Score the evidence or score the pitch with ecommerce SEO review methodology

What to take away

A shop in Leeds comparing three agencies should score written evidence, not slide decks, and should record what each supplier would not disclose.
Separate three inputs in every scorevendor claims, independent sources such as the ASA non-compliant advertisers list, and your own editorial judgement.
Weight criteria before you see any pitch, so price cannot quietly override gaps in measurement or data handling.
Publish what was not assessed. A review that hides its limits is marketing, not methodology.
Keep the whole file, scores and notes together, so a later dispute can be traced to a dated record. The parent guide to reviews and comparison methods for 2027 sets out how the wider review programme is organised.

Why UK buyers need a written method

UK store owners may compare suppliers through calls, proposals and references. The problem is memory: by the time a contract is signed, it may be unclear which supplier answered the measurement questions or what evidence supported the answer.

A written method fixes that. It states the criteria, weights, accepted evidence and scoring scale before any supplier is told what matters. It also makes the review repeatable when another supplier is added later.

A business-population estimate does not assess a supplier's search performance. This method compares the evidence for a specific store brief, not the size of the potential market.

What we assess and what we do not

We assess documents a supplier or client supplies: proposals, case notes, measurement plans, data protection statements, contract terms and named references. For search-performance claims, we record the date, scope, metric and source, and check whether the evidence supports the claim being made.

Rankings, organic traffic and revenue:ask for Google Search Console data showing queries, pages, clicks, impressions and average position, alongside Google Analytics 4 (GA4) landing-page sessions and configured conversions or revenue. Record the date range and attribution limits; one measure does not establish the other.
Technical audits, speed and indexation:check the audit scope and sample URLs. A Screaming Frog crawl can surface crawl and on-page issues; Google Search Console can help check indexing and Core Web Vitals, while PageSpeed Insights reports page performance. Record what was actually tested, not just a claim that an audit was completed.
Links and visibility estimates:Ahrefs and Semrush can provide third-party backlink and visibility estimates. Treat these as estimates from those tools, not as Google data or proof of rankings.
Product pages and feeds:assess evidence of work on product and category pages, structured data and product feeds. Where Merchant Center is used, ask for its feed and item diagnostics and record which products and markets were covered.

We run no hands-on tests of live work, no audits of client accounts and no supplier interviews on a reader's behalf. Where a claim cannot be verified from the supplied documents or public evidence, we label it a vendor claim and score it accordingly.

Separating claims, independent sources and judgement

Every criterion gets one of three labels.

Vendor claimthe supplier's own statement, scored on specificity and evidence offered.
Independent evidencea relevant, dated source outside the supplier's claim, such as a regulator or an analytics export supplied for checking.
Editorial judgementour reasoned view, scored on stated reasoning and disclosed limits.

The ASA non-compliant online advertisers list can inform a check of advertising conduct; it does not verify search-performance results. The ICO B2B marketing guidance helps frame questions about consent, lawful basis and suppression lists for direct marketing. The Competition and Markets Authority's guidance on online reviews and endorsements is relevant when presenting reviews or endorsements; it is not proof of an agency's performance.

Scoring weights we apply

Weights are fixed in advance. A typical UK store brief assigns 25 per cent to measurement and reporting. Evidence quality, data protection and marketing compliance take 20 per cent each, and price transparency takes 15 per cent. The measurement criterion should examine the quality and relevance of the search evidence, not just whether a report exists.

Each criterion is scored 0 to 4. Zero means no evidence was supplied. Four means a dated, checkable document answers the question. Multiply each score by its weight and add the weighted scores; record the totals alongside the raw scores so a reader can reweight them.

For an illustrative Leeds shop brief, suppose a supplier provides dated, checkable evidence for every criterion, including search reporting, technical checks, feed work and its data-handling and marketing controls. Score each criterion 4; applying the stated weights gives a weighted total of 4 on the 0–4 scale. This demonstrates the calculation, not a finding about a real supplier. If evidence is missing, score that criterion 0.

Price sits at 15 per cent deliberately. It matters, but a low quote with no measurement plan is a risk, not a bargain. Teams comparing store types should read the notes on ecommerce SEO product comparison in England before fixing their own weights.

Glossary

Evidence weightthe share of the total score a criterion can contribute.
Vendor claiman assertion supplied by the provider without independent support.
Scoring scalethe fixed 0 to 4 band used for every criterion.
Non-assessed itemanything outside the stated scope, recorded so readers are not misled.
Organic trafficvisits attributed to unpaid search, measured in an analytics platform according to its configuration.
Indexationwhether a search engine has included a page in its index; a crawl alone does not establish this.
Core Web VitalsGoogle's measures of real-world page experience, which should be assessed for the stated pages and available data.

Common questions

Do you test the providers you review?

No. We assess supplied documents and public sources only, and we say so in each review. No hands-on testing, account access or trial campaign is claimed.

How is price treated in the score?

Price is one weighted criterion, usually 15 per cent, scored on transparency and completeness rather than being lowest. Hidden cost lines reduce the score.

What happens when evidence is missing?

Missing evidence scores zero for that criterion and is listed as a non-assessed item. The gap is reported, not filled with an assumption.

Can a reader change the weights?

Yes. Raw scores are published alongside weighted totals, so a reader can apply their own weights and see how the ranking shifts.