Sample report · fictional website

Northwind Shoes, its competitors and all numbers are invented. The structure, checks, limits and scoring method are exactly those of a real report. Get a free preview for your website

northwind-shoes.exampleFull report
northwind-shoes.example

Full SEO report · July 21, 2026

northwind-shoes.example

Measured index · not a Google score
  • 2
    Critical
  • 2
    High
  • 3
    Medium
  • 3
    Low

Confirmed issues by priority. Also in the report: 4 recommendations · 1 hypothesis.

01 · Executive summary

A solid store that Google can’t fully see

Northwind Shoes has real product knowledge, two stores with a free fitting service and clean technical basics: all 26 pages we checked load in one step and have their own titles. What holds the site back is that Google is being told the wrong things: an SEO-app setting points all 8 sampled categories to the homepage, and the store pages ask not to be indexed. On phones, product photos take almost 6 seconds to appear. None of this is expensive: the two critical fixes take less than half a day.

Site platform: Shopify (theme Dawn 15). The fix instructions are written for it.

What already works

  • HTTPS everywhere: http:// and www redirect in one 301 step, and a random URL returns a real 404.
  • All 26 sample pages return 200 and have one descriptive main heading.
  • Product descriptions written by staff who fit shoes — specific, not copied from manufacturers.
  • Free gait analysis and free 30-day returns: strong, rare selling points.
  • Structured data has 0 errors, and name, address and phone match on every page.
  • Desktop speed is good: homepage LCP 1.2 s in the lab.

Do these first

  1. Step 1 · F01Category canonical tags point to the homepageStore admin (setting) + developer (theme check) · low effort, 1–2 hours
  2. Step 2 · F02Store pages carry “noindex”Store admin · low effort, 15–30 minutes
  3. Step 3 · F03Slow mobile product page (LCP 5.8 s)Store admin (app embeds) + front-end developer · medium effort, 1–2 days

This is the order of work, not a promise of rankings. We recommend a recheck on the same pages 30–60 days after the fixes.

GEO · AEO · LLMO · AI SEO

AI search at a glance

Access for AI crawlers
Open to every crawler behind AI answers
AI search readiness
5 of 13 checks passed
Visits from AI assistants
Google Analytics is not connected
AI search visibility (GEO) →

02 · Measured index

80 out of 100 — what the number means

The index summarises 10 fixed rules measured on the sample (26 URLs): 9 page checks and the mobile lab performance of one representative page. The rules and weights are the same for every site and every recheck. It is our diagnostic of this sample — not a Google score and not a ranking forecast. Content, competitors, local SEO, structured data, links and AI search are assessed in words, not in the number.

15 findings: 10 confirmed issues (2 critical, 2 high, 3 medium, 3 low), plus 4 recommendations and 1 hypothesis.

Category results (0–100) for the sample (26 URLs), July 21, 2026. Weights in brackets. Performance is the median Lighthouse mobile score of the representative page (3 lab runs). Weighted check coverage 95.96 %: mobile usability is checked on the 5 pages rendered in a real browser, not on every sample page.

The 10 rules behind the number

Rules of the measured index (method seorecheck-observed-v1) with tested units and results
RuleWeightTestedResult
Page answers with HTTP 200sr-http-ok1526 of 26100
Page may be indexedsr-indexable1526 of 2688 · 3 fail
Canonical tag is consistentsr-canonical-consistent1026 of 2669 · 8 fail
Title is presentsr-title-present826 of 26100
Title is uniquesr-title-distinct726 of 2677 · 6 fail
Main heading is presentsr-heading-present526 of 26100
Meta description is presentsr-description-present526 of 2685 · 4 fail
Images have alt textsr-image-alternatives526 of 2696 · 1 fail
Usable on a phone (browser check)sr-mobile-usable55 of 2680 · 1 fail
Mobile lab performance (3 runs)sr-lab-performance251 of 155 · median

03 · Priority actions

What to fix first

Confirmed issues, ordered by expected impact on search visibility and effort. Select one to jump to its cause, fix and check. Recommendations and hypotheses follow below the table.
Confirmed issues in priority order
#IssueScopeWho · timeBenefit
F01Category canonical tags point to the homepagecriticalEffort: low8 / 8Store admin (setting) + developer (theme check)1–2 hoursIndexing
F02Store pages carry “noindex”criticalEffort: low3 / 3Store admin15–30 minutesLocal visibility, store visits
F03Slow mobile product page (LCP 5.8 s)highEffort: medium2 / 2Store admin (app embeds) + front-end developer1–2 daysSpeed, conversion
F04Duplicate titles on colour pageshighEffort: medium6 / 8Store admin + developer2–4 hours (titles) · 1–2 days (merge)Click-through, indexing
F05Product markup without offersmediumEffort: low8 / 8Developer2–4 hoursRicher results, click-through
F06Links to removed productsmediumEffort: low5 / 26Store admin1–2 hoursConversion, link value
F07Size picker hard to use on mobilemediumEffort: medium1 / 5Front-end developer3–5 hoursConversion, mobile usability
F08Missing meta descriptionslowEffort: low4 / 8Store admin1 hourClick-through
F09Icon buttons without accessible nameslowEffort: low26 / 26Developer30 minutesAccessibility
F10Parameter URLs in the sitemaplowEffort: low61 / 412Store admin (+ developer for robots.txt)30 minutesCrawl focus
  1. F01 Critical · low effort
    Category canonical tags point to the homepage

    8 of 8 sampled category pages · 1–2 hours

  2. F02 Critical · low effort
    Store pages carry “noindex”

    3 of 3 store pages · 15–30 minutes

  3. F03 High · medium effort
    Slow mobile product page (LCP 5.8 s)

    2 of 2 speed-tested pages · 1–2 days

  4. F04 High · medium effort
    Duplicate titles on colour pages

    6 of 8 sampled product pages · 2–4 hours (titles) · 1–2 days (merge)

  5. F05 Medium · low effort
    Product markup without offers

    8 of 8 sampled product pages · 2–4 hours

  6. F06 Medium · low effort
    Links to removed products

    5 of 26 sample pages · 1–2 hours

  7. F07 Medium · medium effort
    Size picker hard to use on mobile

    1 of 5 browser-rendered pages · 3–5 hours

  8. F08 Low · low effort
    Missing meta descriptions

    4 of 8 sampled product pages · 1 hour

  9. F09 Low · low effort
    Icon buttons without accessible names

    26 of 26 sample pages (one shared header) · 30 minutes

  10. F10 Low · low effort
    Parameter URLs in the sitemap

    61 of 412 sitemap entries · 30 minutes

Share of checked pages or URLs affected by each confirmed issue (affected ÷ checked). Colour shows priority.

