The 2026 Canadian SEO Audit Checklist.

1. robots.txt allows Googlebot, Bingbot, and the AI crawlers you want indexed

   Verify GoogleBot, Bingbot, GoogleOther, ChatGPT-User, OAI-SearchBot, PerplexityBot, and Claude-Web aren't disallowed at the path level. Most Canadian sites we audit have at least one over-broad disallow rule that blocks AI engines from key pages.

2. XML sitemap is current, complete, and submitted in Search Console

   Every indexable URL should appear; sitemap should be regenerated on every deploy; submission status in Search Console should be 'Success' with no errors.

3. Canonical tags resolve to themselves on every indexable page

   Render the page, view source, confirm the rel=canonical points to the same URL. Self-referencing canonicals are mandatory; conflicting or missing canonicals are the single most common indexation issue we find.

4. HTTPS is enforced site-wide with HSTS preload

   All HTTP requests 301 to HTTPS. HSTS header is set with max-age >= 31536000 and includeSubDomains. No mixed content warnings on any indexable page.

5. Core Web Vitals pass on the homepage and top 10 landing pages

   LCP < 2.5s, INP < 200ms, CLS < 0.1 on mobile. Pull the Core Web Vitals report in Search Console and confirm 'Good URLs' covers your top revenue pages.

6. Mobile-first usability passes on top revenue pages

   No clipped text, no horizontally scrolling content, tap targets >= 48×48px. Run the page on a real mid-tier Android over Canadian 4G to validate the experience your buyers actually have.

7. International targeting and hreflang are configured (if multi-locale)

   Bilingual (EN/FR) Canadian sites need bidirectional hreflang and an x-default. Most Quebec-targeted sites we audit have unidirectional or missing hreflang, which suppresses French SERP visibility.

8. Pagination, faceted navigation, and parameter handling are deliberate

   Choose noindex+follow, canonical to view-all, or rel=prev/next deprecation handling — but pick one approach and apply it consistently. Inconsistent handling fragments link equity across thousands of URLs.

9. Structured data validates with no errors in the Rich Results Test

   Run the homepage, top service page, top blog post, and top location page through validator.schema.org and Google's Rich Results Test. Zero errors. Warnings are acceptable but should be reviewed.

10. JavaScript-rendered content is reachable in the rendered HTML

    View-source the most important page; if your H1, primary copy, and internal links don't appear in the raw HTML or the rendered HTML in Search Console's URL Inspection, you have an SSR/prerender problem capping rankings.

1. Every indexable URL has a unique title tag, 50–60 characters

   Crawl the site (Screaming Frog, Sitebulb, or equivalent) and export the title tag column. Sort by duplicates. Truncated, missing, or duplicated titles get fixed first; aim for distinct title tags lead-loaded with the primary query.

2. Every indexable URL has a unique meta description, 140–160 characters

   Same crawl, same exercise. Descriptions don't directly rank but they materially affect CTR — a 1% CTR lift on a page receiving 5,000 monthly impressions is meaningful.

3. H1 is unique, present, and matches the user's likely query phrasing

   Exactly one H1 per page; H1 should mirror the search intent the page serves. If your H1 is your brand name, you're competing with yourself for navigational queries while losing the query you actually want.

4. Heading hierarchy (H2 / H3) is logical and scannable

   No skipped heading levels (H1 → H3 with no H2). Headings reflect the actual content sections, not decorative styling. AI engines and Google use heading structure heavily for passage extraction.

5. Image alt text is descriptive and contextual on every meaningful image

   No 'image1.jpg', no 'logo', no empty alts on content images. Hero, decorative, and infographic images get distinct alts that describe what's actually depicted.

6. Internal links use descriptive anchor text — not 'click here' or 'learn more'

   Audit your top 50 pages: every internal link's anchor should describe the destination. Generic anchors waste a per-link signal that Google uses for relevance.

