Table of contents

SEO (search engine optimisation) is the practice of getting a site found in search results. The technical part sits underneath the writing and the links. If Google cannot fetch a page, nothing on that page matters, and if Google cannot tell what your business is, it has little to match against a customer's search. This guide explains each layer in plain terms, gives you an 18-line checklist with a free way to run each check, and shows what we found when we audited our own site.
What is technical SEO, and how is it different from on-page SEO?
Technical SEO is everything that decides whether a search engine can find, fetch, read and understand a page. On-page SEO is what the page says and how it is written. Off-page SEO is what other sites say about you.
The three overlap, but the question each one answers is different.
| Technical SEO | On-page SEO | Off-page SEO | |
|---|---|---|---|
| The question | Can Google reach and understand this page? | Does this page answer what the searcher wants? | Do other sites vouch for this business? |
| Examples | robots.txt, sitemap, HTTPS, page speed, schema markup, canonical URLs | Titles, headings, copy, images with alt text, internal links | Links from other sites, directory listings, reviews, mentions |
| Who usually fixes it | A developer, or a person comfortable in the site settings | The writer or business owner | The marketer or agency |
| When it shows up | Pages missing from Google, wrong details in results | Pages indexed but ranking poorly | Competitors with the same pages outranking you |
So technical SEO is about access and meaning, and on-page SEO is about content. A site needs both: a well-written page that Google cannot index earns nothing, and a perfectly configured site with thin pages has nothing worth indexing.
Google's own SEO starter guide draws the same line. It treats discovery, crawling and canonical URLs as groundwork, and says that creating "content that people find compelling and useful will likely influence your website's presence in search results more than any of the other suggestions in this guide". Technical SEO makes that content eligible to be seen.
Technical SEO now also covers how a site describes itself to machines: structured data (a standard way of labelling what is on a page), entities (the real-world things a page is about) and the AI answer tools that read the same signals. The middle sections spend the most time there, because small sites most often have gaps there.
Does a small business website need technical SEO?
Yes, but a small site needs far less of it than the guides written for large ecommerce sites suggest. Most of the work is a one-off check of about ten items, then a repeat after any big change.
A 20-page site rarely has a crawl budget or international targeting problem. Those topics fill enterprise guides and can be skipped. What goes wrong on a small site is more basic:
- A page is set to "noindex", or blocked in robots.txt, by a developer during the build and never switched back.
- The site has two versions of the same page, such as the address with and without "www", and Google has to guess which is the real one.
- The business details in the site's markup disagree with the Google Business Profile.
- Pages exist but nothing links to them, so Google has no route to them.
Each can stop a page appearing in search, and none is visible in a browser. If a page you expect to rank is missing from Google, start with why your business is not showing on Google, then come back here for the technical layer.
For a one-page site, see SEO for a single-page website. For a general health check on any site, how to check your website is working is the gentler starting point.
Google says a sitemap may be unnecessary if a site is "small", which it defines as "about 500 pages or fewer", and "comprehensively linked internally". Most local business sites fit that description. The same page adds that "in most cases, your site will benefit from having a sitemap", so we build one anyway.
Can Google find and index your pages?
A page can appear in Google only if Google can crawl it (fetch it) and index it (store it for retrieval). Google's technical requirements list three conditions: Googlebot is not blocked, the page returns an HTTP 200 (success) status code, and the page has indexable content. It adds that "indexing isn't guaranteed" even when all three are met.
Googlebot is not blocked. The usual cause is a robots.txt file, a plain text file at the root of your site that tells crawlers which URLs they may fetch. Google's robots.txt guide states that the file is "used mainly to avoid overloading your site with requests; it is not a mechanism for keeping a web page out of Google". A page blocked in robots.txt can still show up as a bare URL with no description. To keep a page out of results, use noindex, and let Google crawl the page so it can see the instruction. Mixing these two up is an easy way to leave a page out of Google by accident. Our robots.txt checker reads your file and also shows whether AI crawlers such as GPTBot are blocked.
The page works. Google indexes pages served with an HTTP 200 status. Error pages are not indexed. The URL Inspection tool in Search Console shows the status for any page, which catches a "page not found" message served with a 200 status.