Recommendations and hypotheses

Improvements and ideas, not defects. A hypothesis needs your data to confirm it.

04 · Technical & indexing

Can Google find and keep your pages?

Checked: robots.txt, sitemap.xml (412 URLs parsed), all 26 sample pages (status, redirects, canonical tags, robots directives, headers), 4 probes (http→https, www, trailing slash, a random URL) and raw HTML against the rendered page.

F01criticalConfirmed issue8 of 8 sampled category pages

Category pages tell Google the homepage is the “real” page

All 8 category pages in the sample carry two canonical tags. The first names the page itself; the second — which Google may treat as the deciding one — names the homepage. The sitemap lists 6 more categories and 28 paginated category URLs on the same template; we did not fetch those, so check them after the fix instead of assuming.

Evidence

/collections/womens-running

<link rel="canonical" href="https://northwind-shoes.example/collections/womens-running">

<link rel="canonical" href="https://northwind-shoes.example/"> ← added by the SEO app

Why it matters
A canonical tag says “this is the main version of the page”. When a category names the homepage instead, Google can drop the category from results and show the homepage — or nothing — for searches like “women’s trail running shoes”. Categories are usually a store’s best landing pages for shopping searches.
Cause
The SEO app’s “Canonical URL” setting is “Site root” for all pages, and the theme also prints its own correct canonical, so every category ends up with two conflicting tags.
How to fix

In an app or pluginNeeds a developer

  1. In the SEO app’s settings, switch “Canonical URL” to “Page URL” — or turn its canonical output off and keep the theme’s tag.
  2. Once the fix is live, request indexing for the main categories in Google Search Console (URL Inspection → Request indexing).

For the developer

  1. Make sure layout/theme.liquid prints exactly one canonical tag from Shopify’s canonical_url, and that no app snippet adds a second one.
  2. Check a paginated category (?page=2): its canonical is its own URL, not page 1.
<link rel="canonical" href="{{ canonical_url }}">

Official guide: Shopify: SEO metadata in themes (canonical_url)

Who · effort
Store admin (setting) + developer (theme check) · Low effort, 1–2 hours
How to verify
Open a category, press Ctrl+U and search for rel="canonical": exactly one result, with the page’s own URL. In Search Console → URL Inspection, “User-declared canonical” equals the inspected URL.
Show 8 affected sampled category pages
  • /collections/womens-running
  • /collections/mens-running
  • /collections/trail-running
  • /collections/walking-shoes
  • /collections/wide-fit
  • /collections/boots
  • /collections/sale
  • /collections/new-arrivals
  • Same template, not fetched: 6 more categories and 28 paginated URLs listed in the sitemap
F10lowConfirmed issue61 of 412 sitemap entries

The sitemap lists filter and page-number URLs

61 of the 412 URLs in your sitemaps contain query parameters, such as ?page=3 or ?filter.v.option.size=10. All of them are in a second sitemap that a sitemap app publishes and lists in robots.txt next to Shopify’s own sitemap.xml. A sitemap should list the main version of each page.

Evidence

61 of 412 <loc> entries contain “?”

https://northwind-shoes.example/collections/trail-running?filter.v.option.size=10

Why it matters
A sitemap tells Google which URLs you consider important. Filter combinations are near-copies of their category and take attention away from products.
Cause
A sitemap app adds every URL it finds in the navigation, including filter links — although Shopify already generates sitemap.xml with every product, collection, page and post.
How to fix

In an app or pluginNeeds a developer

  1. In the sitemap app’s settings, exclude URLs with query parameters — or uninstall the app (Settings → Apps → the app → Uninstall): Shopify’s own sitemap.xml already covers the store.
  2. Keep paginated categories reachable through normal links; they do not need to be in a sitemap.

For the developer

  1. If the app leaves its “Sitemap:” line in templates/robots.txt.liquid after uninstalling, remove that line.

Official guide: Shopify Help: Find your sitemap · Shopify Help: Uninstall apps

Who · effort
Store admin (+ developer for robots.txt) · Low effort, 30 minutes
How to verify
The sitemaps listed in robots.txt contain no entries with “?”.
Show 5 affected sitemap entries
  • /collections/womens-running?page=2
  • /collections/womens-running?page=3
  • /collections/trail-running?filter.v.option.size=10
  • /collections/boots?filter.p.vendor=Northwind
  • /collections/sale?page=2
  • + 56 more parameter URLs in the app’s sitemap

Checked and fine

  • http:// and www. redirect to https://northwind-shoes.example in one 301 step; a random URL returns a real 404.
  • All 26 sample pages return 200; titles, headings and canonical tags are identical in raw and rendered HTML.
  • robots.txt blocks only cart, checkout and account pages.

05 · On-page & metadata

What searchers see before they click

Checked: title, meta description and main heading (H1) on all 26 sample pages.

F04highConfirmed issue6 of 8 sampled product pages

Colour versions of a shoe compete with each other

6 of 8 sampled product pages share their title and meta description with another URL, because every colour is a separate product page with identical text — for example /products/trailpace-gtx-womens and /products/trailpace-gtx-womens-moss.

Evidence

/products/trailpace-gtx-womens

<title>TrailPace GTX Women’s | Northwind Shoes</title>

/products/trailpace-gtx-womens-moss

<title>TrailPace GTX Women’s | Northwind Shoes</title>

Why it matters
When several pages look identical, Google picks one to show — possibly a colour you rarely stock. Links and reviews are split across the copies instead of strengthening one page.
Cause
Colours were created as separate products instead of variants of one product, and titles are generated from the product name only.
How to fix

In the admin, no codeOwner's decision

  1. Short term: Products → open each colour product → Search engine listing → Page title and Meta description: add the colour (“TrailPace GTX Women’s — Moss”).
  2. Long term: merge the colours into one product with variants and delete the old colour products. Then Content → Menus → View URL redirects → Create URL redirect from each old colour URL to the main product (a redirect only works once the old URL returns 404).

