- Keep an unexplained improvement out of a causal verdict.
Collect evidence that another reviewer can inspect
Request the supplier's dated scope and a relevant work sample with permission to examine it. Label those materials first-party. For independent corroboration, seek a client-controlled record that addresses the same assignment and dates. A public-sector standard cannot corroborate a supplier's claimed delivery result.
Inspectable evidence to collect
- Supplier's dated scope
- Relevant work sample with permission
- Label supplier materials first-party
- Client-controlled record on same dates
- Save starting conditions and actions
- Record exceptions and final artefacts
Design any future trial around a bounded task, with approved material and recorded access permissions. Save the starting conditions, actions, exceptions and final artefacts. Distinguish a demonstration led by the seller from work completed by the evaluator. Neither should be described as ordinary production experience without justification.
Decide what can be published
Where a review is used in advertising, CAP Code rule 3.7 requires documentary support for objective claims. That advertising rule is not a universal certification of editorial reviews.
Have the future reviewer trace each proposed conclusion to its evidence and record disagreement. Publish the scope, limitations and unresolved questions next to any eventual verdict. The immediate action is to appoint that reviewer and approve a test brief.
We refuse to publish named supplier verdicts, numerical scores, rankings or recommendations until that reviewer signs off a documented assessment against a frozen brief. We also refuse to publish objective claims that rely only on supplier-supplied material with no client-controlled corroboration.
Before you act
- Freeze the question and the assessment rules first.
- Separate essential requirements from preferences before scoring.
- Choose the evidence window before collecting submissions.
- Keep outputs, outtakes and outcomes as separate records.
- Label supplier materials first-party and seek client-controlled records.
- Publish scope, limitations and unresolved questions beside any verdict.
Common questions
What does the article say about the status of this review?
No reviewing body has been appointed and no named human reviewer has signed it off. No agency was commissioned, no product account was exercised and no customer was interviewed, so hands-on coverage is nil.
How should a reviewer treat an unanswered question about a supplier?
The text says not to turn an unanswered question into a failure. Essential checks should use pass, fail and unverified, with a written reason beside each result. Attractive extras must not compensate for a failed essential requirement.
What evidence can support a published verdict?
Each proposed conclusion must trace to its evidence, with disagreement recorded. The scope, limitations and unresolved questions should sit next to any eventual verdict. A public-sector standard cannot corroborate a supplier's claimed delivery result.
