Fix the buying promises.
Clear the path to purchase.

UNITED50 storefront audit

Ecommerce SEO and AI search readiness

Private decision report | u50.com | Evidence collected 11 October 2026

Executive assessment

Fix the store's buying promises and shopping dead ends first. The public storefront has conflicting shipping and return rules, homepage promotions leading to wholly sold-out products, and incorrect size guidance. These issues can confuse shoppers before additional SEO or AI-search work has a chance to help.

The technical foundation is largely present. All 103 sitemap HTML pages returned 200 in the recorded crawl, all 56 product pages supplied parseable Product and Offer data, canonical tags were present, and the reviewed-product rating is correctly linked in the rendered sample. The next search priorities are accurate shared facts, useful category guidance, consistent brand identity and clearer product paths.

Decisions to make first

P1 - Approve the actual shipping rate, free-shipping threshold, destinations and timing; reconcile the return fee and initiation route. Findings U50 01 and U50 02.

P1 - Confirm inventory and replace unavailable homepage destinations with useful in-stock or restock paths. Finding U50 03.

P1 - Validate size-guide mappings and measurement labels against product specifications. Finding U50 04.

P2 - Repair inquiry, product-discovery and cart clarity, then improve identity, category content, product data, stale URLs and claim evidence. Findings U50 05 through U50 11 and U50 13.

P3 - Measure accessibility, mobile usability and performance before prescribing broad changes. Finding U50 12.

Status and limits

This is the completed public-storefront portion of the audit. A reversible cart test passed add, quantity update and zero-quantity removal, and the cart was reset to empty. Shopify admin access and the private configuration audit remain pending. Checkout, payment, customer accounts and real returns were not tested. There is no current ranking, conversion, Core Web Vitals, AI-visibility or ROI result.

All owners and effort estimates are proposed. No implementation dates are commitments, and no production changes are authorized by this report.

Scope and evidence

Sitemap contained 104 URLs: 103 HTML pages plus the agent-discovery document. HTML inventory: 56 products, 25 collections, 20 standalone pages, homepage and one blog index.

All 103 HTML pages returned HTTP 200. The blog request had a transient client/network failure during the first pass and succeeded on a later serial read. No storefront HTTP 429 was observed in successful recorded reads. This does not verify Google/OpenAI provider access, or prove all intermittent throttling is fixed.

Desktop cloud Chrome interaction at a 1165 × 757 CSS-pixel viewport covered homepage, women hub, leggings collection, Renew collection, several product pages, direct variant landing, search, empty and populated test cart, size guide, shipping/returns and wholesale.

Raw HTML, response headers, robots, sitemap, canonical and schema were inspected. Structured data was JSON-parsed; Google Rich Results Test/URL Inspection were not run.

No mobile-device/emulated-mobile journey, checkout/payment test, customer account test, actual return initiation, Search Console, GA4, Merchant Center, private server logs or first-party conversion data was verified. Product and policy causes inside Shopify remain unconfirmed.

Mobile emulation could not be completed: the browser API had no viewport/device-emulation capability, and native Chrome displayed an organization restriction preventing DevTools on this page. That restriction was respected; it is a test-environment limitation, not a storefront defect.

PageSpeed Insights public mobile API returned HTTP 429. No valid Lighthouse score or Core Web Vitals field measurement was obtained. That API failure is not evidence of a U50 speed defect.

Crawl inventory

Evidence standard: “Verified” describes an observed page, interaction or machine-readable response. Recommendations describe expected mechanisms, not measured business uplift. Index eligibility is distinct from confirmed indexing; parseable schema is distinct from Google rich-result validation.

Priority standard: P1 is direct customer-promise or purchase-discovery risk; P2 is meaningful content, discovery or usability work; P3 is optimization or further verification. Effort assumes the underlying business rule is known.

Customer promises

U50 01 Shipping promises conflict

Priority: P1. Status: verified defect. Confidence: high. Effort: small after an operations decision. Suggested owner: ecommerce operations + theme developer.

Evidence:

https://u50.com/policies/shipping-policy gives a $10 standard charge below $150, then states a $17 FedEx two-day standard rate in its delivery section. It also leaves a carrier-selection note referencing Left Brain.

