Service page SEO is the work of making one page the clearest answer to one commercial query: the service named, the area served, the proof, the price and the next step, in an order a reader and a retrieval system can both follow. Most service pages fail before any of that, by competing with a sibling page.
What follows is an anatomy: seven blocks, what each one is for, and a free way to check it. Platform facts are stamped as of October 2026 and attributed to the document that states them, because one convention service pages still carry stopped being true this year.
| The block | What it is for | Free way to check it |
|---|---|---|
| 1. The URL and the query it owns | One page per commercial intent | A site: search for the phrase on your own domain |
| 2. The first 100 words | The answer a reader and an engine both lift | Read the top and look for the answer |
| 3. Proof a stranger can check | Replaces adjectives with verifiable fact | Ask what can be checked without calling you |
| 4. A price, a range or a floor | Answers the question the visit started with | Search your own page for a number |
| 5. The objection FAQ | Price, timeline, and what if it does not work | Read it aloud as the buyer's questions |
| 6. One call to action | One next step, repeated | Count the distinct actions on the page |
| 7. A title nothing else on the site uses | Stops two of your pages chasing one query | Search Console Performance, sorted by page |
Decide the page should exist: one query, one page
A service page earns its own URL when it owns a query no other page on the site owns. Run a site: search for the phrase you think it owns, plus your domain. If two of your pages come back, they are splitting one signal, and the fix is a redirect rather than more writing.
Google's AI-optimization guide, last updated July 2026, addresses the temptation to spin out a page for every variation of how people might search. Doing so primarily to manipulate rankings, the guide says, "violates Google's scaled content abuse spam policy". Its spam policies page names the older form of the same pattern: "Doorway abuse is when sites or pages are created to rank for specific, similar search queries."
The honest test is whether a reader searching one phrase and a reader searching the other want a different page. Heat pump repair and furnace replacement do. Heat pump repair and heat pump repair near me do not, and the second one should be a sentence on the first page.
The first 100 words are the page
Open with a 40 to 60 word answer that names the service, the area and what the reader gets, with the service as the grammatical subject of the first sentence. Surfer's analysis of more than 100,000 AI citation placements, collected on 2 June 2026, found 38% of citations came from a page's first 100 words.
In the same analysis, pages that confirmed and answered the query early were cited 45% of the time against 23%. Surfer states the data is correlational and sells a product built on the finding, so read it as a strong pattern rather than a proven mechanism. The practical version costs nothing: put the answer where the hero image usually goes.
Proof a stranger can check, not adjectives
Proof on a service page is anything a stranger can verify without calling you: a named project, a photograph of your own work, a review with the reviewer's name attached, a licence number, a named tool you run. Adjectives are not proof, and an answer engine has nothing to quote from them.
Google's guide draws the same line between commodity and non-commodity content, and its example is specific: it contrasts a piece like "7 Tips for First-Time Homebuyers" with one written out of what the author actually did. On a service page the equivalent is one job described properly, including the part that went wrong and what it cost to fix.
Publish a price, a range, or the thing that sets it
Publish a number. If a single price is impossible, publish a range with the two or three variables that move it, or publish the floor. A page that refuses the question sends the reader to a competitor who answered it, and gives a retrieval system nothing to lift when someone asks what this costs.
Our own version is a floor rather than a quote: retainers start at $3,000 a month, and the discovery call is where the scope that sets the real number gets worked out. A floor is honest and it filters the enquiries that were never going to fit. A missing number does neither.
The objection FAQ, now that it buys no rich result
Answer the three objections that stop the call: price, timeline, and what happens if it does not work. Put them in visible text. The rich result that used to justify the block is gone: Google's documentation log records that the FAQ rich result stopped appearing in Search on 7 May 2026.
The narrowing started earlier. Google's August 2023 announcement limited FAQ rich results to "well-known, authoritative government and health websites", and the 2026 change finished it; the feature's documentation was removed that June. Keeping the markup costs nothing, and Google's guidance is that unused structured data causes no problems for Search. The block now earns its place by answering a buyer in words on the page, which was always the better reason to write it.
One call to action, and a title nothing else on the site uses
Give the page one action and repeat it: call, book, or request a quote, not all three. Then check the title. Google's title-link documentation tells site owners to "Avoid repeated or boilerplate text in title elements", and a title duplicated across two service pages is the cannibalization problem showing up in the results.
That documentation also lists what Google draws the title link from: the title element, the main visual title on the page, heading elements, og:title, other prominent content, and the anchor text of links pointing at the page. Make the first of those distinct and the rest usually agree. Three competing buttons split a visitor's attention the way two pages split a query.
What a service page looks like when it runs as a system
A page is a one-time artifact; a system is what keeps it true. Mirastart builds service pages as server-rendered HTML, so the opening answer sits in the source before any script runs, keeps one page per query by design, and puts rankings, calls and bookings on one screen, so a regression shows up the week it happens.
The booking system behind this site's own discovery-call page calculates real availability and confirms automatically; the online booking and repair-status software Quick Auto NC runs its bays on is ours, and so is the Carolina Sky Painting website. A service page that ends in a form nobody answers is a page that ranks and loses the job anyway.
Sources
- Google Search Central, "Optimizing your website for generative AI features on Google Search" - Google's statement that creating separate content for every possible variation of how people might search, done primarily to manipulate rankings or generative AI responses, violates its scaled content abuse spam policy; the commodity versus non-commodity content contrast and its "7 Tips for First-Time Homebuyers" example; that unique, compelling and useful content influences presence more than the guide's other suggestions. Page footer reads Last updated 2026-07-10 UTC.
- Google Search Central, "Spam policies for Google web search" - The doorway abuse definition quoted here: pages created to rank for specific, similar search queries. Page footer reads Last updated 2026-08-28 UTC. The policies describe practices, not thresholds; Google does not publish how similar is too similar.
- Surfer, "Why Almost 40% of AI Citations Come from Your First 100 Words" - Over 100,000 AI citation placements across more than 10,000 AI Overviews responses, collected June 2, 2026: 38% of citations come from a page's first 100 words, and pages that confirm and answer the query early were cited 45% of the time against 23%. The 38% figure is Google AI Overviews only. Surfer states the data is correlational and sells a product built on the finding.
- Google Search Central, "Influencing title links in Google Search results" - Google's instruction to avoid repeated or boilerplate text in title elements and to give each page distinct, descriptive title text, and its list of what the title link is generated from: the title element, the main visual title, heading elements, og:title, other prominent content, other page text, anchor text and WebSite structured data. Read via search result summary on October 2, 2026; developers.google.com is blocked to direct fetches from this environment.
- Google Search Central Blog, "Changes to HowTo and FAQ rich results" - The August 2023 announcement limiting FAQ rich results to well-known, authoritative government and health websites, and stating that for all other sites the rich result would no longer be shown regularly. Read via search result summary on October 2, 2026.
- Google Search Central, "What's new in Google Search documentation" - Google's documentation change log: the deprecation notice added to the FAQ rich result documentation because the feature no longer appears in Google Search as of May 7, 2026, and the removal of that documentation in June 2026; also the note that structured data which is not being used does not cause problems for Search. Read via search result summary on October 2, 2026.