A five-star average is a starting clue, not a completed background check. One reviewer may care most about a beautiful home page. You may care about migrating 400 URLs, preserving form data, training three editors, or getting support after an employee leaves. The useful question is not “do people like this designer?” It is “what evidence shows this provider can handle a project like mine, under conditions I can accept?”

Use three evidence bucketsConfidence grows when public feedback, observable work, and private references tell a consistent story.
01Public patternReviews across time, sources, project types, and star levels.
02Observable proofReal business identity, live website, credited scope, and dated case details.
03Direct accountA recent client's description of process, problems, launch, and support.

No bucket is perfect. Public reviews can be incomplete or manipulated. A live site may have changed after launch. A reference is usually selected because the relationship went well. Use the overlap, and keep the checks proportional to the decision. A modest brochure site may need a shorter review than a high-cost platform migration, but both deserve more than a star count. You can find the rest of the selection process in the hiring and project planning hub.

Begin with identity, not sentiment

Verify a sample of public reviews

01

Choose reviews across time

Include recent and older feedback, detailed and brief comments, and at least one review below the top rating when available.

02

Look for a real business connection

Check whether the reviewer name, company, and project appear on a credible business site or other public record. Do not dig for private personal information.

03

Find the work

Open the business website and compare it with the provider's portfolio or case study. Note whether the provider states its exact role.

04

Check timing

A review from years ago may still be genuine but may not describe the current team, process, technology, or support model.

05

Record what is verifiable

Separate facts such as “the site launched” from judgments such as “communication was amazing” and outcome claims such as “leads doubled.”

Read the pattern instead of collecting adjectives

Turn repeated review language into a question you can test.

Review patternPossible meaningWhat to verify
Clients repeatedly mention explanations and updatesCommunication may be a real strengthAsk which channel, cadence, and decision records the provider uses
Most praise focuses on speedThe process may be efficient or the projects may be smallAsk what scope and client readiness made the timeline possible
Reviews praise creativity but say little about outcomes or operationsVisual work may be strong while delivery evidence is thinTest forms, mobile tasks, content depth, and post-launch responsibilities
Several reviews mention delays but still rate highlyThe provider may recover well, or the schedule may be unreliableAsk who caused the delays and how expectations were reset
Every review uses nearly identical phrasesThe provider may have given customers a prompt, or the feedback may lack independenceCompare dates, identities, detail, and other sources before drawing a conclusion
A negative review receives a calm, specific responseThe provider may handle conflict constructivelyLook for accountability without exposing client information or attacking the reviewer

Read responses as carefully as reviews. A thoughtful response can acknowledge the experience, correct a factual issue without humiliating the customer, and offer a private route to resolve details. A hostile response, disclosure of confidential information, or threat over ordinary criticism is a risk signal. A generic “we are sorry you feel that way” repeated beneath every complaint is not proof of repair.

Know what review policies can and cannot tell you

The Federal Trade Commission's Consumer Reviews and Testimonials Rule addresses practices including fake or false reviews, buying reviews conditioned on positive or negative sentiment, certain undisclosed insider reviews, falsely independent company-controlled review sites, certain review suppression, and fake social influence indicators. Google Maps policy says reviews should reflect genuine, unbiased experiences and prohibits paid, incentivized, conflict-based, and other fake engagement. These rules help you recognize unacceptable conduct; they do not certify that every review still visible on a platform is accurate.

  • A “verified” platform label may confirm a platform event, not every statement in the review or the quality of the work.
  • A large review count can reflect years in business, project volume, a deliberate request process, or a mixed service line; it does not automatically indicate better fit.
  • A small review count can mean a small practice, confidential clients, a new business name, or limited follow-up; ask for context rather than filling the gap with assumptions.
  • Reviews hosted only on the provider's own site are selected marketing content. Ask where the originals appeared and whether the wording is complete.
  • Testimonials from employees, family members, partners, or compensated endorsers need clear context. Do not treat an undisclosed connection as an independent customer account.
VISUAL CHECKPOINT · PlanningUse three evidence buckets

Confidence grows when public feedback, observable work, and private references tell a consistent story.

Ask for references that match your risk

Request two or three references after you narrow the field, not as a first-step burden for every provider. Explain what makes a useful match: a similar budget range, a comparable migration, the same type of integration, a regulated industry, multiple stakeholders, or ongoing care. Ask for at least one client whose launch is recent enough to reflect the current process and, if long-term support matters, one client who has worked with the provider for a year or more. The provider should obtain permission before sharing contact details.

Use a short reference conversation with open questions

Ask for examples and sequences rather than yes-or-no praise.