The policy grants free standard shipping at $150 or more; the announcement bar and PDP summary say orders must be over $150. Exactly-$150 carts are ambiguous.

https://u50.com/pages/shipping supplies 2–4 business days of processing plus 2–3 after dispatch, but does not explain the conflicting cost claims. Its visible H1 is “Returns & Exchanges,” despite the shipping URL/title.

https://u50.com/pages/international-orders says international shipping is not offered; the formal shipping policy explicitly includes Canada. Clarify whether that page means countries beyond Canada.

Why it matters: shoppers and shopping agents can reach different versions and cannot reliably determine total cost, timing or destinations. This is a factual consistency issue; actual checkout charges were not tested.

Action: designate one approved shipping rule set, confirm configured zones/rates/free threshold/processing times, then synchronize policy, support page, banner, PDP, FAQ and feed/channel data. Remove draft wording and correct the H1.

Retest: authorized test carts at $149, $150 and $151 after discounts in supported US/Canada destinations agree with every displayed promise; processing and transit time are clearly distinguished. No purchase needed for this validation.

U50 02 Return rules and the portal disagree

Priority: P1. Status: verified defect. Confidence: high. Effort: small. Suggested owner: customer service/operations + developer.

Evidence:

https://u50.com/products/victoria-bodysuit and the sampled PDP template promise free returns within 30 days.

https://u50.com/pages/returns-exchanges permits deduction of return-shipping cost unless otherwise specified and links its return-portal instruction to https://u50.com/, not a return workflow.

https://u50.com/policies/refund-policy instead tells buyers to begin from their purchase-confirmation email and omits the fee-deduction bullet. The two policy locations are not synchronized.

Why it matters: unexpected deductions undermine purchase confidence and a broken initiation path increases support workload. Whether the PDP free-return promise overrides that conditional deduction, and which workflow is actually configured, remain operations/admin questions.

Action: confirm intended return fees, window, exclusions and initiation route; publish one consistent summary with a working deep link or clear email-based steps; link the PDP summary to the full applicable policy.

Retest: an authorized existing/test order can reach the correct return-entry screen; customer-service instructions, both policy locations and PDP fees agree. Do not create or submit a real return merely to test.

Merchandising and fit

U50 03 Promotions lead to unavailable products

Priority: P1. Status: verified storefront state; stock cause unknown. Confidence: high. Effort: small for merchandising; replenishment depends on operations. Suggested owner: merchandising + inventory operations.

Evidence:

Homepage features https://u50.com/products/coastline-short at $85. Its PDP shows a disabled sold-out button; all 32 variant offers in public JSON-LD are OutOfStock.

Homepage's Renew CTA reaches https://u50.com/collections/united-renew, whose only two products are https://u50.com/products/first-light-boardshort and https://u50.com/products/breakwater-boardshort. Each has 24 OutOfStock offers. First Light's UI remained sold out after changing size.

No restock-notification action was visible on the inspected Coastline/First Light PDPs. Collection cards show prices without a sold-out label in the inspected listing representation.

Why it matters: scarce homepage attention is spent on paths that cannot currently convert in the observed US/USD storefront.

Action: verify real inventory/channel availability first; promote in-stock alternatives or clearly present launch/restock status. Add an approved back-in-stock flow when appropriate and mark unavailable product cards. Do not falsify availability or delete a useful temporarily sold-out URL.

Retest: each primary shopping CTA has an available purchase path or an explicit useful restock/alternative path; product UI, Offer data and catalog stock agree.

U50 04 Size guidance contains errors

Priority: P1. Status: verified defect. Confidence: high. Effort: small–medium. Suggested owner: product/fit specialist + theme developer.

Evidence:

On https://u50.com/products/victoria-bodysuit, Size Guide opens the Bottoms tab and a 24-inch inseam row, although a Bodysuits & Dresses tab exists and the product describes a 3.5-inch short.

https://u50.com/pages/size-charts → Men → Tops lists an Inseam row with 31–35.5-inch ranges. This is not a useful measurement label for tops; the intended measurement must be confirmed rather than guessed.