The page has indexable content. Text in a supported file type, that does not break Google's spam policies. A page built mostly from images, or one that loads its words only after a script that fails, has little to index.
An XML sitemap lists the pages you want indexed, and Google says it "tells search engines which pages and files you think are important". It does not guarantee crawling or indexing. Our sitemap checker confirms yours loads and lists the right URLs. Canonical URLs deal with duplicates: Google chooses "a single URL (the canonical URL) to show users, per piece of content", and the starter guide suggests redirecting duplicates to the preferred address, or using rel="canonical" where a redirect is impossible.
Reading the indexing statuses in Search Console
Search Console's Page Indexing report sorts your URLs into statuses, and two cause the most worry. "Discovered, currently not indexed" means Google knows the URL exists but has not fetched it yet. "Crawled, currently not indexed" means Google fetched the page and has not added it to the index. Neither is an error message.
Start by linking the page from other pages on your site, since Google lists "making your content easily findable through internal links" among the fundamentals. Check that the page says something the rest of your site does not. Then request indexing through the URL Inspection tool and wait, because Google says indexing is never guaranteed.
Speed, mobile and security: what to check
Page speed, a working mobile layout and HTTPS are the page-experience basics.
Google groups speed into three measures called Core Web Vitals. Largest Contentful Paint (LCP) measures loading, and Google says to "strive to have LCP occur within the first 2.5 seconds". Interaction to Next Paint (INP) measures responsiveness, with a target "of less than 200 milliseconds". Cumulative Layout Shift (CLS) measures visual stability, aiming for "less than 0.1". Google says it "highly" recommends that site owners achieve good Core Web Vitals for success with Search, and that this aligns with what its core ranking systems seek to reward.
For a small business site the practical advice is short. Check the Core Web Vitals report in Search Console, which uses real visitor data and shows which pages fail. Fix the biggest cause first, such as oversized images, a slow host or heavy scripts.
View your own pages on a phone and tap every button, since that is how many visitors will use them. Serve the whole site over HTTPS, so the browser shows a padlock and no page triggers a warning, and make sure every old HTTP address redirects to the HTTPS one. Our free SEO checker flags many of these basics in one pass, and how to check your website is working covers what to look at by hand.
Schema markup: labelling your pages for machines
Schema markup is code added to a page that labels what the page describes, in a vocabulary search engines share. Google calls the result structured data, "a standardized format for providing information about a page and classifying the page content". It gives search engines "explicit clues about the meaning of a page" in place of leaving them to infer it from prose.
The shared vocabulary is schema.org. The format Google recommends is JSON-LD, a block of structured text in the page's code that is invisible to visitors. Google's introduction to structured data lists JSON-LD, Microdata and RDFa as supported and says JSON-LD is "Recommended". The block sits in a <script type="application/ld+json"> tag, so it can be added or edited without touching the page design.
A plain example makes it concrete. This is a made-up business, shown for illustration only:
{
"@context": "https://schema.org",
"@type": "Plumber",
"name": "Hartley Plumbing",
"telephone": "+44 20 7946 0000",
"address": {
"@type": "PostalAddress",
"addressLocality": "Bromley",
"addressCountry": "GB"
}
}
A person sees a plumber in Bromley. The block says the same thing in a form a program does not have to guess at.
Google's structured data policies say that structured data "enables a feature to be present, it does not guarantee that it will be present", and that missing required properties make a page ineligible for rich results (the enhanced listings with extra details). Schema helps Google understand the page and can make it eligible for those features. It is not a ranking switch. Google also requires that you do not "add structured data about information that is not visible to the user, even if the information is accurate".
For a local business, the types that matter are Organization for who you are, LocalBusiness or one of its more specific subtypes for what you do and where, and Service for what you offer. Google's local business guidance says to "use the most specific LocalBusiness sub-type possible". Our schema generator builds all four from a short form.
What is schema markup explains the format line by line, and local business schema gives the template to adapt for a UK trade or service firm.
FAQ schema, in short
FAQPage markup labels question-and-answer pairs on a page, and it has to match text a visitor can see. Our own home page carries one. Write the answers for people first and mark them up second. Our FAQ schema generator produces the code, though no special display in Google results is guaranteed.
