SEO tips that still work in 2026: twenty-five, with the reason for each

You want SEO tips that still work and are worth the hour you have. Here are twenty-five that match what Google documents, with the reason for each. No tricks, just things you can ship on a small site.

By , founder of Porteur · Updated 14 September 2026 · Markdown

What Google rewards as of 2026

Google finds pages by crawling links and sitemaps, renders them, then indexes some of them. It ranks by relevance, quality, usability and context. There is no payment or submit form to rank. A public page linked from a known page is found.

The ranking systems Google names include link analysis and PageRank, helpful content guidance folded into core systems, BERT, MUM, passage ranking, and SpamBrain. What you control on a product site: answer the query with original, trustworthy content, earn useful links, be fast and usable, and be crawlable and indexable.

Pages: write and mark up pages that answer the query

  • Write one page per intent: match the search. Example: put pricing on /pricing, not in a PDF. Reason: Google ranks pages that answer the query type. Mistake prevented: thin pages that do not match intent.
  • Put the primary term in the title, early. Example: “Email API pricing” for /pricing. Reason: titles help systems understand the topic. Mistake prevented: clever titles that hide the subject.
  • Keep titles to about 60 characters. Reason: longer titles are truncated in many results. Mistake prevented: cut-off titles that depress CTR.
  • Use a clear h1 that matches or supports the title. Example: h1 “Email API pricing” on /pricing. Reason: headings help structure and relevance. Mistake prevented: mismatched h1 that confuses the page focus.
  • Answer the question in the first paragraph. Example: /guides/getting-started starts with the install command. Reason: Google surfaces helpful content that gets to the point. Mistake prevented: burying the answer below fluff.
  • Use plain URLs. Example: /guides/create-invoice, not /G123?ref=home. Reason: crawl and deduplication work better with stable, descriptive slugs. Mistake prevented: parameter bloat and duplicate URLs.
  • Add a short, accurate meta description. Example: 150 characters that reflect the page. Reason: Google may use it in snippets, improving CTR. Mistake prevented: auto-pulled nonsense lines in results.
  • Avoid keyword stuffing. Write naturally. Example: Say “invoice template” where it fits, not ten times. Reason: spam systems downgrade stuffing; density is folklore. Mistake prevented: penalties and poor copy.
  • Use descriptive alt text on images. Example: alt=“Email API dashboard graph”. Reason: images are indexed and improve accessibility; helps understanding. Mistake prevented: blank or spammy alt attributes.
  • Mark up product or software pages with the right schema. Example: SoftwareApplication on /features, FAQ on /faq. Reason: eligible rich results can improve appearance. Mistake prevented: invalid or missing structured data.

A fixed page reads like a clear help doc: title about the topic, h1 aligned, first paragraph answers, headings that map the subquestions, links to next steps, and no filler at the top.

Pages: make them discoverable and renderable

  • Make every page crawlable. Reason: Googlebot must fetch the HTML. Mistake prevented: blocked paths in robots.txt or auth walls.
  • Ship a single canonical URL for each page. Example: /pricing, not both /pricing and /pricing/. Reason: indexing groups duplicates and picks one. Mistake prevented: split signals and wrong canonicals.
  • Use rel=canonical correctly. Example: on /pricing?ref=ad, point to /pricing. Reason: helps Google select the canonical. Mistake prevented: “Alternate page with proper canonical tag” on a URL that should not be alternate.
  • Avoid rendering blockers for key content. Reason: Google renders in a second wave; critical content should be in HTML or hydrate fast. Mistake prevented: empty HTML that never gets indexed fully.
  • Provide an XML sitemap and keep it clean. Reason: sitemaps aid discovery, especially for new sites. Mistake prevented: orphan pages and stale URLs left for months.
  • Fix soft 404s and true 404s. Reason: error states waste crawl and hurt user trust. Mistake prevented: indexed stubs like “No products found” that look like thin content.
  • Localise with proper hreflang if you have versions. Reason: helps serve the right language or region. Mistake prevented: wrong-country pages ranking for brand searches.