7. Every page has at least 3 contextual internal links from other pages

   Crawl with internal-link-counts column. Pages with zero or one inbound internal link are functionally orphaned regardless of being in the sitemap. Add inbound links from topically relevant pages.

8. Primary content is above the fold, not buried under decorative hero

   Real first-paint content should be visible without scrolling. Heavy hero images that push the H1 + primary copy below the fold reduce both ranking signals and conversion rate.

9. Content depth matches the SERP — no 400-word page targeting a 2,000-word query

   Pull the top 10 ranking pages for your target query. Note their average word count, heading count, and inclusion of FAQs, comparison tables, and original elements. Underbuilt content cannot displace overbuilt content.

10. URLs are short, descriptive, and lowercase with hyphens

    No tracking parameters in canonical URLs. No mixed case. No underscores. Short URLs rank slightly better than long URLs and are dramatically more shareable.

1. FAQPage schema is implemented on every page with 3+ FAQs

   Use Schema.org FAQPage with one Question per FAQ, accepted Answer text matching what's visible on the page. AI engines disproportionately cite FAQ-marked answers.

2. HowTo schema is used on procedural / step-based pages

   When a page genuinely teaches a procedure (audit, fix, install, configure), HowTo schema helps both Google rich results and AI citation. Don't add HowTo to pages that aren't actually procedural — that's structured-data spam.

3. Article + Author + Reviewer schema is consolidated under stable @ids

   Every article references the same Person @id for the author and Organization @id for the publisher. Inconsistent author references fragment your E-E-A-T signal across multiple synthetic entities.

4. Original research, opinion, and first-party data is cited with hedging language

   AI engines preferentially cite content with explicit methodology and uncertainty markers ('most projects we audit show…', 'in our 47-site sample'). Definitive uncited claims are increasingly demoted in citation ranking.

5. Comparison and 'X vs Y' content is published with transparent methodology

   If you publish 'Tool A vs Tool B', state how you tested. AI engines cite methodologically transparent comparisons; reskinned competitor copy gets demoted.

6. Robots.txt deliberately allows the AI crawlers you want citing you

   Default Cloudflare and WordPress hardening configs frequently block ChatGPT-User, OAI-SearchBot, PerplexityBot, and Anthropic's crawler. If you want AI engine citations, you need to allow the crawlers.

7. Brand mentions and citations on third-party sites are tracked monthly

   Run brand-name searches in ChatGPT, Perplexity, Gemini, and Claude monthly. Note who is cited alongside you. Track citation share over time as a real KPI, not a vanity metric.

8. Page speed is fast enough that AI crawlers don't time out

   AI crawlers have shorter timeouts than Googlebot. Pages over 3 seconds to first byte are routinely skipped from citation indexes. Measure with curl -w from a Canadian region against your top pages.

1. Google Business Profile is verified, complete, and category-correct

   Primary category should match your highest-revenue service. Secondary categories add coverage. Hours, services, attributes, products, and photos should all be filled. Most GBPs we audit are 40–60% complete.

2. GBP photos are added monthly — at least 5 per month

   Recency of photo uploads is a ranking factor in the Local Pack. Real photos of work, premises, team, and customers materially outperform stock or rendered imagery.

3. Google posts published weekly

   GBP posts (offers, updates, events) are an underused signal. Even a simple weekly update materially improves Local Pack ranking and click-through.

4. Reviews velocity is consistent — at least 2 new reviews per month

   Review velocity matters more than absolute count past about 50 reviews. Build a request workflow that earns 2–4 new reviews per month consistently.

5. All reviews receive a response within 7 days

   Both positive and negative. Response rate and recency are ranking signals; the response also serves as additional indexable content for the GBP profile.

6. NAP (Name / Address / Phone) is identical on every citation

   Audit your citations on Yelp, Yellow Pages, Bing Places, Apple Business Connect, and the top 10 industry-specific directories. Inconsistencies fragment your local entity.

