How to Verify SEO Expert Results and Case Studies

Almost every SEO website shows case studies with big arrows and bigger percentages. Almost nobody who reads them runs a single check before signing a contract. That gap is why weak case studies keep working year after year.

I am an independent SEO expert, and I hold an unpopular position here: I would rather be verified than trusted. Trust is a shortcut. Verification is evidence. The checks below take minutes and cost nothing. This article is one connected checklist: what a real case study contains, how to check ranking claims yourself, how to read reviews, which answers should worry you, and which red flags end the conversation. Every test in it works on the agency pitching you next week. Every test in it works on me, and I want you to run them hardest on me.

Why Most SEO Case Studies Prove Nothing on Their Own

A case study is a claim wearing professional formatting. Formatting is not evidence. Before you weigh anything in a portfolio, identify the exact claim being made, because each claim type needs different proof.

Rankings, impressions, clicks, traffic, leads, revenue, links, and technical health get bundled under the phrase "great results," yet they are different claims with different evidence. Moving one page from position 40 to position 8 for a low-volume query proves ranking ability; it says nothing about revenue. Doubling impressions can coexist with flat clicks when the queries are informational. Naming which result you are actually being shown is the first verification step.

Then separate absolute change from percentage change, and sitewide scope from page-level or query-level scope. Growth that looks dramatic as a percentage can be trivial in absolute visits, a trap I return to below. Also separate causation from chronology: traffic rose after the engagement started, and dozens of other things changed in the same window.

This is why a truthful testimonial can still mislead you. The client may honestly be delighted. Their result may still be unrepresentative of what the typical client receives, or of what your business will receive. Honest reporting and typical results are two different standards, and you deserve to know both.

Start With the Result You Actually Need Evidence For

Start by writing down the outcome you need proven. Local map visibility for one city? Ecommerce revenue from organic search? Rankings preserved through a site migration? Lead flow for a service business? Recovery from a penalty? Each outcome requires different evidence, and a generic "we grew traffic" story matches none of them.

Next, compare the case study's conditions to yours. A result produced under a different business model, market, budget, or starting condition demonstrates competence, and competence matters. It does not predict your outcome. An expert who has never handled a migration can still be excellent at local visibility, and the reverse is equally common. Ask which of their results came from businesses shaped like yours.

Finally, check how old the evidence is. Search itself changes: results pages now carry AI Overviews, more ads, and different result types than they did a few years ago, and tactics that produced wins five years ago operated in a different environment. For a sense of how far the ground has moved, read a 30-year SEO expert on why AI search isn't business as usual. Recent evidence beats old evidence, and old evidence still has to pass every check below.

What a Verifiable SEO Case Study Must Include

A verifiable case study supplies eight things before you ask for them. When a portfolio piece is missing any item below, that gap is your first question.

1. A real client, named or anonymized with a stated reason, published with permission
2. A defined baseline with starting absolute values for every metric shown
3. A complete, seasonally comparable time window
4. A stated scope: sitewide, page-level, or query-level
5. The specific interventions the expert performed
6. Disclosed external changes: algorithm updates, seasonality, PR, paid media, redesigns
7. Current live evidence you can check today
8. An honest statement of whether the result is typical

Baseline, Time Window, and Starting Values

Ask for the start date, the end date, and the starting absolute value behind every percentage. "Organic traffic grew 300%" is incomplete. "Ten thousand monthly sessions grew to forty thousand between March and October, measured against the same months of the prior year" is a claim you can evaluate. Percentages without absolutes hide the base, and the base determines what the growth means.

The baseline also needs history. Request the full pre-engagement trend rather than one weak week selected as the "before" picture, plus disclosure of anything that shaped the starting point: a prior migration, a penalty, a redesign, or a previous agency's exit. Compare like periods, too. A retail site's October measured against its January flatters whatever happened in between.

What Was Actually Done — and What Else Changed

Ask what the expert actually did, in plain language, and which reported change each action plausibly caused. Content work, technical fixes, link acquisition, and information architecture each leave different footprints in the data. Then ask what else changed during the window: algorithm updates, a site redesign, a PR spike, paid campaigns, seasonal demand. Any of these can move traffic without the expert's involvement, and an honest case study names them.

Percentage growth deserves extra scrutiny because low bases inflate it. Ten monthly visits growing to forty is 300% growth. One hundred thousand growing to one hundred ten thousand is 10% growth. The first headline reads better; the second added ten thousand monthly visits.

Two final checks apply here. Confirm the client permitted the case study, because publishing client results without permission tells you how the expert treats confidential data. Then ask whether this result is typical. A single peak cherry-picked from a portfolio of flat accounts is marketing, not evidence.

How Do You Check Ranking and Traffic Claims Yourself?

You check a ranking claim by getting the exact keyword, location, device, and date behind the headline number, then searching it yourself. Open an incognito window, set the location for local claims, and look. A free rank tracker does the same job while removing personalization. Two minutes of effort converts the claim into something either real or not.

Screenshots need metadata before they count as evidence. A valid ranking screenshot states the query, the location, the device, the date, the result type (standard organic, local pack, AI Overview), and how the position was measured. A graph without axis labels and a screenshot without a query are decoration. Rankings also move weekly, so a position captured months ago describes the past, not the present.