Testing your markup
Check markup in three places. Google's Rich Results Test, which its structured data introduction calls "an easy and useful tool for validating your structured data", shows whether a page is eligible for rich results. The URL Inspection tool shows what Google saw on its last visit. Our schema validator finds the JSON-LD on any page and flags invalid markup. Run them after every template change, since a theme update can silently break the block.
Entities: getting Google to know your business is a thing
An entity is a specific, identifiable thing: a person, a place, a company, a product. Entity SEO is the practice of making it unambiguous to search engines which thing your site is about, and linking that thing to what is already known about it elsewhere.
Google described the shift in 2012, when it introduced the Knowledge Graph under the line "things, not strings". Its announcement asked whether a search for Taj Mahal meant "Taj Mahal the monument, or Taj Mahal the musician", and answered: "Now Google understands the difference." The Knowledge Graph is the database of such things and the relationships between them. A string is just letters. A thing has a type, attributes and links to other things.
For a small business the problem is practical. Your business name may be shared with other firms, or with a common word. Without help, a search engine or AI tool has to guess which one your website describes. Three techniques reduce the guessing.
One identity per thing. Give your business a single stable identifier in your markup, called an @id, usually a URL with a fragment such as https://example.co.uk/#organization. Every page that mentions the business points to that same @id instead of describing a new organisation each time. When two pages give the same business two different names, or leave the business anonymous, the graph splits into several half-things.
Links to places that already know you. The sameAs property lists other pages about the same entity: your Companies House record, your Google Business Profile, your LinkedIn page. Google's organisation markup guidance describes it as "the URL of a page on another website with additional information about your organization", and notes that some properties are used "behind the scenes to disambiguate your organization from other organizations". A knowledge base entry such as a Wikidata record gives your entity a stable reference.
Consistent facts everywhere. The name, address, phone number and service descriptions in your markup, on your pages, on your Google Business Profile and in your directory listings should agree. Local business listings and consistent details covers the listings side.
Google's knowledge panels, the information boxes shown for entities in the Knowledge Graph, are "automatically generated", and Google's help page says that businesses with a physical location or service area use a Business Profile instead. You can make your entity easy to recognise. How to get a Google Knowledge Panel covers what that takes and what Google says about claiming one, and our Google Knowledge Panel checker shows whether Google already recognises your business or name as an entity. Entity SEO goes deeper on the whole method, and how to set up your Google Business Profile covers the profile that often serves as the main anchor for a local firm.
Semantic SEO: writing for meaning
Semantic SEO is the practice of writing pages around topics and the relationships between things, in place of repeating a target phrase. A page about boiler servicing should name the boiler types, the area served, the qualifications and the price factors, because those are the things a person asking about boiler servicing wants resolved.
In practice the approach has three habits.
Cover the topic, then link the parts. A topic is a set of questions people ask. A strong page answers the main one first, then the nearest related ones, and links to deeper pages for the rest. This guide does that on purpose.
Name the things the page discusses. A page that mentions "Bromley", "Gas Safe registration" and "combi boiler" is built from recognisable entities. Use each thing's proper name plainly.
State facts as complete sentences. The smallest unit of meaning is a statement with a subject, a predicate and an object, known as a semantic triple. "Hartley Plumbing serves Bromley" is one triple: the subject is Hartley Plumbing, the predicate is "serves", the object is Bromley. Schema markup writes the same triple as data (same made-up example business):
{
"@type": "Plumber",
"@id": "https://hartleyplumbing.example/#business",
"name": "Hartley Plumbing",
"areaServed": { "@type": "City", "name": "Bromley" }
}
The @id is the subject, the property areaServed is the predicate, and the City is the object. The point is that every sentence of fact on your site has a matching line in your markup, and the two should agree. Prose and markup that state different things are worse than no markup.
Nothing in Google's documentation promises that writing in triples or naming more entities raises a ranking. The value is clarity: pages that state complete facts are easier for people to scan, for a search engine to match to a question, and for an AI tool to quote accurately.
