01 / DiscoveryMake the right URLs discoverable
Shopify can expose alternate product paths, filters and search pages. Decide which URL family deserves crawl attention.
01Robots rules preserve valuable assets
Confirm robots.txt does not block CSS, JavaScript, product media or routes needed for rendering.
02XML sitemap contains canonical, live URLs only
Sample products, collections, articles and localized pages; remove redirects, errors and duplicate parameters.
03Navigation reaches commercial pages
Important collections and guides should not depend on search, filters or JavaScript-only discovery.
Give Google one clear version
Every page should state whether it belongs in search and which URL represents the content.
04Canonical tags are self-consistent
Check products in collections, pagination and tagged content. Canonicals must resolve directly to a 200 page.
05Indexable pages return indexable responses
Avoid accidental noindex, soft 404 copy, redirect chains and canonical targets that disagree with sitemaps.
06Facets do not create an infinite crawl space
Inventory parameters for sort, filter, currency and tracking; retain only pages with independent demand.
Make every template explain itself
Titles, headings, copy and structured data should agree on what a product, collection or guide is for.
07One descriptive title and one visible H1
Use unique, concise language. Do not turn every close keyword variation into another page.
08Product schema matches visible facts
Validate offers, price, availability, brand, variants and reviews against the rendered page.
09Thin collection pages gain decision value
Add useful selection logic, compatibility notes and original merchandising context—not filler paragraphs.
Protect the buying journey
Performance work should reduce delay and instability on real storefront templates, not chase a lab score alone.
10LCP media is sized and prioritized
Serve responsive formats, reserve dimensions and avoid lazy-loading the main product or hero image.
11Apps justify their JavaScript cost
Measure third-party scripts by template and remove dormant features before optimizing around them.
12Tap targets and type remain usable
Use 16px mobile body text, 44px controls and stable layouts across phones, tablets and older browsers.
Connect languages, data and decisions
International versions must be real equivalents. Measurement must tell you what to fix next.
13Hreflang clusters are reciprocal
Each localized page references itself and every true alternate, using valid language and region codes.
14Search Console data maps to page intent
Review queries, landing pages and indexation together. Impressions without clicks may signal a weak snippet or intent mismatch.
15Every recommendation has evidence
Record the affected URLs, source data, expected impact, owner and validation date.
Verify the page a buyer actually uses
16Identify one product and its variants
Open an actual product and record the selected size, colour and destination market. Compare the name, price and availability for each option you test. A product-wide statement must not conceal an unavailable variant.
17Answer the purchase decision
Check whether a buyer can determine dimensions, material, compatibility, included items and relevant limitations without contacting support. Use supplier specifications or measured facts; do not invent benefits or expand a manufacturer's description with filler.
18Describe images accurately
Inspect the main image and a detail image at phone width. Alt text should explain useful visual information, not repeat every keyword. Decorative images can have empty alt text; do not put essential specifications only inside an image.
19Keep delivery and return terms consistent
Compare the product page with the policy and checkout for the same destination. Record costs, restrictions and timing only when verified. Clearly distinguish estimates from guarantees and flag missing information for the store owner.
20Match structured offers to visible facts
Use Google's Rich Results Test for an individual product. Compare selected offer price, currency and stock with the page and the corresponding Merchant Center item. Correct the theme or app that owns the output; do not add a conflicting second schema block.
21Separate stockouts from discontinued URLs
For temporary stockouts, retain useful information and clearly control purchases. For permanent retirement, document whether a genuinely equivalent successor exists. Do not redirect unrelated products to the homepage or claim every kept page will retain rankings.
Measure the real interaction before changing code
22Use field data and lab tests for different questions
Review Search Console's Core Web Vitals report for real-user groups, then use PageSpeed Insights to investigate a representative URL. No field data means unavailable, not a perfect score. A single lab run is a diagnostic, not a business outcome.
23Record LCP, INP and CLS separately
Google's good thresholds are LCP within 2.5 seconds, INP at or below 200 ms and CLS at or below 0.1. Check mobile and desktop evidence separately and record the measurement period. Passing these metrics does not guarantee a ranking.
24Repeat a controlled before-and-after test
Keep the template, device profile and network conditions consistent. Test multiple runs before and after a single change. Save the original configuration and reverse the change if buying, navigation or consent behaviour breaks.
25Find the actual largest-content element
Use the performance trace to identify the LCP element rather than assuming it is the hero. Check its transfer size, discovery and loading priority. Do not preload every image: compare the effect of each change on the critical request path.
26Test keyboard and touch paths
Open navigation, select a variant, dismiss an overlay and reach the cart using touch and keyboard. Check visible focus, labels and text zoom. The 16px text and 44px controls above are practical design targets, not a Google indexing rule.
27Audit app changes with the feature owner
List scripts that load on product, collection and article templates. Confirm whether each feature is used before removal. Test on a duplicate theme where possible; a faster page with broken checkout or lost consent is not an improvement.
Validate each market and leave an audit trail
28Use real language equivalents
Compare the actual content and purpose of two alternate URLs. A language homepage is not an equivalent of a translated product or article. Confirm both destinations return the intended page and that reciprocal annotations use full URLs.
29Keep canonical and hreflang roles distinct
A canonical chooses the preferred representative of duplicate content; hreflang describes alternate language or regional versions. Do not point every localized page to English as a shortcut. Check Google's selected canonical in URL Inspection separately.
30Test market selection without forcing a redirect
Visit a localized URL directly while logged out. Confirm a visitor can select another market, understand currency and shipping restrictions, and retain access to the requested content. Document what the store actually serves, not just the language selector labels.
31Check a conversion without calling it a customer
Use a clearly labelled test inquiry or approved test order. Verify front-end confirmation, server storage and analytics independently. Exclude synthetic records from business reporting and do not send personal data in analytics event parameters.
32Close each finding with evidence
Maintain URL, symptom, source, proposed change, owner, rollback and recheck date. Reinspect the affected URL after deployment. Sitemap submission, indexing request acceptance, indexed status and a real order are separate completion states.
How to use this checklist
Start with one product, collection, article and localized page, then expand only when the template finding is confirmed. Mark a check as not applicable when the feature is absent, and explain why. These 32 checks are an editorial workflow, not certification or a guarantee of indexing.
References checked September 12, 2026: Shopify SEO documentation, Google product data, Core Web Vitals guidance and localized URL annotations.
For a stock transition, use the out-of-stock decision guide. For paid implementation scope, see our audit service.
Completion rule:A checked box needs a URL, screenshot, crawl sample or Search Console record. “Looks fine” is not evidence.