Official guide: Shopify Help: Add and update products · Shopify Help: URL redirects

Who · effort
Store admin + developer · Medium effort, 2–4 hours (titles) · 1–2 days (merge)
How to verify
Each colour page has its own title in the page source; after a merge, the old colour URLs answer 301 to the main product.
Show 6 affected sampled product pages
  • /products/trailpace-gtx-womens
  • /products/trailpace-gtx-womens-moss
  • /products/cityglide-2-mens
  • /products/cityglide-2-mens-graphite
  • /products/ridgeline-boot-womens
  • /products/ridgeline-boot-womens-oak
F08lowConfirmed issue4 of 8 sampled product pages

Four product pages have no meta description

4 of 8 sampled product pages have no meta description, so Google builds the snippet from whatever text it finds, such as the shipping bar.

Evidence

/products/trailpace-gtx-womens

<meta name="description"> — missing

Why it matters
A description does not change rankings, but it is your chance to write what searchers read under the title: fit, width, free fitting.
Cause
The Meta description field in these products’ search engine listing was left empty.
How to fix

In the admin, no code

  1. Products → open each of the 4 products → Search engine listing → Meta description: 1–2 sentences on fit, width and what sets the shoe apart → Save.

Official guide: Shopify Help: Add and update products

Who · effort
Store admin · Low effort, 1 hour
How to verify
Each product page source contains one <meta name="description"> with page-specific text.
Show 4 affected sampled product pages
  • /products/trailpace-gtx-womens
  • /products/cityglide-2-mens
  • /products/ridgeline-boot-womens
  • /products/tempo-racer-3

Checked and fine

  • Every page has one descriptive H1.
  • Category and guide titles are unique and describe the page.

06 · Content & search intent

Do your pages answer what shoppers ask?

Checked: all 26 pages read for intent and usefulness; 11 key pages (categories, guides, store pages) reviewed in depth for experience, expertise and trust (E-E-A-T). Word count alone is never treated as a quality signal.

F11mediumRecommendation6 of 8 sampled category pages

Category pages give shoppers no advice

6 of 8 sampled category pages have 0–40 words of text besides the product grid. Both competitor category pages we compared have 310–450 words of fitting advice. The gap is not the word count itself but the answers shoppers look for: fit, width, terrain, when to replace.

Evidence

/collections/wide-fit · 0 words outside menus and product tiles

stride-and-co.example/shop/wide-fit · 450 words of fitting advice

Why it matters
For buying searches, Google tends to rank pages that help people choose. Your staff’s fitting knowledge — your strongest asset — is not visible on the pages that should rank.
Cause
Categories were set up as plain product grids.
How to fix

Texts and mediaIn the admin, no code

  1. Start with wide-fit, trail running and walking shoes: Products → Collections → open the collection → Description: 150–300 words of real advice from your fitters (Dawn shows it above the product grid).
  2. Answer 3–4 questions customers actually ask in the store, in text.
  3. Link each category to the matching guide and to the store fitting page.

Official guide: Shopify Help: Collection settings

Who · effort
Store staff (writing) + store admin · High effort, 2–3 hours per category
How to verify
The text is in the HTML source (not only after JavaScript), and each category links to its guide.
Show 6 affected sampled category pages
  • /collections/mens-running
  • /collections/trail-running
  • /collections/walking-shoes
  • /collections/wide-fit
  • /collections/boots
  • /collections/new-arrivals
F12lowRecommendation3 of 3 sampled guides

Guides don’t say who wrote them or which shoes to buy

The 3 sampled guides have no author, no date and no links to the products they discuss. The advice itself is specific and clearly written from store experience.

Evidence

/blogs/guides/how-to-choose-trail-running-shoes

no author, no date, 0 links to /products/

Why it matters
Readers — and Google — trust advice more when it is clear who gives it and when. Product links turn a reader into a buyer.
Cause
The blog post section of the theme has the author and the date switched off.
How to fix

In the theme or site editorTexts and media

  1. In the theme editor, open a blog post and switch on the author and the date in the blog post section (Dawn: “Show author”, “Show date”).
  2. Start each guide with the author’s role (“fitter at Pearl District”), and link each recommended model to its product page.
Who · effort
Store admin + staff · Low effort, 2 hours
How to verify
Each guide shows author and date in the HTML and links to at least 2 live products.
Show 3 affected sampled guides
  • /blogs/guides/how-to-choose-trail-running-shoes
  • /blogs/guides/best-walking-shoes-for-standing-all-day
  • /blogs/guides/wide-feet-running-shoe-guide

Checked and fine

  • Product descriptions are written by staff who fit shoes: specific, and not copied from manufacturers.
  • Free gait analysis and free 30-day returns are rare, strong selling points (they only need to be visible as text, see F14).

07 · Competitors & search results

Who shows up when shoppers search?

Checked: 4 buying searches run through a web search tool on July 21, 2026, and one comparable page of each of the two local competitors that appeared most often. Web search results are not personalised Google rankings; positions vary by place, device and person.

Search snapshot

Buying searches (4) run through a web search tool on July 21, 2026. The number is the position among the results the tool returned.

Where you and the compared competitors appeared in the web search results
SearchYourosecity-runners.examplestride-and-co.example
“trail running shoes portland”not among the results#2#5
“wide fit walking shoes portland”not among the resultsnot among the results#3
“running shoe store portland”#6#1#3
“shoe fitting portland”not among the results#2#4

Two national chains and one marketplace also appeared; we left them out because they are not comparable to a two-store local retailer.

Competitors compared

Competitors compared
CompetitorAppeared inCompared pageYour page
rosecity-runners.example3 of 4 searcheshttps://rosecity-runners.example/collections/trail/collections/trail-running
stride-and-co.example4 of 4 searcheshttps://stride-and-co.example/shop/wide-fit/collections/wide-fit

One comparable page per competitor

Compared: your /collections/trail-running, /collections/wide-fit with https://rosecity-runners.example/collections/trail, https://stride-and-co.example/shop/wide-fit, as publicly visible on July 21, 2026. We did not measure their traffic or backlinks.

Your pages compared with one comparable page of each competitor
SignalYourosecity-runners.examplestride-and-co.example
Buying advice on the category page (words, without menus and product tiles)38 / 0310450
Product markup with price and availabilityNo (F05)YesYes
Rating markupNoYesNo
City named in title or H1NoYesNo
Named fitter or author on the pageNoYesYes
H01mediumHypothesis1 of 1 category page

Wide-fit shoes could be your quickest content win

Your wide-fit page lists 46 models and has no text. For “wide fit walking shoes portland”, our web search returned a competitor’s wide-fit page that lists 27 models with a 450-word fitting guide — and not yours.