Technical SEO for AI search
AI search tools read the same web that Google does, so the technical groundwork that makes a site findable in Google is also the groundwork for AI answers. Google says so directly in its guide to AI features: "The best practices for SEO remain relevant for AI features in Google Search", and "There are no additional requirements to appear in AI Overviews or AI Mode, nor other special optimizations necessary".
The same page sets the technical bar: to appear as a supporting link, "a page must be indexed and eligible to be shown in Google Search with a snippet". It lists the basics worth keeping in order, including allowing crawling in robots.txt "and by any CDN or hosting infrastructure", linking your content internally, keeping important content in text, making sure "structured data matches the visible text on the page", and keeping Business Profile information up to date.
On schema, Google is plain: "There's also no special schema.org structured data that you need to add." So the honest position is that markup is not a lever for Google's AI features. It is still worth having, because it removes ambiguity about who you are and what you do, and AI tools that retrieve facts about a business benefit from facts that are consistent and labelled. For other AI tools we cannot point to a published statement from their makers about reading schema, so we do not claim one.
Two practical checks apply. First, make sure you have not blocked the AI crawlers you want. Our robots.txt checker shows whether bots such as GPTBot are disallowed. Second, ask the tools. Our AI visibility checker asks AI tools questions about your trade and area and shows which businesses they name, which is the only direct way to see how you appear.
A word on llms.txt. It is a plain text file at the root of a site that lists the pages a site owner wants AI tools to read. Google's AI features guide says you "don't need to create new machine readable files, AI text files, or markup" to appear in its features. We treat llms.txt as optional and cheap, and our llms.txt generator makes one in a minute. Add it after the basics are in place.
For the wider picture, see what is AEO, AEO vs SEO and how to get recommended by AI search.
Worked example: what we found auditing our own site
We ran an entity audit of seopages.co.uk on 20 October 2026 using our nine-check method, found real faults, and fixed every finding the same day. We then re-checked all 199 pages in the sitemap.
The nine checks cover these questions:
- Entity identity: is there one canonical node for the business?
- Graph connectivity: does one thing carry one
@id? - Knowledge-base anchoring: does
sameAspoint to the right references? - Aliases: are alternative names declared in markup, not stuffed into copy?
- Attribute coverage: does the markup answer what people ask about the thing?
- Subject entities: do content pages declare what they discuss, and only that?
- Consistency: is there one name everywhere?
- Topical focus: does each page stay on its subject?
- Honesty gates: is every entity claim true?
The findings were specific, and they were our own mistakes:
- A shared fallback had labelled 25 local area pages as "Programmatic SEO", a service we do not sell under that name.
- One
@idcarried two different names for the same service. - Our blog posts declare Google Business Profile as a subject, and its anchor pointed at a Wikipedia title that redirects into the Google Maps article. That would have merged two different things, so we anchored it to the product's Wikidata item (Q21235252) instead.
- The free audit page described itself as covering paid ads, which it does not.
An earlier audit on 25 August 2026 had found a thin entity home, provider stubs that created a fresh anonymous organisation on each page, and served places with no anchor to say which town was meant.
Here is a trimmed version of the live markup on our home page after the fixes. It shows the structure: one Organization node, a stable @id, an official identifier, a founder with their own @id, and a service with its own @id that points back at the same organisation.
{
"@context": "https://schema.org",
"@type": "Organization",
"@id": "https://www.seopages.co.uk/#organization",
"name": "SEO Pages",
"legalName": "SEO Pages Ltd",
"identifier": {
"@type": "PropertyValue",
"propertyID": "GB-COH",
"value": "14129018"
},
"founder": {
"@type": "Person",
"@id": "https://www.richardthomaspryce.com/#person",
"name": "Richard Thomas-Pryce"
},
"sameAs": [
"https://www.linkedin.com/company/seo-pages",
"https://www.google.com/search?kgmid=/g/11zggns48z"
],
"disambiguatingDescription": "SEO Pages is a UK digital marketing agency (SEO Pages Ltd, Companies House 14129018), not the generic term for search-optimised web pages.",
"makesOffer": [{
"@type": "Offer",
"itemOffered": {
"@type": "Service",
"@id": "https://www.seopages.co.uk/services/ai-visibility#service",
"provider": { "@id": "https://www.seopages.co.uk/#organization" }
}
}]
}
None of these faults showed in a browser. Every page looked right. They appeared only when we read the graph the way a machine reads it, and a search engine or AI tool resolves a name collision by guessing.