The women's generic Bottoms chart uses one 24-inch inseam across sizes while serving varied shorts/leggings products. Clarify body measurements versus garment-specific measurements.

Why it matters: fit uncertainty can discourage orders or cause avoidable size-related returns. No return-rate lift or reduction has been measured.

Action: correct the product-to-size-guide mapping; validate all measurement labels/ranges with the product team; expose product-specific inseam, fit and model measurements where available; keep body and garment measurements distinct.

Retest: bodysuit, bra, tee, hoodie, short and legging samples open the intended guide; inch/cm conversion and keyboard interaction work; dimensions match approved tech packs.

Inquiry and shopping journeys

U50 05 Wholesale instructions lack the promised form

Priority: P2. Status: verified defect, persists from August and October 7 audits. Confidence: high. Effort: small. Suggested owner: wholesale lead + developer.

Evidence: https://u50.com/pages/wholesale-inquiries asks visitors to complete a form below, but the live page has a PDF download and no wholesale inquiry form. Footer email is available, so this is a broken stated journey, not a claim that all contact is impossible.

Action: add a short accessible inquiry form or revise the instructions to explicitly download, complete and send the PDF to a verified destination. State what happens next and a realistic response window.

Retest: the page presents one unambiguous route; owner receives an approved test inquiry without accidental marketing enrollment. No inquiry was submitted during the audit.

U50 06 Product discovery adds avoidable friction

Priority: P2. Status: verified behavior; conversion recommendation. Confidence: high for observations, unmeasured for outcome. Effort: medium. Suggested owner: merchandising + UX developer.

Evidence:

Homepage's Shop All Women/Men links land on https://u50.com/collections/women and https://u50.com/collections/men, which are subcategory hubs, with zero product links in this run. They do not actually show all products. The hub design itself is not a defect; the promise/path needs clarity.

https://u50.com/pages/seasonal-looks has many outfit captions/images but zero product hyperlinks in source HTML. It lacks an H1.

Search for leggings expands to https://u50.com/search?q=%28leggings*+OR+tights*+OR+joggers*%29&q_original=leggings and returns 53 results, including tees/hoodies. The top four are relevant leggings/capri, so relevance is partially working; broad secondary results and the raw Boolean query in breadcrumb/title reduce clarity.

Category sampling showed no filter/sort controls. On four-product leggings this may be reasonable; prioritize larger ranges and search results rather than adding needless controls everywhere.

Action: offer a true all-products option or label the hub accurately; link outfits to purchasable items; tune synonym/ranking behavior, keep customer-facing search labels human-readable, and consider available-size/category/stock filtering where useful.

Retest: shoppers can browse all products and buy a depicted outfit without manually searching names; repeat a fixed search panel covering exact product, leggings, hoodie, fit, color and misspelling. Compare search-to-PDP and PDP-to-cart rates when analytics is available.

Cart behavior and estimates

U50 13 Make cart costs and removal clearer

Priority: P2. Status: verified behavior plus usability recommendation. Confidence: high. Effort: small. Suggested owner: theme developer + ecommerce operations.

Evidence: a temporary Black/M Victoria Bodysuit at https://u50.com/products/victoria-bodysuit?variant=54217607774521 added correctly at $125. In https://u50.com/cart, changing quantity from 1 to 2 and selecting Update recalculated subtotal to $250. The cart offered no visible Remove control; setting quantity to 0 and selecting Update successfully emptied it. The quantity spinbutton had no accessible name in the inspected tree. The summary labeled $125/$250 as Estimated Total and said tax is calculated at checkout, but had no shipping line or shipping-calculation disclaimer in the summary. A discount field says discounts are applied at checkout; no code was submitted.

Action: add an explicit accessible Remove action and product-specific quantity label; clearly state when shipping is calculated or show a trustworthy estimate/progress indicator, using the corrected policy threshold. Make pre-checkout estimates visibly provisional. Avoid claiming shipping is free just because an estimated total omits it.

Retest: add, quantity update and removal work with mouse and keyboard; totals, discount status and shipping/tax qualification are clear; temporary cart is returned to empty. This test passed basic add/update/zero-quantity removal, with cleanup verified at 23:52:57 UTC on October 11. Checkout, delivery destination, taxes, discount application and payment remain untested.