Evidence

/collections/wide-fit · 46 products, 0 words of text

stride-and-co.example/shop/wide-fit · 3rd result

27 products, 450 words

Why it matters
You stock more wide-fit models than the competitor page that appears, and you offer free fitting — but nothing on your page says so.
Cause
Not confirmed yet: how many people search for wide-fit shoes in Portland is unknown without Search Console or keyword data.
How to fix

Texts and mediaOwner's decision

  1. Write the wide-fit page first when you do F11.
  2. Give our service account read access in Search Console, so the recheck can measure impressions for wide-fit searches.
Who · effort
Owner + staff · Medium effort, 3–4 hours
How to confirm
Search Console impressions for “wide fit” searches before the change and 60 days after it.
Show 1 affected URL
  • /collections/wide-fit

08 · Structured data

Price, stock and store details in search results

Checked: 61 JSON-LD blocks on 26 pages, validated against schema.org and Google’s documented requirements and compared with what the page shows.

F05mediumConfirmed issue8 of 8 sampled product pages

Product markup has no price or availability

All 8 sampled product pages have Product markup (JSON-LD) with name, image and brand, but no offers: no price, currency or availability. The price is visible on the page, just not in the markup. Both competitor pages we compared include it.

Evidence

/products/trailpace-gtx-womens

"@type": "Product", "name": "TrailPace GTX Women’s", "brand": …

"offers": — missing

Why it matters
Without offers, Google cannot show price and stock for your products in search results or use the pages for merchant listings. Shoppers compare prices before they click.
Cause
The theme outputs a minimal Product block, and the SEO app’s product markup is switched off.
How to fix

Needs a developer

  1. Hand the steps below to your developer; afterwards test 3 product pages in Google’s Rich Results Test.

For the developer

  1. Replace the theme’s minimal Product block with Shopify’s own structured_data output (see the code): it builds a ProductGroup with an Offer — price, currency, availability, url — for every variant.
  2. Keep the values identical to what the page shows, and remove any second product markup that an app adds.
<script type="application/ld+json">
  {{ product | structured_data }}
</script>

Official guide: Shopify: structured_data filter

Who · effort
Developer · Low effort, 2–4 hours
How to verify
The Rich Results Test on 3 product URLs shows valid product snippets and merchant listings with the visible price.
Show 8 affected sampled product pages
  • /products/trailpace-gtx-womens
  • /products/trailpace-gtx-womens-moss
  • /products/cityglide-2-mens
  • /products/cityglide-2-mens-graphite
  • /products/ridgeline-boot-womens
  • /products/ridgeline-boot-womens-oak
  • /products/harbor-walker-wide
  • /products/tempo-racer-3

Checked and fine

  • 0 parse errors; Organization and BreadcrumbList markup is correct on every page.
  • No markup describes content that visitors cannot see.

09 · Speed

How fast the store feels on a phone

Checked: Lighthouse 13.5 through our measurement helper. Mobile: simulated slow 4G and 4× CPU slowdown, 3 runs per URL (median and range). Desktop: 1 run. These are lab values, not real-visitor data. Real-visitor data for the whole site comes from the Chrome UX Report (phones, 28 days) and is shown separately.

Largest Contentful Paint (LCP): when the main content appears. Median of 3 mobile runs, 1 desktop run. The dashed line marks 2.5 s, Google’s “good” threshold for real-visitor data. Lab data, not real-visitor data.
Lighthouse lab measurements, median of runs (range)
Page · device (runs)PerformanceLCPCLSTBTFCPSpeed Index
Homepagemobile (3) · /5553–584.1 s3.9–4.40.04390 ms2.2 s4.0 s
Homepagedesktop (1) · /911.2 s0.0160 ms0.6 s1.1 s
Product · TrailPace GTXmobile (3) · /products/trailpace-gtx-womens3936–425.8 s5.5–6.30.02610 ms2.6 s5.1 s

Lighthouse 13.5.0, simulated mobile (median of 3 runs, range in brackets) and desktop. Lab data, not real-visitor data.

Real visitors (Chrome UX Report)

Real Chrome users, 2026-06-22 – 2026-07-19. Shown next to the lab data; not part of the measured index.
Metric75th percentileRatingGood / needs improvement / poor
LCP3.1 sneeds improvement68 % / 22 % / 10 %
INPnot enough data——
CLS0.05good88 % / 8 % / 4 %
FCP2.0 sneeds improvement61 % / 27 % / 12 %
TTFB900 msneeds improvement59 % / 30 % / 11 %

Real Chrome users, 2026-06-22 – 2026-07-19. Shown next to the lab data; not part of the measured index. Real visitors see the main content sooner than the lab’s slow-phone profile, but a quarter of phone visits still wait more than 3 seconds.

F03highConfirmed issue2 of 2 speed-tested pages

Product pages take 5.8 s to show the main photo on a phone

On /products/trailpace-gtx-womens the main photo — the largest element on screen — appears after 5.8 s (median of 3 mobile lab runs, 5.5–6.3 s). The homepage needs 4.1 s on mobile and 1.2 s on desktop. All 8 sampled product pages use the same template, image format and scripts, but only this one was speed-tested.

Evidence

LH-product-mobile-1…3 · /products/trailpace-gtx-womens · mobile

LCP 5.8 s (5.5–6.3) · Performance 39 · TBT 610 ms

LCP element: <img loading="lazy" src="…/trailpace-moss.png"> 1.9 MB, 3000×3000 px

Why it matters
Shoe shoppers compare several stores on their phone. A product photo that takes almost 6 seconds to appear makes the page feel broken, and many people go back to the results. Page experience is one of Google’s signals, although speed alone never guarantees rankings.
Cause
The main product photo is a 1.9 MB PNG (3000×3000 px) sent to 412 px wide screens, with no srcset. It is marked loading="lazy", so the browser deliberately fetches it late. Three font files block the first render, and the reviews widget adds 410 KB of JavaScript before the page is usable.
How to fix

In the theme or site editorNeeds a developer

  1. Online Store → Themes → Edit theme → App embeds: switch off the app embeds you no longer use — each one adds scripts to every page.