Traffic claims follow the same logic. Ask which tool produced the numbers, whether the scope is sitewide or page-level, and whether the source is the client's analytics or the agency's own reporting. Then ask for current evidence: a keyword the client ranks for today, a page earning traffic this month. Live evidence you can re-check beats archived screenshots, and any expert who maintains results has fresh examples ready. I know which of my results a prospect can re-check today, and I expect the same readiness from anyone I would hire.

Can You Trust SEO Testimonials and Reviews?

You can trust testimonials after reading them for pattern, specificity, and history, not before. Fabricated reviews follow patterns you can spot in minutes.

Fabrication signals cluster. A run of five-star reviews posted inside one short window, written in near-identical generic language ("great to work with, highly recommend"), from profiles with no other review history, is a purchased pattern. Genuine review profiles look different: they accumulate steadily over years, they vary in length and tone, and they contain specific details ("fixed the indexing problem two agencies missed") that generic praise never includes. Video testimonials carry weight only when a named, findable person is speaking.

The client check is just as fast. Search the named business, confirm an active website, and confirm a Google Business Profile matching the details in the case study. A client who cannot be found anywhere is a story, not a reference. Initials-only clients and "a leading retailer" deserve less trust, though anonymization is sometimes legitimate; what you should require is an explanation, ideally alongside a client you can speak with directly. Prospects have asked to speak with my clients before hiring me, and that request has never bothered me. An expert bothered by verification has told you something important.

Does the Expert's Own Presence Back Up Their Claims?

The expert's own website is their always-on case study, and it should meet the standard they sell. You can inspect it without anyone's permission.

Visibility first. The expert's site need not rank first for every head term in the industry; the SEO niche is among the most contested in search, and plenty of strong practitioners concentrate on client work. What the site should show is intentional structure: connected topic coverage demonstrating command of the field, not random posts chasing last month's trend. If they teach keyword research, their content should reflect keyword research discipline. If they sell content strategy, their library should show a content strategy at work.

Then read their current thinking, because search has moved and their advice should have moved with it. Start with what an SEO expert actually does and what SEO experts think about AI Overviews; my SEO expert hub collects the rest of the series. Those are my pages, and this test applies to them fully. Search my name, read the cluster, and check whether the site practices what those pages teach. I would rather you arrive as a client already convinced by evidence you verified yourself than won over by a pitch.

What Questions Should You Ask Before Hiring?

Ask process questions, not promise questions. A practitioner with a real process answers in specifics; a practitioner without one changes the subject.

Questions That Reveal a Real Process

Four areas of questioning expose whether a process exists: keyword research, technical SEO, measurement, and reporting. Ask how they find the queries worth winning, and listen for method rather than tool names. Ask how they approach technical SEO, and listen for crawlability, indexing, speed, and site structure rather than "we optimize everything." Ask how success gets measured and what appears in the monthly report, and listen for metrics tied to your business rather than vanity counts. Whether you hire an SEO consultant or a full agency, the same questions apply; scale changes deliverables, not the existence of a process.

Answers That Should Worry You

Vague answers are the warning sign. "Proprietary methods," an inability to explain proposed changes in terms you understand, and resistance to showing a sample report all signal a process that cannot survive scrutiny.

Google's own hiring guidance, published in its "Do You Need an SEO?" documentation, agrees: ask for explanations, make sure you understand the proposed changes, and walk away from ranking guarantees. From the search side, Google's own podcast on SEO expertise covers the same ground. Currency matters as well, because search now includes AI-driven results and expertise that stopped updating years ago is aging fast; SEO experts moving into AI search shows what that transition looks like when a practitioner takes it seriously. Ask the expert directly how AI search changes their work. "Haven't needed to think about it" tells you when their expertise stopped.

Which Red Flags Should End the Conversation?

Six signals end my evaluation of an SEO provider, and I apply them to my own industry without exception.

1. Guaranteed rankings. Google states plainly that no one can guarantee a #1 ranking on Google and that organic positions cannot be purchased. A guarantee is a disqualifier, not confidence.
2. Instant results. "Traffic doubles in 30 days" describes either a temporary spike or a tactic that risks penalties. Durable SEO compounds over months.
3. Evidence without metadata. Screenshots missing query, date, location, or device are decoration, not proof.
4. Anonymous clients without explanation. Initials and "a leading retailer" sometimes have legitimate reasons behind them. An expert who cannot explain the reason, or connect you to any verifiable client, fails the check.
5. Evasive methodology. "Proprietary process" is not an answer. You deserve to understand what gets done to your site before it is done.
6. Unconfirmed methods. Ask directly whether the work qualifies as white hat SEO. A hedged answer, or no answer, means walk away, because penalties land on your domain, not theirs.

The Bottom Line

Verification costs minutes and prevents months of wasted budget. Name the result you need proven, demand baselines with absolute values, re-check one ranking claim yourself, read reviews for pattern and specificity, inspect the expert's own presence, and keep asking process questions until you understand the answers. None of this requires SEO knowledge. It requires treating a case study as a claim to be tested rather than a trophy to be admired.

Use the full checklist on me. Search my name, check what ranks, read the hub, and re-check anything I publish about my results. I would rather be verified than trusted, and the clients I want are the ones who verify. When you want a first step that produces evidence rather than opinions, start with an SEO audit: it documents where your site stands today, the priority fixes, and what verifiable progress looks like. Then judge me on the work.

Frequently Asked Questions