Observed test sequence

Black M Victoria Bodysuit - Added one item at $125 with the selected variant intact.

Quantity update - Changed 1 to 2 and selected Update; displayed subtotal became $250.

Removal - Set quantity to 0 and selected Update; the empty-cart state returned.

Cleanup verified - 11 October 2026 at 23:52:57 UTC. No order, delivery destination, discount code or payment was submitted.

Interpretation: core cart arithmetic and the available removal path worked in this sample. The recommendation concerns discoverability, accessible naming and cost communication. It is not a finding that the cart or checkout is generally broken.

Identity and category content

U50 07 Brand identity is inconsistent in structured data

Priority: P2. Status: verified data-quality issue. Confidence: high. Effort: small. Suggested owner: SEO/content + developer.

Evidence: homepage Organization and WebSite names are u50.com, rather than the visible UNITED50 brand. Organization is minimal (name, URL, sameAs) with no logo/contact details. Some sameAs URLs differ from current footer profiles: TikTok @united.50 versus @united50apparel; YouTube @unitedfifty versus @United50official; Pinterest /unitedfifty versus /United50Apparel. Product brand is UNITED FIFTY.

Action: agree on UNITED50/U50 naming and legitimate aliases, verify owned social destinations, and use consistent Organization/WebSite/Brand references. Add a real logo and public support contact; use only verified organization details. Stable absolute IDs can link entities; an @id on every node is not mandatory.

Expected mechanism: clearer entity disambiguation and fewer conflicting brand facts. This is not a promised AI mention or ranking increase.

Retest: rendered and raw homepage data, storefront footer and product brand represent the same verified business; schema parses and aligns with visible text.

U50 08 Categories need buying guidance and metadata

Priority: P2. Status: verified omissions + content recommendation. Confidence: high. Effort: medium. Suggested owner: ecommerce content/SEO.

Evidence: 38 of 103 sitemap HTML pages have absent/empty meta descriptions, including 19 of 25 collections and core service/about pages. Men/Women descriptions are short brand slogans. https://u50.com/collections/leggings has a heading and four cards, but no substantive comparison of rise, length, support, pockets or use case. Many PDP meta descriptions are mechanically clipped at 320 characters; this is not a Google length-rule violation, but they can be made more deliberate. Training/regular Forest tee variants share the same description.

Action: start with commercially important categories and write concise, truthful buying help using available product facts. Distinguish training fit versus regular fit, lined versus unlined shorts, 7/8 versus capri, and the four fabric families. Add relevant product/guide links and unique titles/descriptions. Avoid filler or minimum-word-count targets.

Retest: priority landing pages answer real selection questions, all claims are approved, snippets summarize the correct intent, and Search Console indexing/query/CTR changes are reviewed with seasonality and merchandising changes recorded.

Note: 0 of 25 collection pages emitted JSON-LD. CollectionPage/ItemList can be an optional semantic enhancement; their absence is not an invalid Google product result and should not outrank useful content.

Product data and URL hygiene

U50 09 Product data can be richer

Priority: P2. Status: enhancement to a working baseline. Confidence: high. Effort: medium. Suggested owner: developer + catalog/SEO.

Evidence:

All 56 sitemap products emit parseable Product/Offer and BreadcrumbList JSON-LD: 1,140 variant offers with price, currency, stock and URLs. This is not full rich-result validation.

The direct Black/M https://u50.com/products/victoria-bodysuit?variant=54217607774521 selected matching color, size and imagery. In-page size changes did not update the URL, so sharing it loses the new selection.

North Laguna's rendered rating shares the base Product @id. Initial HTML lacks the rating and linked ID; the old detached-rating assertion does not hold for this rendered sample.

ProductGroup/explicit variant attributes, standard shipping and return policy schema were absent from the inspected raw nodes. These are enhancements, not automatic failures of the existing Product/Offer markup.

Action: synchronize catalog, stock, prices, visible products and schema. Add supported variant and approved shipping/return semantics where useful; prefer initial-HTML shopping data when feasible. Check rendered rating linkage before adding schema. Consider updating the URL on variant selection while preserving the product-group canonical.