For the developer

  1. Remove loading="lazy" from the first product image and give it fetchpriority="high" (see the code).
  2. Output product images with Shopify’s image_tag at widths 480, 960 and 1600 px: the CDN then serves WebP/AVIF at the right size instead of the 1.9 MB PNG.
  3. Preload the one font weight used above the fold, set font-display: swap and drop the two unused weights.
  4. Load the reviews widget when the reviews section scrolls into view.
{{ product.featured_media | image_url: width: 1600 | image_tag:
    loading: 'eager', fetchpriority: 'high',
    widths: '480, 960, 1600', sizes: '(min-width: 990px) 50vw, 100vw' }}

Official guide: Shopify Help: App embeds · Shopify: image_tag filter

Who · effort
Store admin (app embeds) + front-end developer · Medium effort, 1–2 days
How to verify
Repeat 3 mobile Lighthouse runs on the same URL with the same settings: median LCP under 3.5 s as a first milestone, then 2.5 s or less; the product image transfers under 200 KB.
Show 2 affected speed-tested pages
  • /products/trailpace-gtx-womens
  • /

Checked and fine

  • Desktop homepage: Performance 91, LCP 1.2 s.
  • Layout is stable: CLS 0.04 or lower on both pages.

10 · Mobile & accessibility

Can everyone use the store on a phone?

Checked: 5 pages rendered in a real browser as a phone (Pixel 7, 412 × 915 px, mobile Chrome) and on desktop (1366 × 900 px) — readability, sideways scrolling, tap targets and accessible names; accessible names also in the HTML of all 26 pages.

Screenshot B-S10-mobile, simplified for this fictional store: /products/trailpace-gtx-womens at 412 px. A real report shows the actual screenshots of your pages.
F07mediumConfirmed issue1 of 5 browser-rendered pages

The size picker and size chart don’t fit a phone screen

On the product page on a phone (412 px wide), size buttons are 30 × 30 px with 4 px gaps, and the size chart opens as a 900 px wide image that makes the whole page scroll sideways. All 8 sampled product pages use this template; we rendered one of them.

Evidence

/products/trailpace-gtx-womens · 412 × 915

size buttons 30 × 30 px, gap 4 px

document width 900 px > viewport 412 px (size chart image)

Why it matters
Choosing a size is the step right before “Add to cart”. Small targets cause wrong taps, and a chart you have to pan across is hard to read — at the exact moment a shopper decides.
Cause
The theme’s size buttons have a fixed 30 px size, and the size chart is an image inserted into the product description.
How to fix

Needs a developer

Make size buttons at least 44 × 44 px with 8 px spacing. Replace the size chart image with an HTML table that fits a 360 px screen (it may scroll inside its own box). Give the chart a text caption so screen readers can use it (see F14).

For the developer

  1. Make size buttons at least 44 × 44 px with 8 px spacing.
  2. Replace the size chart image with an HTML table that fits a 360 px screen (it may scroll inside its own box).
  3. Give the chart a text caption so screen readers can use it (see F14).
Who · effort
Front-end developer · Medium effort, 3–5 hours
How to verify
At 412 px width the product page does not scroll sideways, and each size button measures at least 44 × 44 px in the browser’s developer tools.
Show 1 affected URL
  • /products/trailpace-gtx-womens
  • 7 more sampled product pages use the same template (not rendered)
F09lowConfirmed issue26 of 26 sample pages (one shared header)

Cart and search icons have no names for screen readers

The search, account and cart buttons in the header are icons without text or aria-label on all 26 pages (one shared header). Lighthouse’s accessibility audit of the homepage flags the same 3 buttons.

Evidence

<button class="header__icon"><svg …></svg></button> × 3

“Buttons do not have an accessible name” (3)

Why it matters
Screen-reader users hear “button, button, button” and cannot find the cart. It is quick to fix and part of basic accessibility.
Cause
The theme header renders SVG icons inside empty buttons.
How to fix

Needs a developer

In the header section (sections/header.liquid in Dawn), give the three icon buttons names: aria-label "Search", "Account" and "Cart (n items)" — see the code.

For the developer

  1. In the header section (sections/header.liquid in Dawn), give the three icon buttons names: aria-label "Search", "Account" and "Cart (n items)" — see the code.
<button class="header__icon" aria-label="Search">…</button>
<a href="{{ routes.account_url }}" class="header__icon" aria-label="Account">…</a>
<a href="{{ routes.cart_url }}" class="header__icon" aria-label="Cart ({{ cart.item_count }} items)">…</a>
Who · effort
Developer · Low effort, 30 minutes
How to verify
Lighthouse accessibility on the homepage no longer lists “Buttons do not have an accessible name”.
Show 1 affected URL
  • /
  • and the other 25 sample pages (same header)

Checked and fine

  • Text is readable without zooming on all 5 rendered pages, and the main button is visible on the first screen.
Screenshots
A real report shows the actual browser screenshots of your pages next to the findings they illustrate; this fictional store has none.

11 · Images

Are photos helping or slowing you down?

Checked: 214 images on 26 pages — alt text, format, file size and displayed size (from the browser renders and Lighthouse).

Checked and fine

  • 213 of 214 informative images have alt text, and product photos have descriptive alts. Decorative icons correctly use empty alt.
  • Category thumbnails are served as WebP at the size they are displayed.
Covered elsewhere
The 1.9 MB product photo is part of F03; the size chart image without alt text is part of F14.

12 · Internal links

Checked: every internal link on the 26 pages, cross-referenced with the sitemap. Link targets missing from the sitemap were requested within the page budget (3 of 7).

F06mediumConfirmed issue5 of 26 sample pages

The homepage and guides link to products that no longer exist

5 of 26 pages link to 7 product URLs that are not in the sitemap. We requested 3 of them (page budget), and all 3 return 404. The other 4 follow the same pattern (old model years) but were not requested.

Evidence

/blogs/guides/how-to-choose-trail-running-shoes

<a href="/products/trailpace-gtx-2023"> → 404 (E-L01)

<a href="/products/cityglide-1-mens"> → 404 (E-L02)

Why it matters
Visitors who follow the advice in your guides land on an error page instead of a shoe they can buy, and the link passes no value to a live product.
Cause
Old models were deleted without redirects, and the links in guides and homepage tiles were never updated.
How to fix

In the admin, no codeTexts and media

  1. Content → Menus → View URL redirects → Create URL redirect: from /products/trailpace-gtx-2023 to /products/trailpace-gtx, and the same for each removed model (or to its category). The old URLs already return 404, so the redirects work at once.
  2. Update the 7 links: in the guides (Content → Blog posts) and in the homepage tiles (Online Store → Themes → Edit theme).
  3. In future, create the redirect as soon as you delete a product.

