SEO analysis for your website.

An SEO analysis works through five layers, and the order is not arbitrary: without indexing there is no position, without a position there are no clicks, and without the right landing page there is no enquiry.

• Indexing. Which URLs Google knows, which are included, which are excluded deliberately or by accident. The Search Console indexing report separates these cleanly: "crawled, currently not indexed", "discovered, currently not indexed", redirects, duplicates and active exclusions. Each of those categories has a different cause and a different treatment. We have written up the reasons behind the most common one in a separate article on "crawled, currently not indexed".
• Visibility and positions. Which search queries the website actually appears for, with how many impressions, at what average position, at what click rate. What matters here is the distribution, not the mean. A domain with tens of thousands of impressions beyond position 20 does not have a click-rate problem, it has a position problem — and the two diagnoses lead to completely different actions.
• Matching search query to page. Which URL ranks for which term, and whether that is the URL that should be ranking. This is where it shows whether several of your own pages compete for the same term — keyword cannibalisation, one of the few SEO effects that can be fixed reliably by deleting and merging.
• Keyword gaps. Which terms comparable providers serve that you have no matching page for. Not every gap is an opportunity. Only the combination of volume, competitive difficulty, buying intent and the question of whether you can say anything evidenced on the topic turns a gap into a task.
• Technical obstacles with a search bearing. Redirect chains, canonical contradictions, blocking entries in robots.txt, sitemaps containing dead or excluded URLs, content that only appears via JavaScript, and internal linking that buries important pages unreachably deep.

Everything beyond that — load time optimisation in detail, accessibility, imprint and privacy policy, forms, abandoned purchases — does not belong in the SEO analysis. It belongs in the website audit. That separation is not a formality. It stops an analysis ending with thirty findings, twenty-six of which have nothing to do with the question of why nobody is finding the site.

Six checks cover the most common causes, cost nothing and take about thirty minutes together. The precondition is verified access to Google Search Console for your domain. If that access is missing, that is already the first finding, and everything else waits until it exists.

• 1. Open the indexing report. Compare the number of indexed pages with the number of pages that should be indexed. If they differ sharply, first establish the direction. Too few means a blockage. Too many means filter, search or parameter addresses have landed in the index.
• 2. Set the performance report to twelve months. Sort search queries by impressions and look at the position column. Everything between position 4 and 20 with meaningful impressions is your shortest route to more clicks. That does not need a new page, it needs a better one.
• 3. Hold the pages report against the queries report. Filter one important search query and look at which URL is being served for it. If it is not the page you expected, you have found a matching problem — the most common undetected finding there is.
• 4. Open robots.txt and the sitemap. Both are one click away in a browser. Check whether whole directories are blocked by accident, and whether the sitemap actually contains the pages that should rank — and only those.
• 5. Look at the field values for load time. Alongside the lab measurements, PageSpeed Insights shows the real user data from a rolling 28-day window. The benchmarks are 2.5 seconds for Largest Contentful Paint, 200 milliseconds for Interaction to Next Paint and 0.1 for Cumulative Layout Shift, each at the 75th percentile.
• 6. Type your most important search query in yourself. In a private window, with a Swiss location. Look at what kind of page the first five results are: provider pages, comparison lists, directories, forums, an AI summary above them. If your page type does not appear there, that is not an optimisation problem. It is a format problem.

Anyone who works through those six points has covered a considerable share of what many paid "SEO analyses" deliver as their result. That is exactly why the instructions are here and not behind a form.

Free tools show states, not connections. The value of a professional analysis lies almost entirely in linking data sources that look unremarkable on their own. Five kinds of finding turn up regularly, and none of them appears in an automatically generated report.

• Self-cannibalisation across several URLs. When four of your own pages serve the same term, every signal splits four ways and none gets strong enough. That only becomes visible when the Search Console query report is broken down by URL over months and the positions are laid side by side: same query, changing landing page, all of them mid-table. A review tool reports nothing, because formally each individual page is fine.
• The wrong search intent. A page can be technically flawless and still not rank, because it answers the wrong question. Someone searching for a price wants a number, not a service description. Someone searching for a comparison wants a comparison, not a self-portrait. That can only be judged by reading the actual results page — no tool does it for you. What the term means is explained in our glossary entry on search intent.
• Index bloat. Filter combinations, sort parameters, internal search results and print views quickly generate several thousand reachable addresses on a mid-sized website. They do no harm by existing. They do harm by binding the crawler's attention and diluting signals. The finding only emerges from comparing the crawl result, the sitemap and the index status.
• Losses after an intervention. After a relaunch, a domain change or a rebuild of the navigation, visibility rarely drops at a stroke. It drains over weeks. That is only recognisable by comparing equally long periods before and after the intervention, broken down by URL group — and it is then usually traceable to a manageable number of missing or wrongly set redirects.
• Competitor comparison at URL level. Not "who is stronger", but: which specific page does the competitor hold this term with, how long is it, which additional questions does it answer, which page type does it serve. Out of that comes a requirement for your own page, which you can then meet or deliberately reject.

