How should you verify a GEO supplier's case studies and references?

Treat proof as a due-diligence record, not a logo wall

A case study can help a buyer understand relevant experience, but a client logo, testimonial or headline percentage is not enough on its own. The useful question is what the evidence proves about the supplier's technical ability, delivery controls or experience for this procurement—and what it does not prove.

Make the check proportionate to the value, complexity and risk of the engagement. Current GOV.UK procurement guidance says conditions of participation may assess legal and financial capacity or technical ability, including qualifications and experience, but must be proportionate. It also distinguishes those conditions from award criteria. For a GEO purchase, verify only the evidence needed for a defined decision instead of demanding a general disclosure of every client relationship.

Create one eight-part verification record

Complete the record before relying on a case study in a shortlist or award decision. The supplier can provide the first draft, but the buyer should record what was independently checked, what remains confidential and what is only asserted.

Check scope and comparability before the headline result

Confirm the supplier's actual role. A result may combine the buyer's internal team, another agency, a platform change, paid media, existing authority or work completed before the named supplier arrived. Record who produced the audit, content, technical implementation, approvals and release, and whether the proposed team or subcontractors are the same people whose work is being cited.

Then test comparability. A single-language commercial site may not evidence delivery for a multilingual university, regulated organisation or public body. Capture the markets, languages, audience, content types, technical environment, review constraints, sample and starting condition. Relevant differences do not automatically invalidate a case study, but they should change the weight placed on it.

Make every metric reproducible and bounded

Ask for the exact observation rather than a polished percentage: the baseline and follow-up dates, source, query or question set, sample-selection method, numerator, denominator, treatment of missing data and any changes to the method. Screenshots without dates or denominators may illustrate an observation but rarely establish a repeatable measurement record.

Separate delivery evidence from external outcomes. A supplier can document what it audited, recommended, changed, tested and handed over. Crawling, indexation, rankings, AI mentions, citations, traffic and enquiries also depend on external platforms, competitors, users, the buyer and time. Treat those outcomes as dated observations with limitations, not as a transferable guarantee.

• Match the metric to the claim: visibility, citation, factual accuracy, traffic and enquiry measures are not interchangeable.
• Retain the original data or export, method version and reporting cut-off where the buyer is permitted to do so.
• Record whether the figure describes the whole programme, a selected sample or one successful example.
• Mark confidential or unavailable evidence explicitly instead of replacing it with an unsupported certainty.

Run a short, structured reference check

Use a reference to verify a specific requirement, not to collect a general endorsement. Civil Service procurement guidance notes that references should normally support a particular contract and that opinions should be backed by contemporaneous records. With the supplier's and referee's permission, send the scope and questions in advance and keep the call focused on facts the referee is authorised to discuss.

Ask whether the described scope, dates, team and deliverables are accurate; how evidence and approvals were handled; how corrections, access, risks and dependencies were managed; what was handed over; and whether the reference record needs a caveat. Do not ask the referee to disclose confidential material. Record who answered, their relationship to the work, the date, unresolved gaps and any follow-up evidence.

Separate capability checks from proposal scoring

Write down whether each item is a condition for participation, evidence for an award criterion, or supporting due diligence. Avoid scoring the same claim twice. A case study may demonstrate minimum relevant experience, while the proposal's method, people and controls are assessed separately against the current requirement.

Use consistent evidence statuses such as Verified, Partially verified, Not verifiable and Not relevant. Confidentiality is not the same as failure: it may mean the buyer needs redacted evidence, an authorised reference, a bounded pilot or a lower evidence weight. Keep the rationale, conflicts and approval in the procurement record.

Ask SALIENS for an inspectable evidence plan

SALIENS provides GEO audits, AI visibility monitoring, content and technical improvements, and ongoing GEO delivery. A scoped engagement can define relevant evidence, a dated baseline, bilingual fact approval, change and test records, measurement limitations, reference permissions, and a closing decision pack. Where a buyer needs more confidence, a bounded pilot can test proposed delivery against pre-agreed acceptance criteria.

Email info@oxfordintelligence.co.uk with your website, markets and languages, the decision the evidence must support, required comparators, confidentiality constraints and review date. SALIENS is operated by Oxford Intelligence Limited. Case studies and references can support due diligence but do not guarantee crawling, indexation, AI mentions, citations, rankings, traffic or enquiries. This guide is a procurement framework, not legal or commercial advice.