Official guide: Shopify Help: URL redirects

Who · effort
Store admin · Low effort, 1–2 hours
How to verify
All 7 old URLs answer 301 to a live page, and the 5 pages link directly to pages that return 200.
Show 5 affected sample pages
  • /
  • /collections/trail-running
  • /collections/sale
  • /blogs/guides/how-to-choose-trail-running-shoes
  • /blogs/guides/best-walking-shoes-for-standing-all-day
  • 7 link targets, e.g. /products/trailpace-gtx-2023, /products/cityglide-1-mens

Checked and fine

  • Every category is linked from the main menu, and every sampled product is reachable in 2 clicks from the homepage.
  • Breadcrumbs link products back to their category.

13 · Local SEO

Can nearby shoppers find your two stores?

Checked: the store pages, name/address/phone consistency on all 26 pages and local structured data. Your Google Business Profile was not checked (it needs owner access).

F02criticalConfirmed issue3 of 3 store pages

Your two store pages ask Google not to index them

The store index and both store pages (Pearl District and Hawthorne) contain <meta name="robots" content="noindex">, and none of them is in sitemap.xml. These are the only pages with your addresses, opening hours and the free gait-analysis fitting.

Evidence

/pages/stores/pearl-district

<meta name="robots" content="noindex">

0 of 412 entries start with /pages/stores

Why it matters
People search for “running shoe store Portland” or “shoe fitting Hawthorne” when they want to visit a store today. These pages answer that search, and they ask Google to leave them out. Google respects noindex, so the pages cannot appear in results while the tag is there.
Cause
The three store pages have Shopify’s seo.hidden metafield set to 1 — most likely left over from before the stores opened. With it, Shopify adds noindex and leaves the pages out of sitemap.xml.
How to fix

In the admin, no code

  1. Online Store → Pages → open each store page → Metafields: clear the “seo.hidden” value → Save. Shopify then removes the noindex tag and puts the pages back into sitemap.xml by itself.
  2. Keep the footer links to the three store pages.
  3. Request indexing for both store pages in Search Console.

Official guide: Shopify Help: Hide a page from search engines

Who · effort
Store admin · Low effort, 15–30 minutes
How to verify
The page source of /pages/stores/pearl-district contains no noindex. Search Console → URL Inspection shows “Indexing allowed: Yes”.
Show 3 affected store pages
  • /pages/stores
  • /pages/stores/pearl-district
  • /pages/stores/hawthorne
F13lowRecommendation2 of 2 store pages

Store pages lack store markup and text opening hours

Neither store page has ShoeStore (LocalBusiness) markup, and opening hours are shown only in an image. Name, address and phone match the footer on all 26 pages.

Evidence

/pages/stores/pearl-district

<img src="…/pearl-hours.png" alt=""> ← opening hours

JSON-LD: Organization only

Why it matters
Once F02 is fixed, markup and text hours help Google show correct details for your stores. Text hours are also readable for screen readers and AI assistants.
Cause
The opening hours were uploaded as a designed image.
How to fix

Texts and mediaNeeds a developer

  1. Online Store → Pages → each store page: write the opening hours as text instead of the image.
  2. Keep your Google Business Profile hours the same (we could not check them).

For the developer

  1. Add ShoeStore markup to the store page template with address, phone, geo coordinates and opening hours — identical to the visible text.

Official guide: Shopify Help: Add and edit pages · Google: Local business structured data

Who · effort
Developer + owner (confirm hours) · Low effort, 1–2 hours
How to verify
The Schema Markup Validator shows a valid ShoeStore with the same hours as the page.
Show 2 affected store pages
  • /pages/stores/pearl-district
  • /pages/stores/hawthorne

Checked and fine

  • Name, address and phone are identical in the footer of all 26 pages.

14 · GEO · AEO · LLMO · AI SEO

Can AI assistants find and quote your answers?

Checked: robots.txt rules for AI crawlers, whether key facts (returns, sizing, shipping, store hours) exist as crawlable text, llms.txt (optional, not used by Google Search) agent readiness (can an AI agent read the page and find the buttons it needs) and Bing’s page checks on the same 26 pages (the rules of Bing Webmaster Tools; Microsoft Copilot answers from the Bing index). Nobody can guarantee AI citations.

ChatGPT, Perplexity, Gemini, Microsoft Copilot and Google's AI Overviews and AI Mode answer the question themselves and name a few sources. They can only pick pages they are allowed to read, can understand without guesswork and have reason to trust. Google's requirements are those of Search — the page must be indexed and eligible to show a snippet — plus one condition: the site must be included in Search generative AI features in Search Console. So this chapter is not a separate trick: it shows what stands between your site and AI answers. Nobody can guarantee being cited.

GEO (generative engine optimization) is also called AEO, LLMO or AI SEO: it is the same work.

AI search readiness

5 of 13

checks passed

Partly: 5 · not passed: 3

  • Access for AI crawlers2/2
  • Text without JavaScript1/1
  • Your business as an entity1/2
  • Answers on key pages1/8

A count of checks, not a score: it does not predict citations or traffic, and an untested check is not counted as a failure.

Visits from AI assistants

We cannot see where your visitors come from: connect Google Analytics in your dashboard, and the next report will show how many come from ChatGPT, Gemini, Perplexity and Copilot.

What we checked

  • Search and answer crawlers may read the pages

    Passed

    Allowed for all 9 crawlers that feed AI answers

  • No snippet limits that keep pages out of AI answers

    Passed

    No nosnippet, max-snippet:0 or noarchive on 23 pages

  • Main text is in the HTML, without JavaScript

    Passed

    The main text is in the server HTML on 5 of 5 pages

  • Organization or LocalBusiness markup

    Passed

    Organization “Northwind Shoes”

  • Official profiles linked in the markup (sameAs)

    Not passed

    The markup lists no profiles

    What to do: Online Store → Themes → Customize → Theme settings → Social media: add the Instagram and Facebook profile links; Dawn adds them to the Organization markup as sameAs.

