SEO best practices for 2026: the checklist we run first

Start an SEO review with the pages the business needs customers to find. For each one, establish whether Google has discovered it, fetched it, indexed it and shown it for relevant searches. These are separate stages, and a single traffic total cannot explain all of them.

This checklist is for a business website with service pages, supporting articles and a working enquiry route.

Read Search Console before changing the page

Review indexing status, the last recorded crawl and the selected canonical. Compare impressions and clicks by page over a recorded period, then examine queries where available. A live URL test checks current fetchability; the stored inspection record describes Google’s earlier indexing assessment.

Flat impressions with fewer clicks can have several causes: query mix, result position, competing results, search features or the snippet. Do not attribute the change to AI answers from that pattern alone. Record the affected pages and dates before proposing a fix.

Verify the public response and main content

Confirm a successful response for pages intended to be public. Check robots rules, indexing directives, canonical and sitemap inclusion. Inspect the main content in rendered HTML and Google’s test output. Investigate server or CDN failures when there is an observed failed request, rather than repeatedly changing settings that already pass.

Google’s recrawl guidance explains the available request methods. Requests do not guarantee inclusion. Keep the test date and last crawl date so a later result can be assessed accurately.

Make important pages easy to find

Check for orphan pages, broken links, unnecessary redirect chains and navigation that hides a core service. Add links in relevant hubs and articles where the next page helps the reader. Use descriptive anchors; avoid forcing every page into the same keyword pattern.

Match titles and headings to the actual offer

Use the page’s search queries as evidence of how people describe the subject, then assess whether the content answers those needs. A title should identify the offer and its useful distinction. An H1 should introduce the page clearly. Do not add locations, delivery times or expertise claims that the business cannot support.

Review related pages together to catch overlapping intent. A country page needs accurate delivery arrangements and useful local context, rather than a copied page with only the place name replaced.

Check completeness and project evidence

A service page should explain who the work suits, inputs required, deliverables, exclusions, responsibilities and the next step. Use approved project descriptions for evidence. Label illustrative workflows and remove unsupported result claims.

The people-first content guidance is a useful review framework. It does not specify a preferred word count. Keep useful pages and improve actual gaps; a mass deletion or date refresh is not a substitute for that assessment.

Review local information where the business is eligible

Keep the real business name, address, contact details and opening hours accurate. Check profile eligibility and the location’s customer-facing arrangements. Use the local SEO checklist for that work. Do not invent an office to target a city.

Measure performance and enquiries

Use field performance data when available, supported by lab diagnostics. Investigate the slow interaction, image or layout shift observed on the page. Our which changes actually move the numbers guide explains the diagnostic process. Good scores alone do not establish better rankings or sales.

Test the contact route on mobile and desktop. Check whether a recorded form event represents an accepted submission, a click or a qualified enquiry. Keep test traffic and unrelated hosts out of business reporting.

Plan one reviewable release

Prioritise failed access, missing information and broken enquiries before cosmetic changes. Record the intended outcome and preserve existing URLs unless a reviewed redirect or retirement is necessary. Follow the release with public response checks, priority URL inspection and a dated comparison of search and enquiry data.

For deeper diagnosis, read what we do about citations and the scope of AI search optimisation. Discuss SEO retainer work when the findings need ongoing implementation and review.

Decide which evidence is missing

For a discovered page with no recorded crawl, check whether the URL is linked from relevant public pages and present in the canonical sitemap. Confirm that the public response is usable and record when it was published. Repeatedly rewriting the same paragraph does not show whether discovery has improved.

For a crawled page that is excluded, inspect the stored crawl, selected canonical and actual content available at that time. Compare the live page for changes. Review whether it serves a distinct need and provides enough accurate detail for a customer to assess the offer. Do not assume every exclusion is a content-length problem.

For an indexed page with impressions but few clicks, inspect query relevance, position and the result’s presentation before changing the offer. For a page that receives visits without enquiries, test the next step and assess whether those visitors match the intended audience. Keep these findings in separate rows of the review so an editorial change is not reported as a server fix, or a successful fetch as an enquiry improvement.