Use reviews to judge the experience of working with an SEO agency, and case studies to examine the results it can explain. Neither proves what will happen for you. Check who the work was for, when it happened, what changed, how results were measured and whether a relevant client can confirm the account.
Five stars and a graph going upwards. It looks reassuring.
But those two things might answer different questions. The review might tell you the agency was helpful during a website build. The graph might show more visitors, not more customers. Neither necessarily tells you whether a monthly SEO fee is a good decision for your business.
You do not need to become an SEO specialist to check the evidence. You do need to ask what the evidence actually supports.
A useful review describes a real experience in enough detail to help you understand the service. Was it ongoing SEO, a new website, a one-off audit or Google Ads? Were they talking about the first meeting or a working relationship over time?
Look for specifics about communication, delivery, understanding the business and how problems were handled. A short thank-you can be genuine, but it tells you less than an account of what happened after several months.
Check dates as well. An agency can change its team, service or ways of working. Older reviews are not worthless, but ask whether they describe the people and service you would be buying now.
A review is someone’s account. Even a genuine, detailed review is not an independently checked profit-and-loss statement.
No, not reliably. Repeated wording, a sudden cluster or vague praise can be reasons to look more closely. They are not proof that a review is fake. A new business may ask customers for reviews together; a genuine customer may only write one sentence.
Look across more than one source, and compare the review with the service being sold. If there is a concern, ask a neutral question rather than turning a suspicion into an accusation.
The CMA’s current guidance covers fake reviews, concealed incentivised reviews and misleading publication of consumer reviews. Google Maps also prohibits reviews that are not based on a genuine experience and reviews posted because of incentives. Platform rules and the law are not interchangeable. [1][2]
For a buyer, the practical question is simple: is this genuine experience, and is anything important about the relationship or incentive being left out?
A good case study makes the result understandable, not just impressive. It should let you follow the connection between the starting problem, the work and the claimed outcome.
Ask for:
An anonymised example can still be useful. Client confidentiality is a reasonable limit. The agency should still be able to explain the method, the time period and the definitions without exposing private customer data.
A screenshot is supporting material. On its own, it may not show the full period, filters, channel or business outcome.
Start with what the business needed. Rankings and visitors can be useful evidence of progress, but they are not the same as enquiries, customers or profit.
| Claim | Evidence worth asking for | What it does not prove by itself |
|---|---|---|
| Rankings improved | Named relevant searches, location, dates and the tracking method | That the searches attracted suitable customers |
| Organic traffic grew | Analytics or Search Console source, comparison period, filters and brand/non-brand context | That visitors made worthwhile enquiries |
| Enquiries increased | A definition of an enquiry, source tracking and treatment of spam or duplicates | That every enquiry became a paying customer |
| Revenue increased | Recorded sales, attribution method, comparison period and costs | That SEO alone caused the increase, or that the work was profitable |
| A client recommends the agency | A relevant, recent account of the work and relationship | That the same approach will work for your business |
The aim is not to demand a perfect scientific experiment. It is to stop a small piece of evidence being made to carry a much bigger claim.
A percentage can sound dramatic when the starting number is tiny. Ask for the actual numbers as well as the percentage, and check that periods are comparable.
Seasonal demand matters. So do a new service, changed prices, a larger sales team, a website rebuild or paid campaigns running alongside SEO. None automatically invalidates the case study. They change what you can safely conclude from it.
Also ask whether tracking changed. More recorded enquiries after fixing a broken tracking setup may partly reflect better measurement, not more new business.
If several channels helped the same customer, attribution rules decide how credit is assigned. Google Analytics explains that data-driven attribution can distribute credit across multiple contacts before a conversion. That is a useful reminder to ask how the agency counted its results. [3]
An agency does not need an identical client to be useful. It does need to explain why its experience is relevant.
A national ecommerce shop and a local service business have different demand, buying processes and measures of success. A successful campaign with a large budget is not a plan for a small business with limited capacity.
Ask what carries across: the customer problem, search competition, location, type of website or work required. Then ask what does not.
If the answer is only that both businesses wanted to rank on Google, keep asking.
Ask any agency to put you in touch with a current client. The introduction should be agreed with the client. Do not assume the agency can hand over a client’s private contact details or analytics.
Useful questions include:
Ask about the relationship as well as the headline result. An agency can produce an attractive report and still be difficult to work with.
Google’s own advice on hiring an SEO includes checking business references and asking whether clients received useful service, found the SEO easy to work with and saw positive results. [4]
Pause if the agency guarantees a number-one Google ranking, cannot explain where its figures came from, or turns every sensible question into a sales objection. Google says no one can guarantee a number-one ranking. [4]
Be wary of a case study that leaves out the dates, a graph with no clear labels, or a traffic claim presented as proof of sales. Ask for clarification first. A weak explanation is a reason to keep investigating, not permission to label a firm dishonest.
Equally, a refusal to disclose private client information is not evidence of poor results. Look for an explanation that protects the client while still giving you something useful to assess.
A small review count does not decide the issue. Ask about relevant experience, who will do the work and how the proposed plan follows from your website and business.
A clearly scoped audit or initial piece of work can help you assess how they think. Agree the deliverable, cost, access and next decision before starting. Do not give full account control simply to obtain a sales proposal. Google’s guidance suggests read-only Search Console access for an initial SEO audit. [4]
The same questions apply to Bamsh. Do not take a testimonial as proof that every result will be repeated. Ask us to explain the relevant work, the evidence and the limits.
If you are comparing proposals, use what should be included in a monthly SEO fee? alongside the reviews. You are choosing a working relationship and a plan, not the best-looking screenshot.
No. A star rating summarises submitted ratings, not a technical assessment. Read whether the reviews describe search engine optimisation or another service, and whether the customer had time to judge it. Customer satisfaction and SEO results are both relevant, but they are different evidence.
Read the complaint and the response. Is there enough detail to understand the issue? Does the response explain what was done without exposing private information? One negative review does not settle the quality of an SEO company, and a polite reply does not settle the facts. Ask about any pattern that matters to the work you would buy.
It can still help if the dates, methods, baseline and meaning of the results are clear. Ask for another way to assess the work, such as a redacted report or a reference who has agreed to speak. Do not require private account access as proof. Confidentiality is compatible with a useful explanation; it is not a reason to accept a result that cannot be explained.
Reviewer profiles can give context to Google reviews on a Google Business Profile, but they do not prove identity or results. Do not judge by the number of reviews alone. Combine the profile, the detail in the review and a relevant reference call. For a UK SEO agency, a local business case may be more relevant than a large international campaign, but location alone is not proof of a good fit.
Ask whether the improvement followed technical SEO, content changes, a website rebuild or several pieces of digital marketing at once. The agency should explain why the work addressed the problem, not present keyword research or a report as an outcome. Search rankings and a better user experience can help, but the claim still needs evidence of the particular result being sold.
[1] Competition and Markets Authority: Fake reviews guidance, April 2025. https://www.gov.uk/government/publications/fake-reviews
[2] Google Maps: Prohibited and restricted content, fake engagement. https://support.google.com/contributionpolicy/answer/7400114?hl=en
[3] Google Analytics: Cross-channel data-driven attribution. https://support.google.com/analytics/answer/11538060?hl=en
[4] Google Search Central: Do you need an SEO? https://developers.google.com/search/docs/fundamentals/do-i-need-seo