7. LocalBusiness schema is on the homepage and contact page with areaServed

   Use LocalBusiness or a more specific subtype with explicit areaServed properties listing every neighbourhood / city you genuinely serve. AI engines use areaServed to filter citation eligibility.

8. Embedded Google Map on the contact page using the verified place ID

   Map embed should use the Place ID, not just a search query, so the embed surfaces your actual GBP listing. This subtly reinforces the entity association.

1. Organization or LocalBusiness schema is present site-wide

   Either via header injection or on every relevant page. Establishes the canonical entity reference that all other schema can point at via @id.

2. Every article uses Article schema with author, publisher, datePublished, dateModified

   Author and publisher should be @id references to the same Person and Organization nodes used elsewhere. Inconsistent author entities fragment E-E-A-T.

3. Service pages use Service or ProfessionalService schema with provider + areaServed

   Service schema with explicit provider (your @id), serviceType, and areaServed lets Google and AI engines understand what you offer and where.

4. BreadcrumbList schema mirrors visible breadcrumb navigation

   Implement on every page. Google uses BreadcrumbList to display navigational SERP enhancements and to understand site hierarchy.

5. Reviews are surfaced via Review or AggregateRating schema, never fabricated

   Connect to verifiable third-party review sources (Google Business Profile, Trustpilot, etc.). Self-reported review schema without a verifiable source is a manual-action risk.

6. Product / Offer schema on ecommerce pages with availability + price + currency

   Drives product rich results and shopping-graph eligibility. Always include priceCurrency (CAD where applicable) and itemCondition.

7. VideoObject schema on every embedded video with thumbnailUrl, uploadDate, duration

   Video carousel SERP eligibility requires complete VideoObject markup. Most Canadian sites with embedded videos ship VideoObject schema on zero of them.

8. All schema validates in the Rich Results Test and validator.schema.org

   Zero errors. Warnings reviewed. Run on representative pages monthly; structured data drift after CMS updates is the most common cause of silent rich-result loss.

1. Backlink profile is reviewed monthly for toxicity and growth

   New referring domains tracked month-over-month. Toxic links assessed; disavow file maintained for confirmed manipulative spam. Don't disavow defensively — only confirmed bad neighbourhoods.

2. At least one substantive new content asset shipped per month

   Editorial cadence matters more than peak velocity. A monthly substantive piece (>2,000 words, original framing or data) outperforms quarterly bursts followed by silence.

3. Content refreshes — top 10 traffic pages reviewed every 90 days

   Update dates, refresh data, add new sections, expand FAQs. Refresh signals + content depth gains compound.

4. Search Console + GA4 dashboards review weekly

   Weekly review catches ranking drops, indexation issues, and conversion trend shifts within days, not weeks. Monthly-only review is too slow for early diagnosis.

5. Conversion tracking is connected end-to-end — organic visit to closed revenue

   Most agencies optimize for organic traffic because that's what they can show. The agencies you should hire optimize for organic-attributed pipeline. Connect the chain.

6. Quarterly competitive citation-share benchmarking

   Quarterly run brand + category queries through ChatGPT / Perplexity / Gemini / Google AIO and track citation share against your top 3 competitors. Trend matters more than snapshot.

How to use this checklist

This is the working checklist we use on real Canadian engagements. We treat it as a sequence: technical first, then on-page, then schema and AI/AEO, then local (where applicable), then content cadence. Working out of order — particularly tackling content velocity before fixing technical foundations — is the single most common reason SEO programs underperform.

Score your site honestly. If you're below 70% on technical or on-page, prioritize fixing those before any new content investment. If you're above 90% on the first two but below 50% on AI/AEO and schema, the highest-leverage next move is closing those gaps — they compound the value of every page you've already built.

If you'd like an honest written audit of your site against this exact checklist — typically a 12–18 page document delivered within three business days — request one here. No sales pitch, just findings. We'll tell you what's working, what's broken, and the priority order to fix it in.