Public review standard
How GrowthLayer reviews CRO agencies
GrowthLayer does not assign a universal “best agency” score. We verify what a provider publicly discloses, label who made each claim, and evaluate fit for a buyer's specific traffic, budget, maturity, and job to be done.
Reviewed by Atticus Li · Methodology version 2026.07.22
The publication gate
An agency can appear in the broader directory while its review is pending. It cannot become an indexable, evidence-reviewed profile until every gate below passes.
- 1. At least two current, provider-owned source pages support visible claims.
- 2. At least one current source contains a named public proof case.
- 3. Pricing is labeled as published, directly confirmed, or not publicly listed.
- 4. The profile states both best-fit and not-ideal-for scenarios.
- 5. A named reviewer, review date, and buyer verification questions are present.
- 6. Published pricing cannot be used without a current pricing source.
How the reviewed sample is selected
The candidate set is a purposive sample of specialist and adjacent providers a CRO buyer may encounter across enterprise experimentation, SaaS, ecommerce, landing-page, and lower-traffic decisions. It is not a census, random sample, or claim to be the market's largest directory.
Payment, sponsorship, an introduction arrangement, or a provider request cannot buy inclusion, placement, a proof grade, or a fit result. Providers that fail the current source, proof, freshness, or publication gates remain unscored until the missing evidence is resolved.
Ecommerce named-proof relevance
Business-model fit and case-study relevance are reviewed separately. GrowthLayer manually classifies one visible named proof claim for each exact-fit ecommerce provider; it does not infer relevance from a client logo or a general service claim.
- Named ecommerce case
- The reviewed public proof names an ecommerce or direct-to-consumer business and describes an ecommerce journey or outcome.
- Ecommerce case · integrated scope
- The named case is ecommerce, but CRO was delivered with other material work such as redesign, development, SEO, paid media, or branding. The public result cannot be attributed to CRO alone.
- Named case from another business model
- The provider has current first-party evidence that it serves ecommerce, but the named proof selected for review demonstrates SaaS, lead generation, subscriptions, or another model instead.
A classification remains current only while the proof source is provider-owned, provider-reported, and inside the evidence freshness window. Missing current proof falls back to “comparison pending”; it does not become a favorable assumption or a claim that the agency lacks ecommerce experience. The reviewed provider set is purposive, not a market census.
Agency proof grades
Proof grades measure the completeness and independence of public evidence. They do not measure agency talent, strategic quality, or the result a future client will get.
- Grade A
- An independently reviewable outcome with a named client, baseline, timeframe, and method.
- Grade B
- A named, quantified provider case study with incomplete public methodological detail.
- Grade C
- A named testimonial, client logo, unquantified case, or materially confounded multi-channel outcome.
- Grade D
- No current, reviewable public proof was found.
A named provider case study is still provider-reported unless GrowthLayer can inspect independent underlying evidence. A high percentage or famous logo does not make a claim independent.
Claim attribution
verified_primaryVerified on a current first-party page
provider_reportedProvider-reported; not independently verified
independent_supportSupported by an independent source
editorial_inferenceGrowthLayer editorial inference
unknownUnknown
Freshness rules
90 days
Profile review
90 days
Pricing evidence
180 days
Core service and proof evidence
When evidence expires, the claim is not silently carried forward. A profile can be removed from recommendations or indexation until the source is reviewed again.
How scenario matching works
Unknown values receive no score. They are never converted into a favorable guess. Hard budget, traffic, service, and proof conflicts can exclude a provider even when other criteria match.
An eligible shortlist result requires at least 80% weighted fit-model coverage and no hard conflict. An unknown budget can remain eligible only when pricing is visibly labeled unresolved for direct confirmation. Results are displayed alphabetically; “match across evidenced attributes” is not a first-to-third ranking or a universal rating. It does not measure delivery ownership, contract terms, team capacity, or predict conversion lift.
What not to infer
- — “Contact for pricing” does not mean expensive; it means no current public figure was used.
- — Missing traffic guidance does not prove a provider can or cannot serve low-traffic sites.
- — Grade B proof can describe a strong result, but the public method is not independently reviewable.
- — Inclusion is not endorsement, and placement is not a paid ranking.
- — A fit result is a shortlist input, not a substitute for diligence and reference checks.
Corrections and conflicts
Providers may request factual corrections without payment. A correction must include a current source or direct evidence that can be attributed. GrowthLayer does not accept payment for inclusion or placement. If an introduction ever produces compensation, that relationship must be disclosed to the buyer and cannot change proof grades, publication gates, ordering, or fit rules.
GrowthLayer also sells a five-day evidence audit for mature internal teams. The assessment may show that offer only after it calculates agency matches from the buyer's answers. This financial interest cannot affect provider inclusion, proof grades, alphabetical ordering, match eligibility, or match scores.