Competitor research for corporate communications: keep register facts apart from claims

Corporate communications competitor research should compare evidence for one buyer decision, not organisations that use similar words. Northmere Services is a fictional organisation researching proposed opening-hour changes at one England support site. The checklist compares research routes and supplier claims; it reports no market share, demand, capability or reputation.

Use pass, fail or unresolved for each line. Attach a source ID, access date, evidence owner and recheck trigger. Never turn absence of evidence into a favourable score.

Fix the comparison boundary

• Define the buyer job: population mapping, qualitative interviews, questionnaire design, fieldwork, analysis or accessible reporting. These are different services.
• Record the current in-house or manual route. A supplier is not the baseline merely because it has a website.
• State the English site and eligible stakeholder populations. Employees, customers, investors, suppliers, communities, journalists and regulators must not be counted as one audience.
• Freeze common evidence fields: legal identity, scope, method, delivery unit, data roles, access, rights, security, acceptance, correction, export and exit.

Separate four evidence classes

Register fact: Companies House data products can show fields such as company name, number, status, registered office and SIC code. The official register disclaimer says filed information is not verified or validated by Companies House. A record therefore does not prove current trade, England delivery, research competence, solvency or suitability.

Supplier statement: copy the exact claim, page, date, service name and geographic wording. Mark testimonials, portfolios and case studies as first-party material. Do not rewrite a claim as an independent finding.

Independent authoritative record: use a regulator, court, government or standards source only for what it establishes. The absence of a ruling is not approval. A code applying to official statistics is not proof that a private supplier followed it.

Observed buyer test: admit this class only when an authorised, controlled exercise has actually occurred. No such test occurred for this article. All performance fields are unresolved.

Check lawful and ethical collection

• Ask where every personal contact or participant record came from, its purpose and who controls it.
• Require the participant-information text, recruitment route, consent or other assessed basis, objection route, suppression process and retention schedule.
• Confirm whether employee responses can be linked to managers, human-resources files or performance decisions.
• Identify potential special-category data and the named reviewer responsible for the Article 6 and Article 9 questions.
• Reject a proposal that treats publicly visible business contact details as unrestricted research or marketing permission.

The ICO's right-to-be-informed guidance sets out privacy information that may be required when personal data is collected. The page currently notes changes to UK data law, so publication-day review is necessary. Its worker-monitoring guidance is also relevant when a proposed tool or method can observe workers beyond the declared research purpose.

Test method quality and access

• Demand the population, frame, recruitment numbers, refusals, withdrawals, non-response and exclusions behind any claimed finding.
• Inspect question wording, order, pilot evidence, coding rules, contradictions and uncertainty.
• Ask which accessible routes were offered and which configured artefacts were tested.
• Keep a claim about a research platform separate from evidence about the supplier's researchers.
• Stop if raw data cannot be exported, corrections cannot be traced or deletion cannot be evidenced.

The Government Analysis Function's questionnaire design guidance explains how question order, respondent burden and cognitive testing can affect survey work. It is a public-sector method source, not certification of a bidder. Use it to form evidence questions and leave unknowns visible.

Close without a reputational verdict

Compare admitted evidence in a contradiction log. A polished methodology page may conflict with a vague proposal; a registered entity may use another trading name; a claimed accessible platform may not have been tested in the proposed configuration. Give the bidder a correction route and record changed evidence.

Do not publish ratings or allegations from this desk exercise. The decision can approve further due diligence, reject an evidenced mismatch or remain on hold. Named corporate, privacy, IP, accessibility, security and research reviewers must clear their own gates before procurement or publication.