TopicQuestion to askListen for
Starting pointWhat problem were you trying to solve, and what did the provider agree to deliver?A scope similar enough to inform your decision
SelectionWhy did you choose this provider, and which expectation was most important?The evidence that mattered before the project
CommunicationHow did you know what was happening and what the team needed from you?Cadence, documentation, ownership, and clarity under pressure
ChangesTell me about a change or surprise and how price, timing, or scope was handledA written process rather than informal billing or silent delay
DeadlinesWhich milestones moved, why, and how was the revised plan communicated?Specific accountability on both sides
QualityWhat did you have to correct before and after launch?Testing, responsiveness, and a realistic view of defects
SupportWhat happens today when you need an update or something breaks?Current response path, limits, and account access
HindsightWhat would you prepare, ask, or write differently if you started again?The advice a polished testimonial leaves out

A 15-minute conversation is usually enough. Introduce yourself, confirm that it is still a good time, explain the kind of project you are considering, and keep your questions focused. Do not ask the reference to reveal confidential pricing, credentials, disputes, customer data, or internal business information. Take notes on facts and examples rather than trying to capture a private call for marketing use.

Verify claimed results without demanding a guarantee

Put an outcome claim in context

01

Define the metric

“More leads” could mean form submissions, calls, qualified opportunities, booked work, or every message including spam.

02

Find the baseline and period

Ask what the measure was before launch, when measurement began, and how long the reported change lasted.

03

Name other changes

Advertising, seasonality, pricing, service expansion, staffing, reputation, and sales follow-up can affect the same result.

04

Confirm the provider's role

The web designer may have built the site while another team produced content, managed paid media, or handled conversion testing.

05

Ask what is transferable

A thoughtful provider should explain the process it can repeat while refusing to promise that your market will produce the same number.

The reference can confirm whether a case study fairly describes the work, but it should not be asked to prove your future return. Use results as evidence of a method and an honest reporting practice. If a provider promises a specific ranking, traffic increase, or sales number before reviewing your business, competition, offer, data, and constraints, the promise deserves more scrutiny than the testimonial.

Cross-check the story against live work and documents

A reference may say the provider was organized. Ask the provider to show the communication rhythm, milestone view, or redacted example of a decision record. A review may praise mobile design. Test the current mobile site, while remembering the client may have edited it. A case study may describe a custom build. Confirm what was custom and which platform or reusable system supported it. The portfolio review guide gives you a structured live-site test, while the provider's website design process should explain how similar work would be handled for you.

Red flags and reasonable explanations

Signal to investigateContext that may explain it
Sudden clusterMany vague top ratings appear within a few daysThe provider may have sent a one-time request to past clients; identities and detail should still be credible
No referencesThe provider refuses every direct client conversationConfidential work or a young business can limit options, but other proof should be offered
Only perfect storiesNo project ever changed, slipped, or needed correctionA curated list will lean positive, but candid discussion should still include constraints and learning
Review mismatchReviewers or projects cannot be connected to the claimed serviceThe provider may have rebranded, subcontracted, or served a confidential client; ask for a clear timeline and role
PressureYou are told to ignore concerns and sign immediately because many reviewers are happyStrong demand can affect scheduling, but it should not remove due diligence or contract review

Due diligence should produce a decision, not an endless investigation. Write your three highest project risks, the evidence you found for each, and the remaining uncertainty. If a provider has consistent public feedback, relevant live work, candid references, a clear contract, and reasonable answers about the gaps, you have a stronger basis than a star average alone. If the story changes each time you ask, pause before paying a deposit.

How many web designer references should I call?

Two or three well-matched references are usually more useful than a long list. Choose references that reflect your main risk, such as migration, complex integrations, regulated content, many stakeholders, or ongoing support.

What if a web designer will not share references?

Ask why and request alternative evidence such as independently published reviews, named live projects, redacted process records, partner references, or a smaller paid discovery phase. Confidentiality can be legitimate, but it should not require blind trust.

Can I trust reviews on a designer's own website?

Treat them as selected testimonials. Ask where the original appeared, whether the wording is complete, what project it describes, and whether the provider has permission to use it. Compare it with independent sources.

Should one negative review disqualify a web designer?

Not by itself. Consider the complaint's specificity, the provider's response, whether the problem repeats, and what the contract would do about the same issue. A credible response to a real problem can be informative.

Is it appropriate to ask a reference about price?

Ask whether the final cost matched the approved scope and how changes were handled. Do not pressure a reference to reveal confidential dollar amounts. Your own proposal should stand on its stated scope, price, assumptions, and terms.