Small Business SEO Checklist: 21 Items, 6 Retired

Four of the pages ranking for "small business seo checklist" today are 23-item lists with no way to tell whether you finished an item. They number title tags as item 1 and the Google Business Profile as item 12, which is backwards for a plumber and irrelevant for a remote bookkeeper. Three of the four tell you to add FAQ schema for a rich result that Google stopped showing on May 7, 2026.

This list is ordered, it tells you which half applies to you, and every item ends with something you can run or open to see whether it passed. Six items that appear on the competing lists are in a separate section, because they are retired or were never requirements.

Two honest notes before the list. No one can guarantee a ranking, a position or a date, and any checklist that implies otherwise is selling something. And the items are ordered by what blocks everything downstream, not by difficulty: a page Google cannot index does not care how good its title tag is.

Start with the two gates that decide which half of this list applies

Before item 1, answer two questions. They remove between a third and half of the usual checklist for most readers, which is the point.

Gate 1: can Google fetch and index this page right now? Everything below item 7 assumes yes. If the answer is no, the rest is decoration. A page that is blocked, redirected or marked noindex is invisible no matter how many items you tick, and this is the single most common reason a site does not show up on Google at all.

Gate 2: are you eligible for a Google Business Profile? Google's eligibility page is specific. A business qualifies if it "has a physical location that customers can visit" or "travels to customers where they are." If neither is true, the five local items are not a to-do list for you, they are noise.

A remote bookkeeper, a voice-over artist, an online course business: none of them meet either condition. Google's guidelines also rule out a rented mailing address you do not work from, say that "P.O. boxes or mailboxes located at remote locations aren't acceptable," and allow a coworking office only if it "maintains clear signage," is staffed by your own people, and receives customers during stated hours.

If you travel to customers without a storefront, you are a service-area business and Google says you are "allowed one service-area Business Profile." One. Not one per city.

Six items on the competing checklists that are retired or were never rules

Each of these appears on at least one page currently ranking for this keyword. Each was checked against the primary source in November 2026.

1. FAQPage schema for a rich result. Gone. Google's documentation records that the feature "will no longer appear in Google Search starting May 7, 2026" and that on June 15, 2026 the FAQ documentation itself was removed. The Search Console API drops FAQ search-appearance data in August 2026. The structured data gallery now lists 25 features, and FAQ is not one of them. Writing an FAQ section for readers is still worth doing. Marking it up for a rich result buys nothing in Google.

2. "Exactly one H1 per page." Not a Google rule. The SEO starter guide says that from Search's perspective "it doesn't matter if you're using them out of order" and that "there's also no magical, ideal amount of headings a given page should have." Use headings so a reader can skim. Stop auditing their count.

3. "Submit your sitemap or Google will never find you." Google's own starter guide says "the vast majority of sites" are found automatically as Google crawls the web, and that you "usually don't need to do anything except publish your site on the web." It files sitemap submission as optional and a "little technical challenge." Submit one anyway, because it takes four minutes and gives you a coverage report, but it is not the gate it is sold as.

4. "Block the page in robots.txt to keep it out of Google." The robots.txt documentation is blunt: it "is not a mechanism for keeping a web page out of Google," and a blocked URL "can still appear in search results." The actual mechanism is a noindex meta tag or response header, a password, or deleting the page. Robots.txt manages crawler traffic. These are different jobs and confusing them is how private pages end up indexed with no snippet.

5. "Add Service or ProfessionalService schema for a rich result." There is no such feature. The gallery's 25 entries include Local business and Organization and nothing for Service, Offer or professional services. schema.org/ProfessionalService resolves, but its own page carries the notice that the type "was deprecated due to confusion with Service." Markup for a feature that does not exist is not penalised, it is just unpaid work.

6. "Fix your First Input Delay." FID was retired. Interaction to Next Paint became a stable Core Web Vital in 2024. The current targets, measured at the 75th percentile and split by mobile and desktop, are LCP within 2.5 seconds, INP at 200 milliseconds or less, and CLS at 0.1 or less. A checklist still naming FID has not been touched in two years, whatever its "updated for 2026" line says.

Phase 1, items 3 to 7: prove Google can reach and index the page

This phase is five commands. Run them against your homepage and your two most important pages, not against your whole site.

3. The page returns 200 to a crawler. Not to your browser, which has your cookies and your session.

curl -s -o /dev/null -w "%{http_code} %{redirect_url}\n" \
  -A "Mozilla/5.0 (compatible; Googlebot/2.1; +http://www.google.com/bot.html)" \
  https://example.com/services/

A 200 with an empty redirect field passes. A 301 to a URL you did not expect, or a 403, is the item. Check the www and non-www variants, and http as well as https: a redirect chain that bounces four times is still a fail.

4. robots.txt does not block what matters. Fetch it and read it. It is usually twenty lines.

curl -s https://example.com/robots.txt

Look for a Disallow: that catches a directory you want indexed. A single stray Disallow: / left over from a staging deploy is the most expensive line of text a small business site can carry.

