mac, apple, desktop, computer, laptop, screen, notebook, workplace, work, keyboard, office, macbook, technology, logo, iphone, mouse, watch, clock, media, brand, office work, communication, miruth, mac, mac, brand, brand, brand, brand, brand
Photo by HOCKULUS on Pixabay

Audience research

Part of How to plan branding audience research that can still disagree with you, before you hire an agency

How to research rival branding agencies with bounded evidence and lawful limits

A protocol for branding agencies competitor research: keep register facts, supplier claims, independent evidence and inference apart, within lawful limits.

Branding agencies competitor research should answer a narrow decision, such as whether a defined service is already offered to a specified buyer group. No agency was assessed for this article: it is a collection protocol, not a market finding.

What to take away

  • Define the service, buyer, geography and evidence period before you search for rival branding agencies.
  • Keep register facts, supplier claims, independent evidence and editorial inference in separate columns.
  • Name the legal limits that applythe Companies Act 2006 for the register, the CAP Code for advertising claims, the Copyright and Rights in Databases Regulations 1997 for reuse, and the UK GDPR, the Data Protection Act 2018 and PECR for personal data.
  • A registered company or case study alone does not prove capability or client satisfaction.
  • Publish a comparison matrix only when every agency was checked against the same criteria and date window.

Set the universe before searching

  • Write the service, buyer, England geography and evidence period.
  • Decide whether the unit is a legal entity, trading name, office or delivery team.
  • Build the first long list from named UK sources rather than search guesses: the IPO trade mark register, trade bodies such as the Institute of Practitioners in Advertising and the Design Business Association, trade press such as Campaign and Design Week, and award shortlists such as D&AD.
  • Include a candidate only when dated public evidence meets that rule.
  • Record duplicates, unclear identities, dissolved entities and exclusions.

The Companies Act 2006 requires companies to file the information that makes the register public, and the Companies House guide to searching the register, updated on 9 June 2026, explains that public company data and filings can be viewed without charge. Work through the free 'Search the register' service on the Companies House website, looking a company up by name or number to check the legal record, registered address, officers, filing history and accounts. The register lists legal entities, not marketing businesses, so match every trading name to its company number before treating it as a candidate. Do not infer current trading, service quality or brand expertise from registration.

Keep four evidence classes separate

First, register facts go in a factual column with the retrieval date. Second, an agency website, proposal or case study is a supplier claim, so capture exact wording, page and date without making it an independent result.

Third, an official contract notice, court record or regulator ruling may supply independent evidence within its own scope. Fourth, editorial inference must state the reasoning and uncertainty.

The matrix carries one row per candidate and five columns: legal entity, register fact, supplier claim, independent evidence and editorial inference. A filled row would look like this, using placeholders rather than real data: Register fact, legal name and company number from Companies House, retrieved [date]. Supplier claim, website wording on [page], captured [date]. Independent evidence, none found. Inference, claim unverified.

The CAP Code - the Committee of Advertising Practice's UK Code of Non-broadcast Advertising and Direct & Promotional Marketing - is the instrument the Advertising Standards Authority enforces for non-broadcast claims, and the ASA's misleading-advertising guidance sets out the substantiation and comparison boundaries it applies. It can inform how a rival's claim is assessed; it does not certify an agency. The absence of a published ruling is not evidence of good conduct.

Limit collection and reuse

  • Name the documented purpose and fields before opening sources.
  • Prefer a manual, reviewable sample unless a qualified reviewer approves automation.
  • Read site terms, access controls and licences before copying material.
  • Avoid collecting personal details that do not support the research decision.
  • Store URLs and short factual notes rather than duplicating whole databases.

The Copyright and Rights in Databases Regulations 1997 create the sui generis database right, and the IPO's database-rights guidance explains that copyright and database right can protect different aspects of a database. It is general UK information, so a qualified IP lawyer should review any planned extraction or reuse. The IPO trade mark register is also one of the first places to check a rival's name: search it for the word or logo you found, and record the mark, its classes, its owner and its status before relying on the name in your own comparisons.

Personal data on a public page does not lose protection because it is easy to find. The ICO's data protection principles come from the UK GDPR and the Data Protection Act 2018, so apply them to purpose, minimisation, accuracy, retention and security, and note the lawful basis for holding any named individual.

Keep competitor analysis separate from prospecting. Direct marketing by electronic means is governed by the Privacy and Electronic Communications Regulations 2003 (PECR), which the ICO's direct-marketing guidance covers alongside the UK GDPR; research notes are not consent and do not turn into a contact list.

Record conflicts and unknowns

State who performed the review, whether the publisher has supplier, affiliate or client relationships, and which evidence was inaccessible. The MRS business-to-business research guidance is a professional standard for relevant members, not a compliance guarantee or legal opinion.

Create a change log for every included record, showing page checked, capture date, reviewer, claim and evidence class. Add the next review date, and if a service page changes, preserve the earlier source note but mark the claim superseded.

Publish a matrix only if every included agency has been checked against the same date window and criteria. Leave prices, capacity, outcomes, accessibility, security and client sentiment unknown unless directly supported.

Before you act

  • Write the service, buyer, geography and evidence period first.
  • Record the retrieval date beside every register fact.
  • Capture exact claim wording, page and date.
  • Read site terms and licences before copying material.
  • Keep research notes separate from any prospecting list.
  • Log the reviewer, evidence class and next review date.

Common questions

How should a reader treat an agency website case study?

Treat it as a supplier claim, not an independent result. Capture the exact wording, page and date, and do not convert it into a verified outcome. Only an official contract notice, court record or regulator ruling may supply independent evidence within its own scope.

What should happen when two legal entities share one trading identity?

Do not score them separately until ownership and delivery responsibility are understood. Unknown ownership or inaccessible evidence is a reason to pause comparison, not a negative rating. Record duplicates, unclear identities and dissolved entities as exclusions in the collection protocol.

When should a reader stop collecting competitor information?

Stop if the purpose expands, access is restricted, rights are unclear or personal data starts accumulating without an approved need.

More in Audience research