How to judge SEO advice before you spend money

If you want to judge SEO advice before you spend money, ask one question first: what evidence is this recommendation based on, and what does that evidence actually prove?

That sounds simple, but it filters out a lot of weak advice. A crawler visit is not the same as indexing. Indexing is not the same as an impression. An impression is not the same as a click. A click is not the same as an enquiry. Most confusion around UK search engine optimisation starts when those are treated as if they were interchangeable.

What UK search engine optimisation actually covers

In plain English, search engine optimisation is the work of helping search engines access, understand and present your pages, while also making those pages useful to real people. In a UK setting, that usually means four separate areas.

First, technical access. Can search engines reach the page? Is it blocked in robots.txt, noindexed, redirected, canonically pointed elsewhere, or returning an error? If Googlebot cannot fetch or process a page properly, the page may never get a fair chance in search. Google’s own starter guidance is straightforward on this point, and it also makes clear that useful, accessible content helps discovery, without guaranteeing rankings (Google SEO Starter Guide).

Second, content usefulness. Does the page clearly answer the query it seems to target? Is it original, specific and well organised? Does it explain what the reader actually needs to know, or is it padded with generic copy? This is where many SEO pitches become vague. “More content” is not a recommendation unless someone can explain which questions are currently underserved and which pages should answer them.

Third, local visibility. For businesses serving places rather than the whole country, local search matters. That includes your website, your Google Business Profile, and whether your location, opening information and service details are consistent and clear. Google states that local results are influenced by relevance, distance and prominence, and those positions are not something you can buy from Google (Google local ranking).

Fourth, measurement. This is where sensible decision-making either starts or falls apart. We need to know whether a page is crawlable, indexed, shown, clicked and useful after the click. We also need to know what remains uncertain. Search data often tells us that something changed, not why it changed.

That is the practical scope of UK search engine optimisation. It is not a magic layer that sits above your business and manufactures demand. It is a set of checks and improvements that can increase your chances of being found, understood and chosen.

Which evidence should you ask for first?

Before accepting any recommendation, ask what type of evidence supports it. In practice, most SEO claims rely on one of five things.

Crawling evidence shows that a bot requested a URL. That might come from server logs or crawl reports. This can be useful if someone is diagnosing whether Googlebot, OAI-SearchBot or GPTBot can access content, but it proves only that a request happened. It does not prove indexing, visibility or commercial value. It also matters which bot you are discussing. Search access and AI training controls are not the same choice. OpenAI documents this distinction directly, including the separation between OAI-SearchBot and GPTBot (OpenAI crawler documentation).

Indexing evidence shows whether a page is in Google’s index, or why it is not. This is often where technical work starts. If a key service page is marked “Excluded by ‘noindex’ tag”, that is a clear issue. If a thank-you page is excluded, that may be exactly what you want. Indexing evidence is about eligibility and status, not success.

Impression evidence shows that a page appeared in search results. This is useful because it tells you Google sees the page as relevant enough to show for at least some searches. But impressions alone can be misleading. A page can gain impressions because it is being shown for broader, weaker or less commercial queries.

Click evidence is closer to real demand. It tells you that someone saw your result and chose it. Even so, clicks do not prove quality of traffic, lead quality or revenue. A page can lose clicks because rankings fell, because search demand changed, or because the search results page changed around it.

Business enquiry evidence includes calls, form submissions, bookings or tracked sales. This is the evidence most buyers care about, but it is also the easiest to overclaim. Unless tracking is set up properly, the path from search to enquiry may be partly invisible. Even with tracking, causality can be messy. A person may discover you in search, return later by typing your domain, and convert on a different device.

So what should you ask for first? Usually, ask for the nearest evidence to the actual claim. If someone says “your pages are not getting indexed”, ask for indexing evidence. If they say “your titles are hurting click-through”, ask for impression and click evidence by page and query. If they say “SEO is driving leads”, ask how enquiries are being attributed and what remains untracked.

If you want a structured way to compare these signals, our guide to Search Console evidence and its limits and our search measurement baseline workbook can help you separate what is known from what is assumed.

How to read Google Search Console without jumping to conclusions

Google Search Console is often the first place we look, and it is also one of the easiest places to misread.

Start with the Page indexing report. This report tells you which URLs are indexed, which are not, and the reported reason. Google’s own documentation is clear that excluded pages are not automatically a problem, and that you should investigate the reason before deciding what to do (Google Page indexing report). A duplicate page with a canonical to the main version may be behaving correctly. A paginated URL, a filtered URL or an internal search results page may not need indexing at all.

What this report can prove is status at the URL level, plus some broad patterns. If a whole section is blocked or noindexed, that is usually visible here. What it cannot prove is that indexing a page will make it rank well, or that every indexed page deserves to exist.

Next, the Performance report. This is where people often rush to a story. Impressions rose, so the SEO must be working. Clicks fell, so something must be broken. Average position improved, so revenue should follow. None of those conclusions is safe on its own.

Impressions tell us visibility happened. Clicks tell us some users chose the result. Average position is especially easy to misread because it blends many queries, devices and locations. A page can “improve” on average while losing its best commercial terms. It can also lose clicks while gaining impressions if it starts showing for more marginal searches.

Query data needs care too. Search Console does not show every query, and it is not a full analytics platform. It is best used directionally. We use it to ask better questions. Which pages changed most? Which queries moved with them? Was the shift sitewide, page-specific, or limited to one country or device type?

Country and device filters matter for UK readers. If your business serves the UK, look at UK data specifically before drawing conclusions from global totals. A rise in impressions from outside the UK may not help your sales team at all.