5. No stray noindex. This is the one that hides in a plugin setting or a theme option.

curl -s https://example.com/services/ | grep -io '<meta[^>]*robots[^>]*>'
curl -sI https://example.com/services/ | grep -i x-robots-tag

Either command returning noindex is a fail. The header version is worse because it never appears in the page source, so a visual check misses it entirely.

6. One canonical per page, pointing at itself. Two canonicals, or one pointing at the homepage from every page, both tell Google to ignore the page you want ranked.

curl -s https://example.com/services/ | grep -io '<link[^>]*canonical[^>]*>'

7. Search Console connected, sitemap submitted, one URL inspected. Open URL Inspection and paste the page you just tested. "URL is on Google" passes. "Discovered, currently not indexed" means Google knows about the URL and has not fetched it, which is a signal about site quality and crawl budget, not a technical defect you can fix with a plugin.

Those five checks take about ten minutes and they are the whole reason the ecommerce version of this checklist pairs every item with a test. An item you cannot verify is an item you will tick wrongly.

Phase 2, items 8 to 12: aim each page at a query someone actually types

This phase is where small business SEO is usually lost, and it is not technical. The failure is writing a page for a phrase nobody searches, or writing five pages for one phrase.

8. Each page targets one query with a measured volume. Not a guessed volume. "Affordable plumbing solutions" reads like a heading and gets searched by nobody. The query your customers type is shorter, uglier and often has a place name in it.

9. Run the parent topic check before you write. A parent topic is the broader query a keyword tool resolves your phrase into. When two of your planned pages share one parent, Google is treating them as one result page and you are about to compete with yourself. This keyword is its own parent, which is why it is worth a page rather than a paragraph inside a wider article.

10. The title tag is written for the click. Keyword first because that is what is bold in the result, then the reason to pick you over the nine other blue links. Sixty characters is the practical ceiling before truncation.

11. The meta description is a snippet request, not a ranking factor. Google's guidance is that the snippet usually comes from the page's own content and "occasionally" from the meta description. Write it as "typically a succinct, one- or two-sentence summary," unique per page, and do not expect it to be used every time.

12. Find your own cannibalisation. Two of your pages ranking for one query is a defect you built. The fastest check is a Search Console query filter: pick a query you care about, look at the Pages tab, and see whether two URLs trade positions week to week. If they do, one page should absorb the other and redirect, which is standard work in any content audit.

Phase 3, items 13 to 17: the local half, only if Gate 2 said yes

Skip this entire section if you are remote. For everyone else it outranks Phase 2 in effect, which is exactly why putting the Business Profile at item 12 of 23 misleads readers.

13. Profile claimed, primary category correct. The primary category is the single highest-leverage field in the profile and most owners pick the one that flatters rather than the one that matches. "Dentist" and "Cosmetic dentist" surface for different searches.

14. Name, address and phone consistent, and the address hidden if you are service-area. Here is a contradiction worth naming: Google asks a service-area business to hide its address, while LocalBusiness structured data requires one. There is no clean resolution. Hide it in the profile, keep it in the markup, and do not mix a home address into a directory listing you cannot edit later.

15. LocalBusiness JSON-LD with its two required properties. The requirement is genuinely small. Google's documentation lists exactly two required properties, address and name. Everything else is recommended.

{
  "@context": "https://schema.org",
  "@type": "LocalBusiness",
  "name": "Example Plumbing",
  "address": {
    "@type": "PostalAddress",
    "streetAddress": "12 Mill Street",
    "addressLocality": "Reading",
    "postalCode": "RG1 2AB",
    "addressCountry": "GB"
  },
  "telephone": "+44-118-000-0000",
  "url": "https://example.com/",
  "openingHoursSpecification": [{
    "@type": "OpeningHoursSpecification",
    "dayOfWeek": ["Monday", "Tuesday", "Wednesday", "Thursday", "Friday"],
    "opens": "08:00",
    "closes": "17:30"
  }]
}

Note what is absent: no areaServed, because Google's LocalBusiness documentation never mentions it or service-area businesses at all. Adding properties the documentation does not ask for is harmless and does nothing.

16. Reviews, because prominence is a named ranking factor. Google's local ranking page is unusually direct about the inputs. The three headings are Relevance, Distance and Prominence, and prominence "reflects how well known a business is, including links from other websites and review count and ratings." That is the official sentence behind the advice to ask customers for Google reviews, and it is also why review count without recency fades.

17. Accept that you cannot buy the local pack. The same page states: "There's no way to request or pay for a better local ranking on Google." Anyone selling guaranteed map placement is selling something Google says does not exist. Distance is also a factor you cannot change, which is the honest ceiling on local work for a business far from its city centre. The rest of the levers live in ordinary profile and website optimisation.

Phase 4, items 18 to 20: the content that earns links and citations

18. Publish the pages the queries from Phase 2 demand. One page per query, answering the question in the first two sentences rather than after four paragraphs of context. This is the slow half of the work and the reason blogging still helps SEO when the posts target real queries and does nothing when they do not.

19. Link from the pages that already get impressions. Search Console will tell you which of your pages Google already shows. Those pages have the crawl attention, so an internal link from them to a new page is worth more than one from a page nobody visits. Three to five contextual links, with anchor text that describes the destination, is enough.