Retest: Rich Results Test + Search Console URL Inspection on an in-stock product, mixed-stock product and reviewed product; direct variant URLs select the right state; no duplicate conflicting Product entities; store policies match approved business rules.

U50 10 Stale and utility URLs need deliberate treatment

Priority: P2. Status: verified index-eligible/stale content; actual indexing unknown. Confidence: high. Effort: small–medium. Suggested owner: SEO + content + developer.

Evidence:

/pages/home is a self-canonical 200 duplicate-style home destination.

/pages/swym-wishlist and /pages/swym-share-wishlist are 200, self-canonical and have no noindex/description.

/pages/accessories shows an image-editor placeholder and says accessories are unavailable, while /collections/accessories lists an in-stock-marked $120 tote. Both have the same title; the old page lacks H1.

/blogs/news is a 200 empty blog index with no article links.

A source-link scan across 103 pages found 13 zero-incoming candidates after normalizing collection-product aliases. These include utility pages but also /products/cami-tank and /products/u50-gift-card. Sitemap inclusion still allows discovery; this is not a claim of total orphanhood across every possible rendered/external source.

Action: assign each URL a keep, redirect, noindex or update decision. Redirect only to genuine equivalents; leave noindexed pages crawlable. Update stale availability/event text, link intended sale products from useful collections, and plan helpful content before promoting the empty blog.

Critical caution: the -copy shorts URL is Hillcrest ($70; S6WWS003), while the non-copy URL is Monte Vista ($75; S6WWS001). They are distinct products. Do not merge/delete by handle alone; any rename needs a redirect and link/feed updates.

Retest: URL inventory records keep/redirect/noindex/update decisions, no useful product is lost, sitemaps/canonicals/internal links agree, and Search Console confirms intended treatment over time.

Claim evidence and experience testing

U50 11 Publish evidence behind impact and material claims

Priority: P2. Status: verified evidence gap, not proof a claim is false. Confidence: high for page observation. Effort: medium, dependent on source documents. Suggested owner: brand/product + partnerships.

Evidence: https://u50.com/pages/social-responsibility-impact now names APAF, improving on the August audit, but still lacks a linked confirmation, dated U50 contribution or identifiable outcome supporting its broad campaign/impact language. https://u50.com/collections/united-renew states that certifications are verified, while its two boardshort PDPs list polyester/elastane without naming a recycled percentage or a specific certification in the sampled text. Several PDPs make performance/hypoallergenic claims without linked test details.

Action: publish appropriately scoped evidence: named partner confirmation where authorized, dates/program/contribution, specific fiber/certification details and relevant substantiation for product claims. Separate aspirations from completed results. Do not fabricate proof or turn broad wellbeing language into medical promises.

Retest: each material claim can be traced to approved evidence; buying guides and machine data use the same defensible wording. Useful primary proof may support human trust and reliable answers; citation gains are not guaranteed.

U50 12 Measure mobile accessibility and performance

Priority: P3. Status: mixed observed concern and untested coverage. Effort: medium. Suggested owner: frontend/QA.

Observed: homepage pale navigation/wordmark over light carousel imagery has visibly weak contrast in the desktop capture. This is a visual concern, not a measured WCAG contrast failure. Hero controls, semantic product radios, named search and size-guide dialogs were present; the size guide moved focus into its dialog on the later check. No unsupported blanket accessibility failure should be claimed.

Positive performance hygiene: initial hero image has high fetch priority and responsive sources; later hero images are lazy-loaded. The measured homepage initial HTML body was 245,534 uncompressed bytes; this is not total transfer size, speed or Core Web Vitals.

Action: manually test contrast across every hero slide, keyboard focus/escape/return, zoom, target sizes, image alternatives, and real mobile layouts at representative widths. Obtain field LCP/INP/CLS where available and reproducible lab traces for homepage, collection and PDP; investigate measured bottlenecks instead of assuming app count or HTML size causes slow performance.

Retest: documented mobile/desktop QA passes for key shopping journeys, resolved accessibility issues, and per-template performance evidence. Google good CWV thresholds are LCP ≤2.5s, INP ≤200ms and CLS ≤0.1 at the 75th percentile; no current U50 pass/fail is asserted.

