How to Read UX Agency Reviews: A Buyer’s Guide

Start with the question the review can answer

A client review is one person’s account of a particular engagement. It can describe how the client experienced communication, planning, responsiveness, or delivery, and may surface tradeoffs an agency portfolio leaves out. That makes it useful when you are deciding what to ask an agency or reference next.

The account has a boundary. It may describe a different service, project size, team, client, or period from the work you are considering. It may report satisfaction without saying what shipped, whether users tested it, or how the client measured an outcome. Treat the review as evidence about the engagement it describes, then connect it to your brief where the details support that connection.

Start with a practical question: which part of our decision could this review inform? It might help you understand a team’s communication habits or whether a client found a handoff usable. It is less likely to tell you whether the agency can solve your exact product problem or whether its design improved a user outcome.

Read the score with its count and distribution

A star average compresses several client accounts into a single number. The count tells you how many reviews the platform includes in that figure. It does not tell you how many clients the agency served, how many were invited to respond, or whether the reviewers represent the kind of work you want to commission. Keep the count and platform beside the score whenever you record it.

A larger set gives you more accounts to read, but it is not automatically a representative sample. Clients decide whether to respond, agencies may choose when and whom to invite, and platforms set eligibility and publication rules. A small set can still contain a highly relevant project account; a large set can still be weighted toward a particular service or type of client.

Treat the displayed count as a count of reviews included by that platform, not as a count of projects, satisfied clients, or response rate. The page may not show whether several reviews concern repeat work with one client or how many clients were invited to contribute. If that information matters, look for the platform’s collection policy or ask how the review set was assembled.

Read the rating label carefully. A profile might show an overall average, a service-specific score, a verified-review average, or a platform-adjusted number. Those figures are calculated from different groups or rules. Before comparing two agencies, establish whether the same scale, review set, time period, and calculation are being used.

Look at the individual ratings when the site exposes them. The distribution can reveal whether most reviews are clustered at the top or whether mixed experiences disappear inside the mean. A spread of scores is a prompt to read the context and explanations; it is not a formal performance interval or a statistical guarantee.

• Record the displayed rating, review count, platform, and date checked.
• Note whether the number covers all reviews or a selected subset.
• Keep a platform-adjusted score separate from a raw average.

Check what the verification label actually confirms

A verification label matters when you know what was checked. It can refer to a reviewer’s identity, a business relationship, information about the agency, or profile completeness. Those checks answer different questions. A verified company profile does not mean each project claim has been audited, and a verified reviewer does not make the opinion representative of every client.

Clutch says it checks a reviewer’s identity and work history using reviewer information and online research. Its help page says it may publish a useful review without a Verified label when it cannot confirm identity or project details; it does not disclose every step of the process. Read that label as Clutch’s status for the review, not as an audit of the work or outcome.^[3]

DesignRush describes a different process. Its methodology says agencies can submit client reviews, the platform may collect reviews from third-party sites, and reviewer checks can use LinkedIn, email, or a call. DesignRush defines its separate Verified agency badge as a check of details such as contact information, services, and portfolio ownership. Keep the review status and business badge distinct.^[5]

Understand the platform’s score and directory order

A directory position may use much more than the star average. Review data can sit alongside service focus, case studies, client experience, market presence, awards, profile information, and commercial placement. Before using a “top” list to build a shortlist, read the platform’s current explanation of how its order is produced.

Clutch says its review score considers average rating, recency, verification, and project budget. Its broader company rank also uses client and experience information and market presence. Its current methodology describes default directory placement that can put sponsored and verified providers ahead of providers ordered by organic rank. These are platform-specific ranking rules; the placement alone does not establish project quality or review manipulation.^[4]

DesignRush says it uses a Bayesian calculation with a 3.5-star baseline to smooth the average. Its organic ranking also considers items such as portfolio, awards or press, team biographies, and top services. A smoothing formula can change how a small review set appears as a number. It cannot supply missing project context or make a self-selected group of clients representative.^[5]

Keep these distinctions visible in your notes. One platform can show a raw average, another an adjusted score, and a directory may mix review signals with business data or sponsored placement. Write the platform’s own label beside any score or position you quote, and check that the methodology page is current.

Choose reviews that resemble your project

Recency matters because agencies and teams change. Project leads may have moved on, services may have shifted, and a review may describe an offer the firm no longer provides. A recent review can help you ask about the current team, but date alone does not make the project relevant to your needs.

Read the review’s project context. Look for the service, client type, product maturity, scope, duration, and named roles. A branding or marketing engagement does not establish experience in complex UX research. A small website refresh does not demonstrate work on a multi-role enterprise product. Ask which part of the case resembles your brief and who would do that work now.

In B2B work, the person writing a review is often the buyer or project stakeholder, not a user of the product being designed. A review praising communication is evidence about that person’s reported working relationship. It becomes evidence about usability only when it describes user research or user feedback, and even then you should check the underlying context.^[6]

Google Maps requires reviews to reflect a genuine experience with a place or business and prohibits paid or incentivized review activity. Even a genuine local business review may say little about a specific UX engagement. Read the review text for the service and project details before using it as evidence about design work.

Read the account for delivery details and attribution

Specific descriptions are more useful than broad praise. Look for the client’s problem, the agency’s role, the people involved, constraints, and what was delivered. “Great partner” may be sincere feedback, but it does not tell you whether the agency conducted research, designed an interface, implemented it, or supported only one phase.