A technical SEO checklist for small business sites
This is the working list, 18 checks in the order we would run them on a site of 5 to 100 pages. Each has a free way to run it. Google Search Console is free and worth setting up before anything else.
| Check | Why it matters | Free way to check it |
|---|---|---|
| Key pages return a 200 status | Google indexes only pages served with HTTP 200 | URL Inspection tool in Google Search Console |
| robots.txt does not block pages you want found | Blocked pages are unlikely to appear in results | robots.txt checker |
No important page carries noindex | noindex keeps a page out of Google | URL Inspection tool, or free SEO checker |
| robots.txt does not block AI crawlers you want | Blocked crawlers cannot read your pages | robots.txt checker |
| XML sitemap loads and lists the right URLs | Helps Google discover pages, though it guarantees nothing | sitemap checker |
| Sitemap submitted in Search Console | Gives Google a direct list and a status report | Sitemaps report in Search Console |
| Every important page is linked from another page | Google finds pages by following links | Click through from the home page, or Search Console links report |
| One URL per page, with redirects or canonicals for duplicates | Stops Google guessing which version to show | URL Inspection tool, "User-declared canonical" field |
| Whole site served over HTTPS, HTTP redirects to it | A secure padlock and no warnings for visitors | Browser address bar, and type the HTTP address to test the redirect |
| Core Web Vitals pass on the main pages | LCP under 2.5 seconds, INP under 200 ms, CLS under 0.1 | Core Web Vitals report in Search Console |
| Pages work and read well on a phone | Visitors use the site on phones | Open the site on a phone and tap every button |
| Page Indexing report reviewed monthly | Shows pages Google chose not to index, and why | Page Indexing report in Search Console |
| Each page has a unique title and a clear heading | Titles influence the title link shown in results | Free SEO checker |
| Images have descriptive alt text | Explains images to search engines and screen readers | Free SEO checker |
JSON-LD present: one Organization or LocalBusiness node with a stable @id | Labels who you are for machines | Schema validator |
| Markup matches visible text and your Business Profile | Google requires markup to reflect what visitors see | Schema validator, compared by eye with the page |
sameAs links point to the correct profiles | Ties your entity to places that already know it | Google Knowledge Panel checker |
| AI tools name you correctly for your trade and area | Shows what a customer sees in an AI answer | AI visibility checker |
Work down the list in order. The first nine decide whether pages can be found at all, so fix any failure there before touching markup. Repeat it after a redesign, platform change or hosting move, when noindex tags and redirect rules tend to go wrong. The full set of free tools sits at /tools.
FAQ
What does a technical SEO specialist do?
A technical SEO specialist checks whether search engines can crawl, index and understand a site, then fixes what blocks them. The work covers robots.txt and sitemap settings, redirects and canonical URLs, page speed, structured data and the Search Console reports that show indexing problems.
Does schema markup help SEO?
Schema markup helps Google understand a page and can make the page eligible for rich results, which are enhanced listings with extra details. Google does not guarantee that rich results will appear, and its documentation does not describe structured data as a ranking factor. Treat it as clear labelling, and make sure it matches the visible page.
Does AI use schema markup?
Google says there is no special schema.org structured data you need to add to appear in its AI Overviews or AI Mode. It does ask that any structured data you use matches the visible text on the page. For other AI tools, we are not aware of a published statement either way, so we use markup to remove ambiguity and not as a promised route into AI answers.
How long does technical SEO take?
Finding the problems on a small site takes hours to a few days. We fixed every finding from our own entity audit the same day, then re-checked all 199 pages in the sitemap. How soon Google reflects a fix depends on when it next crawls the pages, and Google states that indexing is never guaranteed, so allow time and check Search Console for progress.
Where to go from here
Start with the first nine checks, then work through markup and entities in the order the sections above give them. If you would rather have the entity and AI-answer side handled for you, see how our AI visibility service works. For the monthly local search work that sits alongside it, see our SEO packages.