Also remember that Google Search Console is about Google Search. It does not report how your pages perform in every AI answer surface, and Google states that its existing search fundamentals apply to AI features, with no special AI text file or special schema required for inclusion (Google AI features). That means anyone selling a separate “AI schema package” as a requirement for Google’s search AI features should be pressed for evidence.

If you need a practical method for diagnosing a shift before buying anything, our article on how to investigate a change in search performance sets out a step-by-step approach.

What good SEO recommendations look like

Good recommendations are specific, testable and tied to evidence. They tell you what to change, where, why, and how success or failure will be judged.

For content, a useful recommendation might say that three service pages overlap, each says nearly the same thing, and none clearly answers the most common pre-sale questions. The proposed action could be to consolidate those pages, rewrite headings around real user tasks, and add missing details such as delivery area, lead times, pricing approach or compliance information where relevant. That is much better than “publish more blogs”.

For internal links, a good recommendation identifies pages that matter commercially but are hard to reach, buried in navigation, or linked with vague anchor text such as “read more”. Google’s starter guidance supports descriptive links because they help users and search engines understand destination pages (Google SEO Starter Guide). A recommendation should therefore point to the exact pages that need stronger internal linking and suggest where those links naturally belong.

For accessibility, the signs of useful advice are practical and user-centred. Clear form labels, readable contrast, keyboard-friendly controls, logical headings and understandable navigation all help people use the site, and can also reduce friction for search visitors. W3C’s accessibility tips are a solid primary reference here (W3C accessibility design tips). Good SEO advice does not treat accessibility as a bolt-on or a score-chasing exercise.

For page experience, sound recommendations distinguish between measurable performance issues and exaggerated promises. Core Web Vitals describe loading, responsiveness and visual stability, but a score is not a ranking or sales guarantee (Core Web Vitals). Useful advice names the issue, such as oversized images, layout shifts from late-loading banners, or slow scripts on template pages, and explains which pages are affected.

For local information, good recommendations check whether your location, service area, opening details and contact routes are consistent across the site and your business profiles. They also recognise UK realities. A business serving multiple towns from one office should not fabricate addresses to chase local pack visibility. That creates trust and compliance problems, and it is not the same as improving relevance.

Most importantly, good recommendations say what remains uncertain. If a page has low impressions, the cause may be weak demand, poor relevance, indexing issues, or stronger competitors. An honest recommendation narrows the possibilities rather than pretending to know more than the data supports.

Which claims should make a UK buyer pause?

Several claims should slow you down immediately.

Guaranteed rankings are the obvious one. No one controls Google’s results pages outside paid advertising, and Google’s own guidance does not promise rankings for compliant sites (Google SEO Starter Guide). A provider may guarantee that they will carry out work, but not that Google will place you in a specific position.

Paid link packages are another danger sign. If the offer is framed as a fixed number of backlinks from generic sites, with little explanation of relevance or editorial context, treat it cautiously. The problem is not just quality. It is that the recommendation usually arrives before any evidence-led diagnosis of your site’s actual issues.

Be wary of AI recommendation guarantees too. If someone says they can guarantee inclusion in AI answers, or that blocking or allowing a particular bot will directly secure citations, ask them to separate crawling, training and answer-surface visibility. OAI-SearchBot and GPTBot are not the same thing, and crawler access is not the same as being cited. The same care applies to Google. There is no separate AI text file required for Google’s search AI features (Google AI features).

Vague traffic-to-sales claims should also make you pause. “More traffic means more sales” is not evidence. Some traffic is irrelevant. Some pages are informational by design. Some businesses have long offline buying cycles that make direct attribution difficult. Advice should explain what kind of traffic is expected to change, and how that change will be assessed.

Another warning sign is a recommendation that ignores ownership and implementation. If the issue is clearly a CMS limitation, a tracking gap or a development bottleneck, then buying another SEO dashboard may not solve anything. Sometimes the next step is not more advice, but practical implementation by your web team or another supplier. If that is where you are, it helps to know that Suffolk Web provides implementation services and quotations, and shares ownership with us through Fleeta Limited. Here on The SEO Blog, our role is to help you judge the evidence before you decide whether any service is justified.

A simple decision process before you buy tools or hire help

Use this checklist before spending money.

First, define the problem in one sentence. Is it poor indexing, weak local visibility, falling clicks, thin service pages, or unclear lead attribution? If you cannot name the problem, you are not ready to buy a solution.

Second, identify the evidence type. Are you looking at crawling, indexing, impressions, clicks or enquiries? If the evidence does not match the claim, stop there.

Third, check whether the issue affects pages that matter. A technical fault on archived blog tags is different from one affecting your main service pages. Not every issue deserves urgent spend.

Fourth, ask what change is actually proposed. Which URLs, templates or profiles will be updated? What will be added, removed or consolidated? “Optimisation” is not a plan.

Fifth, ask what result would count as progress, and what would remain uncertain. Better indexing? Higher UK impressions on a defined page set? Improved click-through from rewritten titles? More tracked calls from local landing pages? The answer should be specific.

Sixth, decide the next step.

If you mainly need understanding, keep learning before you buy. Our SEO and AI search topic guides are the right next step if you are still separating the basics.

If the evidence points to one clear issue, run a bounded test. Change a defined set of pages, record the baseline, and review after a sensible period.

If the diagnosis is clear but implementation is the blocker, seek practical help from the right place, whether that is your in-house team, your developer, or a service provider elsewhere.

Judging SEO advice well is not about becoming cynical. It is about becoming precise. Once you separate access, indexing, visibility, clicks and business outcomes, weak claims become easier to spot, and useful advice becomes much easier to recognise.