AI crawlers in robots.txt

Search and answers

Search and answers
CrawlerWhat it is forAccess
GooglebotGoogleSearch index behind AI answersAllowed
BingbotBing / CopilotSearch index behind AI answersAllowed
OAI-SearchBotOpenAISearch index behind AI answersAllowed
PerplexityBotPerplexitySearch index behind AI answersAllowed
Claude-SearchBotAnthropicSearch index behind AI answersAllowed
ChatGPT-UserOpenAIOpens a page when a person asksAllowed
Claude-UserAnthropicOpens a page when a person asksAllowed
Perplexity-UserPerplexityOpens a page when a person asksAllowed
Google-ExtendedGoogleGemini app answers and training (not Google Search)Allowed

Model training: GPTBot, ClaudeBot, Applebot-Extended, CCBot — allowed. Blocking training crawlers is your choice: it does not remove the site from AI answers.

Your business as an entity

AI answers name businesses they can identify with confidence. Markup that states who you are and links to your official profiles makes that easier.

Organization markup
Organization “Northwind Shoes”
Official profiles (sameAs)
The markup lists no profiles
Wikidata
No item names this site as its official website (common for small and local businesses)

Can AI answer your customers' questions from your pages?

We read the key pages the way an AI answer would use them: is there a clear, specific answer a system could quote, and is it clear who stands behind it?

  • A clear answer at the top of the page

    PartlyPages: 2 of 3

    Product pages open with what the shoe is for; category pages start with the product grid and no sentence on who the range suits.

    «Running, walking and trail shoes, fitted by people who run.»

  • Headings that match customers' questions

    PartlyPages: 1 of 2

    Only the guides use the questions customers ask; the wide-fit category answers none of them in text.

    «H2: How do I know a trail shoe fits?»

  • Specific facts: prices, terms, delivery, contacts

    PartlyPages: 1 of 3

    Prices and delivery are in text; the returns policy and the sizes are in a PDF and an image (F14).

    «$165 · Free shipping over $75»

  • Numbers, sources and quotes

    PartlyPages: 0 of 1

    The guides give specific advice from fitting experience, but no measurements or named sources behind it.

    «Most of our customers need half a size up for downhill running.»

  • Visible, current dates

    Not passedPages: 0 of 2

    The guides show no publish or update date (F12).

    «No date on the page»

  • A named author or expert

    Not passedPages: 0 of 2

    No author or fitter is named on the guides (F12).

    «No author on the page»

  • About, contact and policy pages

    PassedPages: 2 of 2

    About, store addresses, phone and returns are linked from every page.

    «Family-owned, with two fitting stores in Portland.»

  • Lists and tables for specs and comparisons

    PartlyPages: 1 of 2

    Product specs are a list; the size chart, the table customers need most, is an image (F07, F14).

    «Specs: weight 290 g · drop 6 mm · waterproof membrane»

Where AI answers take their sources

Where AI answers take their sources
AI answers inSources come fromCrawler to allow
Google AI Overviews and AI ModeGoogle Search indexGooglebot
ChatGPT searchOpenAI's search index and partnersOAI-SearchBot; ChatGPT-User when a person asks
Microsoft CopilotBing indexBingbot
PerplexityIts own index and live fetchesPerplexityBot; Perplexity-User
ClaudeWeb search and live fetchesClaude-SearchBot; Claude-User
Gemini appGoogle Search indexGooglebot; the Google-Extended token

AI crawlers can read the whole store, and the brand is marked up on every page. What holds AI answers back is the format: returns and sizes sit in a PDF and an image, and the guides do not say who wrote them or when.

Agent readiness

Lighthouse Agentic Browsing
3/3
AI crawlers in robots.txt
All AI crawlers allowed; no Content-Signal line.
llms.txt
Not present (optional; Google Search does not use it).
Markdown version
Pages are served as HTML only (optional).
Discovery files
No /.well-known discovery files (optional; the store has no public API).
Tasks without login
An agent can find products, prices, stores and opening hours; the size chart is an image it cannot read (F14).

Optional files are recommendations, not defects.

Bing and Copilot

Microsoft Copilot and Bing's AI answers draw on the Bing index, and Bing Webmaster Tools checks pages with its own rules, some of them stricter than Google's. We ran Bing's page checks on the same 26 pages; the check names are the ones Bing Webmaster Tools shows, so you can find them in your account.

  • HTML size is too long

    page HTML over 125 KB

    26 of 26 pages · Matters for Google: No, Bing only

    Online Store → Themes → Edit theme → App embeds: switch off unused embeds; your developer removes leftover app code and renders the mobile menu and filters only when they are opened.

  • Meta description missing

    no meta description

    4 of 26 pages · Matters for Google: Partly

    See F08: write the meta descriptions of the 4 products.

  • Title too long

    over 70 characters

    3 of 26 pages · Matters for Google: No, Bing only

    Content → Blog posts → the guide → Search engine listing: shorten the page title to about 70 characters.

  • HTTP 4xx/5xx errors

    addresses answering with an error

    3 of 32 addresses · Matters for Google: Yes

    See F06: redirect the removed products and update the links.

No problems found: Title missing, Meta description too long, Meta description too short, H1 tag missing, More than one H1 tag, Image missing alt attribute, Meta language missing, Temporary redirect (302/307), Meta refresh, NOARCHIVE / NOCACHE.

Bingbot in robots.txt
Allowed
Crawl-delay (Bing follows it)
Not set
Sitemap in robots.txt
Present
Bing Webmaster Tools verification tag
Not found on the homepage (the site may still be verified by DNS, a file or a Search Console import)

Like most Shopify stores, every page is over Bing’s 125 KB HTML limit; Google reads up to 2 MB, so this matters for Bing and Copilot only. The 404s and the missing descriptions are fixed with F06 and F08.

F14lowRecommendation2 of 2 pages

Returns and sizing facts exist only as a PDF and an image

/pages/returns links to a PDF instead of stating the policy, and /pages/size-guide shows the chart as one image with empty alt text. robots.txt allows all AI crawlers, so the problem is the format, not access.

Evidence

/pages/returns · policy only in /files/returns-2026.pdf

/pages/size-guide · <img src="…/size-chart.png" alt="">

Why it matters
AI assistants and Google answer questions like “Northwind Shoes returns policy” from page text. Facts locked in a PDF or an image are harder to find and quote — and they are the questions your team answers every day.
Cause
The content was uploaded from print material.
How to fix

Texts and mediaIn the admin, no code

  1. Online Store → Pages → Returns: state the returns policy (30 days, free, in store or by post) as text instead of the PDF link.
  2. Turn the size chart into an HTML table (shared with F07).
  3. Add a short FAQ with the 5 questions customers ask most.