20. Make the answer extractable for assistants. A chunk that answers one question completely, with the entity named rather than referred to as "it," is what gets quoted when an assistant builds an answer. The technique is the same whether the extractor is an AI Overview or a chatbot, and it is the whole subject of answer engine optimisation. Treat it as a formatting discipline, not a separate channel.

Phase 5, item 21: one measurement that tells you whether any of it worked

21. Query Search Console yourself, with a row limit you chose. The interface caps and rounds. The API does not, and the default is the trap.

rowLimit accepts 1 to 25,000 and defaults to 1,000. Ask for more than a thousand rows without setting it and you get a thousand, silently, with no indication that the rest exists. startRow is zero-based for paging. type defaults to web. dataState defaults to final, so fresh partial data is excluded unless you ask for all. searchType is deprecated in favour of type, and seven dimensions are available: country, device, page, query, searchAppearance, date and hour.

curl -s -X POST \
  "https://www.googleapis.com/webmasters/v3/sites/https%3A%2F%2Fexample.com%2F/searchAnalytics/query" \
  -H "Authorization: Bearer $GSC_TOKEN" \
  -H "Content-Type: application/json" \
  -d '{
    "startDate": "2026-10-01",
    "endDate": "2026-10-31",
    "dimensions": ["query", "page"],
    "rowLimit": 25000,
    "startRow": 0,
    "dataState": "all"
  }'

What you are looking for is one pattern: queries with impressions climbing and clicks flat. That is a page Google is showing and nobody is choosing, which is a title and description problem, not a content problem, and it is fixable in ten minutes per page. The opposite pattern, impressions flat at zero after six weeks, usually sends you back to Phase 1.

Google's own guidance on timing is worth quoting before you panic at week three: "some changes might take effect in a few hours, others could take several months," and you "likely want to wait a few weeks to assess whether your work had beneficial effects." That is the official answer to how long SEO takes, and it is vaguer than any agency pitch.

What this checklist leaves out on purpose

Link building schemes. Guest post networks, paid placements and "competitor link stealing" appear on two of the pages ranking for this keyword, in both cases without mentioning that paid links violate Google's spam policies unless marked. For a business with no content team, the realistic sources are suppliers, local associations, sponsorships you already pay for, and the occasional page good enough to get cited.

Speed as a project. Core Web Vitals are a tiebreaker, not a lever. Hitting LCP under 2.5 seconds matters; a four-week rebuild to move a Lighthouse score from 88 to 96 does not. Get the thresholds, then stop.

Keyword density, alt text everywhere, and the long tail of audit-tool warnings. Most SEO audit tools generate a hundred findings because a hundred findings justify the subscription. Twenty of them are the items above.

Whether to do this yourself at all. The starter guide's own callout says that "if you're short on time or not feeling adventurous, you might consider hiring a professional." The honest version of the question is covered in the DIY breakdown, in what SEO actually costs, and in the comparison against paying for ads instead. Phase 1 is an afternoon. Phase 4 is permanent, and that is the part people stop doing in month three.

FAQ

How do I do SEO for a small business with no budget?

Run Phase 1 and Phase 3 first, because they are free and they are the ones that gate everything else. Five curl commands, a Search Console account, a claimed Business Profile with the right primary category, and a review request in your invoice email. That is a weekend. Phase 2 needs keyword data, and a free tier plus Search Console's own query report will get you far enough to stop guessing. The spending usually starts at Phase 4, because publishing pages consistently is the item that costs either time or money.

What is the 80/20 rule in SEO for a small business?

For a local business it is the Business Profile plus indexability: items 3 to 7 and 13 to 17, which is about half the list and most of the result. For a remote business it is Phase 2, aiming pages at queries with measured demand, because without a local pack to win there is nothing else carrying you. The 20 percent that produces almost nothing is consistent across both: schema beyond the required properties, audit-tool warnings, and speed work past the published thresholds.

Is an SEO checklist still relevant now that AI answers appear in search?

The technical half is more relevant, not less. An assistant cannot cite a page it cannot crawl, so items 3 to 7 are the entry fee for both. What changes is the content half: an answer written as one self-contained chunk with entities named explicitly travels into generated answers, while the same information spread across six paragraphs does not. Click-through from the pages that do get cited is lower than the old blue-link era, which is a real cost and worth weighing honestly before committing a year.

How many items should a small business SEO checklist have?

Fewer than the 23-item lists suggest, because at least five of their items do not apply to you and six are retired. Twenty-one is what survives when every item has to pass two tests: a primary source that still says it matters, and a check you can actually run. Any list longer than that is usually padding its count with tool-specific steps, or with structured data for features Google no longer shows.

Do I need structured data for a small business website?

LocalBusiness if Gate 2 says you are eligible, and Organization otherwise. Those are the two entries in Google's 25-feature structured data gallery that a small business site plausibly earns. LocalBusiness needs address and name to be valid and everything else is recommended. Skip FAQPage for rich results, skip Service and ProfessionalService, and do not let a plugin emit six types on every page: invalid markup for a feature that does not exist is work with no payout.