Separate the working relationship from the product outcome. A client may value clear communication before a product launches; a project may ship on time without improving task completion. If a review reports a business result, note who made the claim and look for a baseline, measurement period, population, and method. Without those details, treat the result as something to verify in a conversation.

Reviews can be especially helpful for collaboration details that a case study leaves out: how a team handled changes, communicated risks, or transferred knowledge. Use those accounts to prepare questions. Do not turn one client’s experience into a universal claim about design quality or user impact.

Consider incentives and the rules around reviews

Check for a disclosure that a reviewer received a benefit, works for the agency, or has another business relationship. That context helps you understand the account; it does not automatically make the account false. When no disclosure appears, do not assume a relationship exists. Ask for clarification only when it could materially change your decision.

In the United States, the FTC’s Consumer Reviews and Testimonials Rule took effect on October 21, 2024. It addresses specified practices including fake reviews, incentives conditioned on positive or negative sentiment, certain insider reviews, and certain review suppression. Its legal scope depends on context; an unusual score or missing badge alone does not establish a violation.^[1]

UK guidance for online review sites says publishers should present a full picture, keep reviews relevant and current, explain how reviews are collected and checked, and disclose commercial relationships. The Competition and Markets Authority says its guidance is not legal advice. Treat these sources as context for understanding review integrity, and consult current rules for the activity and jurisdiction involved.^[2]

Follow up on patterns without jumping to conclusions

A cluster of reviews posted close together, repeated language, or project details that do not line up can justify a closer look. None proves that a review is fake. A legitimate group of clients may post at once, and a public profile rarely shows enough about who was invited or how reviews were collected to establish a motive.

Compare the accounts with other public evidence. Do project dates, services, and client descriptions make sense beside the agency’s portfolio? Does a client-owned source confirm a relationship or launch? If something does not line up, state the specific question and ask the agency or platform through the appropriate channel.

When you report a review, refer to the platform’s published rule and the detail that prompted your concern. In your own research record, keep the observation separate from your explanation. That discipline protects the agency and gives your team a clearer basis for making a decision.

Keep review platforms separate in your evidence record

Averaging ratings from Clutch, DesignRush, Google, and other sites produces a number with no shared definition. Platforms differ in reviewer eligibility, verification, imported content, rating calculations, time weighting, and sorting. The same client account may also appear on more than one site. If you cannot establish that reviews are independent, do not count them as separate client experiences.

If two pages show different totals or averages, keep both observations with their dates and direct links. One may be a filtered view, a newer profile, or an imported set that the other does not show. A discrepancy is a reason to check the platform’s labels and update timing; it does not explain itself and should not be resolved by averaging the numbers.

For each platform you use, note the profile URL, date checked, displayed score, count, date range, and label attached to the rating. Record the platform’s own definition of “verified.” If you cannot find one, write that the method was not located. Choose relevant reviews and note their service, client context, role, evidence, and open questions.

Keep review evidence alongside case studies, client sources, and the proposed team. A review describes a past client’s experience. An agency portfolio describes the agency’s account of its work. A client source may confirm a relationship or launch. A conversation can clarify who would deliver your project. UXComparation uses review platforms for attributed feedback and platform-specific facts, not as an editorial quality score.

• Platform and direct profile URL.
• Score, count, date range, and date checked.
• Meaning of the platform’s verification or profile badge.
• Relevant review details and the questions they raise.
• Any overlap with other sources and any unresolved conflict.

Use a worked example to guide the next conversation

Consider a fictional shortlist. Agency A has many positive reviews, most describing branding and general web work. Agency B has fewer reviews, several of which describe research and complex product workflows. Your brief calls for research on an enterprise product. The difference in review count does not settle the decision; the service context makes Agency B’s accounts more relevant to investigate.

Next, check what those accounts actually establish. Do they identify the agency’s role and the team members? Do they describe a shipped product or a discovery phase? Are the outcomes client-reported, platform-reported, or independently corroborated? Then compare the proposed team and ask both agencies the same questions about the work closest to your brief.

Your written conclusion can preserve both the signal and its limit: “Several recent reviews describe research and product design, which makes the agency relevant for a closer conversation. The reviews help us understand collaboration; they do not independently verify user outcomes.” That gives stakeholders a reason to continue without turning platform feedback into a recommendation.

If the reviews do not answer your core question, make that gap part of the next step. You might ask for a case study with a clearer account of the agency’s role, speak with a client who commissioned similar work, or meet the actual delivery team. Choose the follow-up that would change your decision, and avoid requesting references simply to collect more positive impressions.

If a review raises a question the public record cannot answer, request a relevant reference or a more detailed case walkthrough. Ask about the specific team, scope, decision, and outcome you need to understand. Record what the reference confirms and what remains uncertain.

Use reviews as one part of the buying decision

Use reviews to shape the shortlist conversation. If the accounts fit your project, ask the agency to explain the relevant work and meet the people who would deliver yours. If important details are missing, decide whether a reference, a fuller case study, or a scoped discovery phase would resolve the uncertainty. Leave irrelevant reviews out of the selection rationale.

For the wider decision, compare review accounts with project evidence, the proposed team, scope, and commercial assumptions. Our related guides explain how UXComparation uses these sources.See guide to comparing UX agencies · case study evidence guide · source standards · research methodology.