Official guide: Shopify Help: Add and edit pages

Who · effort
Store admin · Low effort, 1–2 hours
How to verify
Both pages contain the facts as text in the HTML source.
Show 2 affected pages
  • /pages/returns
  • /pages/size-guide

Checked and fine

  • robots.txt allows all AI crawlers.
  • Shipping costs and delivery times are stated as text on every product page.

15 · E-commerce, languages & backlinks

What else we looked at

Checked: Checked where it applies to your site, within the limits of the data sources we can use without your accounts.

E-commerce
Applies. Product markup is covered in Structured data (F05). Google Merchant Center was not checked (it needs account access).
Languages
Not applicable: the site is in English only, so no hreflang is needed.
Backlinks
The domain is in Common Crawl’s public web graph (release cc-main-2026-jan-feb-mar) with a harmonic-centrality rank of about 8.4 million. The graph gives a position, not a list or count of linking sites; paid backlink indexes (Moz, Ahrefs) were not used.

16 · 30/60/90-day plan

A realistic order of work

Grouped so each step unblocks the next. Plan the recheck around day 60, after the speed work is live.
  1. Days 1–30

    Quick fixes that unblock Google

    • F01 canonical tags
    • F02 store pages: remove noindex
    • F06 redirects for removed products
    • F10 sitemap without parameter URLs
    • F09 names for icon buttons
    • Start F03: eager main image, fetchpriority

    Depends on: Store admin access; about 1 developer day

  2. Days 31–60

    Speed and richer results

    • Finish F03: image CDN, srcset, fonts, lazy reviews widget
    • F05 product offers markup
    • F07 size picker and HTML size chart
    • F13 store markup and text opening hours
    • F08 four meta descriptions
    • F04 colour in the titles of colour pages

    Depends on: F13 needs confirmed opening hours from the owner

  3. Days 61–90

    Content that wins buying searches

    • F11 advice on the top 3 categories
    • H01 wide-fit page + Search Console
    • F12 authors, dates and product links in guides
    • F14 returns and size facts as text
    • F04 merge colour variants
    • Recheck on the same 26 pages

    Depends on: Staff writing time; the F04 merge needs a redirect plan

This is a sequence of work, not a promise of search results.

17 · Methodology & limits

What we checked — and what we didn’t

Evidence collected on July 21, 2026 with a polite crawler (one request at a time, at least 1.5 s apart, robots.txt respected), a real browser, Lighthouse and a web search tool: 1,164 requests in total (budget 1,500). Every claim was checked against the saved raw evidence before publishing.
Number of unique URLs or searches per type of check. A URL listed in the sitemap is not a checked URL.

How the audit was run

  1. Read robots.txt, the sitemap and the homepage; detect the business type (local retailer with an online store).
  2. Choose the sample before crawling and record it: homepage, 8 of 14 categories, 8 products (including colour pages), store pages, 3 guides, returns, size guide and about.
  3. Collect evidence once, one request at a time, at least 1.5 s apart, respecting robots.txt; render 5 pages in a real browser and run Lighthouse.
  4. Run 4 buying searches through a web search tool and fetch one comparable page of each of the two local competitors that appeared most often.
  5. Specialist reviews (technical, content, structured data, speed, mobile, local, AI search, agent readiness …) work in parallel on the same saved evidence, with no extra requests to your site.
  6. Every claim is checked against the raw evidence before it goes into the report; unsupported claims are dropped or marked as a hypothesis.

How the sample was chosen

Homepage; 8 of the 14 categories (best sellers, new arrivals, sale, wide fit); 8 products, including 3 colour pages, to test duplication; the store index and both store pages; 3 of the 5 guides; returns, size guide and about — 26 pages. Plus 4 probes (http, www, trailing slash, a random URL) and 3 link targets that were missing from the sitemap.

Lab vs field. Speed values in the index are lab measurements on a simulated slow phone. Real-visitor data from the Chrome UX Report is shown separately when Chrome has enough traffic for it.

Tools and versions: Lighthouse 13.5.0 · Chrome UX Report · Common Crawl

Specialist checks

Which specialist checks ran, and why some could not
CheckStatusNotes
Technical SEOExecuted26 pages, 4 probes, robots.txt, canonicals, headers
Content & E-E-A-TExecuted26 pages read, 11 reviewed in depth (E-E-A-T, intent)
Search experience & competitorsExecuted4 web searches, 2 comparable competitor pages
Structured dataExecuted61 JSON-LD blocks on 26 pages
SitemapsExecuted412 URLs parsed, 26 matched to fetched pages
SpeedExecuted7 Lighthouse runs interpreted, no estimates
Mobile & visual reviewExecuted10 renders (5 pages × mobile and desktop)
ImagesExecuted214 images, through the browser and Lighthouse data
AI search (GEO)ExecutedAI crawler access, key facts as text, llms.txt
Agent readinessExecutedAgent readiness: accessibility tree, robots rules for AI agents, discovery files
Local SEOExecutedStore pages and NAP on 26 pages; Business Profile not accessible
E commerceExecutedProduct markup on 8 product pages; Merchant Center not accessible
BacklinksExecutedCommon Crawl web graph only (a rank, no link list); Moz and Bing not configured
Google data (CrUX, Search Console)ExecutedCrUX field data for the origin (phone); no Search Console access
Language versions (hreflang)Not applicableSingle language (English)
Change trackingNot applicableFirst audit — this report is the baseline

Not checked, and why

Google’s actual index, searches and clicks
No Search Console access. HTTP 200 does not prove a page is indexed.
Real-visitor speed of single pages and INP
Chrome has enough traffic for origin-wide phone data only (shown in Speed); there is too little for single pages and for INP.
Exact Google positions and search volumes
Web search results are not personalised Google rankings, and we used no paid keyword data.
Google Business Profile and reviews on other sites
They need owner access.
Backlink quality
Only the public Common Crawl web graph: a rank, no list of linking sites and no quality metrics.
386 of 412 sitemap URLs
Listed, not fetched. Findings describe the sample and are not extrapolated to the whole site.
Checkout and customer account
Out of scope: we never log in, submit forms or place orders.

Want to know if the fixes worked?

Recheck this website after implementing the changes. Recommended: in 30–60 days.

Recheck website (opens the sample comparison)