Site: structure and speed that help ranking systems

  • Keep click depth low to money pages. Example: /pricing reachable from the home and header. Reason: internal prominence aids discovery and importance. Mistake prevented: buried pages that never get crawled often.
  • Link related pages in body copy. Example: from /guides/getting-started to /guides/authentication. Reason: internal links help relevance and PageRank flow. Mistake prevented: orphan guides that stall.
  • Use breadcrumbs and consistent nav. Reason: helps users and Google understand structure. Mistake prevented: duplicate paths and messy hierarchies.
  • Serve over HTTPS. Reason: HTTPS is a lightweight signal and improves user security. Mistake prevented: mixed content warnings and loss of trust.
  • Meet Core Web Vitals. Aim for LCP 2.5 s, CLS 0.1, INP 200 ms. Reason: page experience is a signal and a conversion driver. Mistake prevented: slow, jumpy pages that bleed users.
  • Control third party scripts. Example: defer chat and analytics where possible. Reason: they delay paint and interaction. Mistake prevented: regressions that push LCP and INP over the thresholds.
  • Avoid intrusive interstitials. Reason: Google lists them as a negative for usability. Mistake prevented: demotions for mobile-hostile layouts.

A fixed site has a flat structure, fast hero images, few blocking scripts, and clear links from the home to /pricing, /docs and /guides. Your core pages load in under three seconds on a phone on 4G, and nothing shifts under the finger.

Measurement: know what moved and why

  • Verify Search Console and check the last 16 months. Reason: it is Google’s own data on impressions, clicks and position. Mistake prevented: guessing from rank trackers alone.
  • Use the Performance report to split by page and query. Example: /pricing lost clicks after 12 August. Reason: impressions, CTR and position show which lever moved. Mistake prevented: chasing the wrong page.
  • Read impressions correctly. An impression is counted when a link to your page appears on a loaded results page. Mistake prevented: thinking zero scroll equals zero impression for web search.
  • Compare the last 28 days to the 28 before, then year on year. Reason: isolates seasonality and updates. Mistake prevented: calling a seasonal dip a penalty.
  • Diagnose drops by date. Check Google’s Search Status Dashboard, Manual actions, Security issues, deploy logs and the results page itself. Mistake prevented: missing a noindex ship or a new rival feature.
  • Use URL Inspection for a single page. Reason: you can request indexing of one URL and see crawl, index and canonical state. Mistake prevented: waiting weeks for a fix to be seen.

Local and media specifics that still matter

  • Create and maintain a Google Business Profile if you serve a place. Example: your studio in Austin. Reason: local results use proximity and profile data. Mistake prevented: missing out on map pack visibility for brand and category terms.
  • Put NAP details on your contact page and footer. Reason: consistency helps local understanding. Mistake prevented: mismatched addresses across the web.
  • Add transcripts and descriptions for videos you embed. Reason: text helps indexing and passage ranking on pages with media. Mistake prevented: video pages with no context.
  • Name images descriptively and compress them. Example: email-dashboard.png, WebP or AVIF, width and height set. Reason: images are indexed and speed affects LCP. Mistake prevented: slow hero images and meaningless file names.

Avoid classic blockers that waste months

  • Never ship staging noindex to production. Check meta robots and X-Robots-Tag. Reason: noindex excludes pages from the index. Mistake prevented: a live site invisible to search.
  • Do not block your site in robots.txt. Reason: robots.txt stops crawling, not indexing of known URLs, but it prevents discovery. Mistake prevented: a site that cannot be fetched.
  • Do not change URL structure without redirects. Use 301s with a map. Reason: signals consolidate over time; 301 preserves them. Mistake prevented: wiping rankings on a rename.
  • Do not chase a fake domain authority score. Reason: Google has said there is no single domain authority number. Mistake prevented: decisions based on folklore instead of content and links.
  • Ignore keyword density formulas. Reason: Google has said there is no target density. Mistake prevented: unreadable copy and spam signals.

Putting it together: a weekly loop that works

  1. Pick one page that matters

    Choose /pricing, /features/email-api or a guide that can convert. Confirm its query with an incognito search and Search Console data.

  2. Tighten the page

    Fix the title and h1, answer first, add internal links to and from related pages, add or fix schema, and compress media.

  3. Ship a related asset

    Publish one thing worth citing: a comparison, a benchmark, or a step by step guide that helps the same user.

  4. Improve speed by one metric

    Cut LCP by compressing the hero and preloading the font, or reduce INP by deferring a third party script.

  5. Measure and note

    In Search Console, mark the date. Watch impressions and CTR for that URL and query for the next 28 days. Roll fixes to the next page.

Questions

Sources

Check my site, free

Drop your homepage URL and get a free check in about thirty seconds that reads your site and the searches around it, and shows three findings you can ship next.

  • Free check, no card
  • Read-only, your own accounts
  • Readable by your agent

Read next