What stands out is how often these findings lead not to work but to the opposite. The most common honest recommendation out of an analysis is to merge three mediocre pages into one good one and redirect the other two.

Four figures hold up: impressions, clicks, the distribution of positions, and the number of search queries a website appears for at all. Everything else is either derived or interpretation.

The best-known trap is average position. It averages over every query and every day and hides exactly the distribution that matters. A domain can show a passable average position and still receive practically no clicks, because the impressions gather beyond position 20, where structurally nobody clicks any more. What is meaningful is therefore the split: how many queries sit in the ranges 1 to 3, 4 to 10, 11 to 20 and beyond — and how that split shifts across quarters.

The second trap is visibility indices from third-party tools. They are useful for the trend of one domain and for comparisons inside the same tool, but they rest on a fixed basket of keywords that the vendor defined. For a Swiss SME with a narrow, specialist topic, that basket often contains barely any relevant terms. The index then hardly moves while the visibility that actually matters commercially rises or falls sharply. There is a simple rule for apparent contradictions between tools: numbers from Search Console are your own measurements, everything else is an estimate about your business.

The third point concerns the click rate itself. Where an AI summary sits above the organic results, the click rate falls even at an unchanged position — the answer already appears on the results page. The same content can gain effect and lose clicks in that situation. Anyone looking only at the click count reads it as a decline. How mentions in such answers can be measured at all is described on our page on Generative Engine Optimisation.

A keyword gap is worth filling when four conditions meet: evidenced search volume, reachable competitive difficulty, recognisable buying intent, and the ability to say something evidenced on the topic. If one of the four is missing, you create a page that nobody searches for, nobody finds or nobody is helped by.

On volume, caution is warranted in both directions. Tools report sometimes very different numbers for the same term, because they use different country databases, different aggregations and in part broad-match sums. Annual averages also contain individual months with conspicuous spikes that look more like a data artefact than real demand. So we work with the average of the last three months as a matter of principle, and we name the source alongside the number.

On competitive difficulty, the Swiss market is often friendlier than the volumes suggest. It happens regularly that a term with roughly fifty to a hundred searches a month carries a very low competition index and at the same time a double-digit cost per click — little competition at high commercial value. Terms like that are considerably more valuable to a small provider than the big head term that twenty agencies with four-figure monthly budgets are fighting over.

The fourth condition is the most uncomfortable. A gap you can contribute no experience, no reference and no numbers to gets filled with generic text — and generic text barely ranks in an environment where AI systems answer directly. In that case we advise against it. What follows from all this for choosing an agency is set out on our page on choosing an SEO agency.

Three peculiarities change how Swiss data reads: small absolute numbers, multilingualism, and a high share of German results inside the same search queries.

First, the order of magnitude. A term with a hundred searches a month is not a niche phenomenon in Switzerland. For many services it is already a head term. That changes the statistics: at those volumes the figures swing relatively hard, individual weeks say nothing, and the decision for or against a page cannot be justified with significance, only with plausibility plus commercial value per enquiry. Anyone deriving a trend from twelve clicks a month is reading tea leaves in a spreadsheet.

Second, the languages. The same service is named differently and searched for at different volumes in German, French and Italian, and the results pages differ completely. An analysis that looks only at German-speaking Switzerland is incomplete for a company doing business in the Romandie. Conversely, a French version is not automatically worth it. It doubles the upkeep permanently and only pays off where demand is evidenced.

Third, competition out of Germany. For many specialist queries, German portals with considerably longer domain histories sit in the Swiss results. What helps against that is less text volume than Swiss specificity: amounts in francs, Swiss legal terms, cantonal differences, Swiss Standard German with "ss" instead of "ß", a Swiss address and phone number. For search engines as for AI systems, those are easily checkable signals that a source really does write for this market.

An analysis is the wrong spend when its result is foreseeable, or when nobody will act on it. Four situations come up regularly.

• The website is new. Under three months, the history an analysis draws its statements from does not exist. What makes sense at that stage is a basic technical review and keyword planning, not a reading of data that is not there yet.
• There is no capacity for content. Most findings from an SEO analysis lead to editorial work: merging pages, rewriting them, adding to them. If nobody in the company can own copy and no budget is set aside for it, the list will sit untouched.
• The business does not run on search. Companies whose customers come through passing trade, referral, tender or field sales gain little from better rankings. A well-kept Google business profile and a reachable website are enough then. The rest of the budget is better spent elsewhere.
• The problem is the website itself. If a page is unusable on a phone or the contact form has delivered nothing for months, a better position changes nothing about the outcome. Then the website audit is the right order of operations, or a rebuild outright.

If you are unsure which of the two applies to you, the free SEO check exists for exactly that. It costs nothing, takes minutes, and in many cases ends with the recommendation not to spend money. What a later collaboration would cost is set out openly under SEO costs and ongoing SEO.