SEO and AI search priorities

Resolve contradictory commerce facts first. Shopify's agent-discovery documents link /policies/shipping-policy and /policies/refund-policy, so the customer-policy inconsistencies also matter to agents.

Google supports ordinary helpful, indexable content and accurate product data for AI/search experiences. Special AI markup and llms.txt are not a Google ranking shortcut. Do not buy an unsupported GEO score or promise citations.

Distinguish OAI-SearchBot (search), GPTBot (potential training), and ChatGPT-User (user-triggered retrieval). Do not blanket-block or grant all AI bots without a specific policy decision. Public robots allows do not prove actual retrieval.

For Shopify, verify Agentic Storefront/Catalog eligibility and channel settings rather than automatically recommending a custom OpenAI product feed. Public infrastructure availability does not establish private channel configuration or acceptance of terms.

Measure ordinary organic search and AI visibility separately: indexing, impressions, clicks, qualified referrals, orders/conversion, cited URLs and repeatable query/prompt panels. No first-party results were available in this public audit.

Recent measurement documentation matters: Google's August 31, 2026 rollout update describes dedicated generative-AI performance reports, and September 24 adds multimodal filtering. Bing's AI Performance reporting is observational citation data. Check actual account availability before promising a dashboard or historical coverage.

What to measure after changes

Record a pre-change baseline for index coverage, organic queries, impressions, clicks, search-to-product and product-to-cart behavior, orders and conversion where first-party access permits.

Track AI discovery separately using actual referral sessions, attributed orders, cited URLs and a repeatable query or prompt panel. Provider reporting is observational and account availability must be checked.

Compare only after enough crawl activity and traffic to interpret the result. Record stock, promotions and seasonality; use matched untreated categories where practical. Reverse changes that harm conversion or accessibility.

The current audit has no first-party performance baseline. Infrastructure files, extra words or additional schema alone cannot establish improved rankings, citations or revenue.

Implementation sequence and remaining access

Suggested order

First - Operations approves shipping and return rules; product and inventory owners validate availability and fit data. These decisions unblock U50 01 through U50 04.

Next - Update customer-facing promises, the return route, promoted products, size guidance and wholesale instructions. Verify the same facts in policies, pages, product summaries and machine-readable data.

Then - Improve navigation and search, priority category content, identity references, product semantics, stale URLs and substantiation. Avoid duplicate schema and destructive cleanup based only on URL names.

Before signoff - Complete authorized checkout, mobile, accessibility, performance and private configuration checks. Retest the affected templates and record a before-and-after evidence trail.

Shopify admin audit remains pending

The requested admin inspection could not be completed while access was blocked. Public observations cannot establish the settings, app configuration, stock source or internal cause. The remaining review should verify these items once access is available:

Shipping profiles, zones, live rates, discount interactions, the exact free-shipping threshold and US/Canada availability.

Return rules, applicable fees/exclusions, the configured initiation route and how an authorized test or existing order reaches it.

Inventory locations, publication/channel availability and whether sold-out homepage ranges reflect intended merchandising.

Theme templates, size-guide assignments, catalog attributes, installed schema sources and any duplicate or conflicting output.

Search and merchandising configuration, navigation, redirects, policy destinations, relevant markets and localization.

Available Search Console, GA4, Merchant Center, Shopify analytics and agentic-channel data or settings. Do not assume account access, activation, eligibility or accepted terms.

Safety boundary: verification does not authorize a live purchase, real return, customer submission, subscription, publication or site change. Any such action needs the appropriate approval. An implementation plan should assign actual owners and dates only after the decisions and access gaps are resolved.

Technical appendix and earlier audit comparison

Working foundations to preserve

http://u50.com/ and https://www.u50.com/ 301 redirect to https://u50.com/.

All 103 sitemap HTML pages had a self-canonical URL; the tested collection-prefixed product URL canonicalized to the root product URL. An intentionally nonexistent URL returned 404.

Product descriptions, fabric composition/weight, prices and offers are server-readable. Sample variant deep links work. Price/stock are represented per offer; actual inventory correctness needs admin confirmation.

Product reviews are visible; homepage shows 41 reviews, but that displayed aggregate should not be mislabeled as 41 product-only reviews. No new catalog-wide rendered review-coverage count was computed.

HSTS, Content-Security-Policy, X-Frame-Options and nosniff headers were observed. Do not repeat the old assertion that all security headers are absent. Security headers are not a direct SEO ranking remedy.

https://u50.com/robots.txt is available, generally allows public content and blocks selected private/crawl-trap routes. No dedicated OAI-SearchBot/GPTBot disallow was found. This does not verify CDN/WAF treatment of actual provider traffic.

/agents.md, /llms.txt and /.well-known/ucp returned 200; the UCP version advertised is 2026-08-25. These are Shopify infrastructure surfaces, not evidence of actual ChatGPT impressions, activated channels or orders.

What changed from earlier audits

August 31 PDF and October 7 public report were used as historical baselines, then material claims were independently rechecked. No prior numerical score is recalculated.

Still present: wholesale form wording, minimal/domain-named Organization, absent descriptions, thin category guidance, missing headings on older pages, empty blog/utility index eligibility, shipping-page wrong H1, public claim substantiation gaps.

Improved relative to August / already recognized October 7: APAF is named; rendered hoodie rating connects by Product @id; security headers are present. Important render dependency remains.

October 7 command-line throttling was not reproduced in this low-concurrency successful public crawl. This is a test-specific result, not proof provider crawling is fixed.

Newly emphasized in this audit: contradictory shipping/return policy versions; rate/threshold conflict; promoted all-out-of-stock products; entire Renew collection unavailable; wrong size-guide mapping and men's top measurement label; shallow Shop All/search/outfit journeys.

Prior -copy-product cleanup needs qualification: distinct current SKU families are visible. Do not treat it as a proven duplicate.

Primary sources and reproducibility

Public-store observations were collected on 11 October 2026. Source links beside each finding identify the affected pages. The companion page-audit.csv contains all 103 sitemap HTML records with status, title, H1 count, description length, canonical URL and raw schema types. Counts are scoped to that inventory, not every possible storefront URL.

Platform documentation below was checked on 11 October 2026. Publication or update dates are included only when supplied by the source. Documented platform behavior supports a recommendation; it does not quantify the expected business result.

1. Google Product variants, updated September 8, 2026. ProductGroup/variant identity and direct-URL requirements.

2. Google Merchant listings. Product/Offer eligibility and supported shopping data; exact update date not extracted.

3. Google ShippingService, updated September 8, 2026. Standard organization shipping versus product exceptions.

4. Google returns, updated September 8, 2026. Accurate standard/exception return-policy representation.

5. Google technical requirements, updated December 18, 2025. Access, working pages and indexable content.

6. Google robots guidance, updated December 10, 2025. Crawl blocking is not reliable deindexing; noindex must be crawlable.

7. Google AI optimization, updated July 10, 2026. Ordinary fundamentals; no special AI files or markup shortcut.

8. OpenAI crawlers, undated. Independent search/training/user-fetch controls.

9. Shopify ChatGPT eligibility, undated. US sales, eligible Catalog, accepted terms and complete policies; actual U50 setup unverified.

10. Shopify Catalog/discovery, undated. Automatic discovery files and authoritative Catalog route.

11. Shopify channel measurement, undated. Sessions, orders, conversion, directional preview limits.

12. Google generative-AI reports, June 3 announcement / August 31 rollout update. Newer reporting guidance supersedes older combined-only descriptions.

13. Google multimodal reporting, September 24, 2026. New multimodal filters; no U50 account data inspected.

14. Bing AI Performance, February 10, 2026. Citation/grounding-query reporting is not a ranking metric.

15. Google Web Vitals, updated October 31, 2024. Field thresholds, not current U50 results.

16. W3C WCAG 2.2. Accessibility test reference; this audit does not certify compliance.

Recent-change research covered 11 September to 11 October 2026, expanded to 13 July where material, alongside current foundational documentation. No unsupported practitioner efficacy claim, inherited backlink count, ranking score or AI-citation frequency is treated as current measurement.