Adidas Audit

Observed Behavior: On the Ultraboost PDP, selecting size “10” in regular width, then toggling to “Wide,” retains size “10” as selected — even when size 10 Wide is out of stock. The “Add to Bag” button remains active with no visual indication that the selected combination is unavailable.

Technical Root Cause: The width toggle updates the displayed price and imagery but does not trigger a re-evaluation of the size availability matrix. The selectedSize state persists across width changes without validation against the new inventory subset.

Business Impact: Users add an unavailable SKU to cart, only to be blocked at checkout with an inventory error. This is especially damaging for performance running shoes where wide-width availability is already limited and customer frustration is high.

Remediation Path: On width toggle, re-fetch or re-filter the availability matrix and clear selectedSize if the current selection is unavailable in the new width. Show a brief “Size no longer available in Wide — please reselect” inline message.

Observed Behavior: During a Yeezy Slide restock on the Confirmed web integration, the “Join Waitlist” button shows a spinner on tap but never resolves. After 15+ seconds, the spinner disappears and no confirmation is shown. Refreshing the page shows the user is not on the waitlist.

Technical Root Cause: The waitlist API endpoint returns 503 Service Unavailable during peak load spikes. The frontend catch handler logs the error to Sentry but does not surface it to the user or offer a retry.

Business Impact: Users who believe they joined the waitlist but didn’t will miss the release entirely. On high-heat drops, this drives traffic to resale platforms and generates negative brand sentiment on sneaker forums.

Remediation Path: Implement a user-visible retry mechanism with a maximum of 3 attempts and exponential backoff. On final failure, show a clear message: “High demand — please try again in a few minutes” with a manual retry button. Add request queuing on the backend to absorb traffic spikes.

Observed Behavior: On mobile PDPs, pinch-to-zoom on the product image sometimes zooms the entire browser viewport instead of the image. The behavior is inconsistent — it depends on whether the first touch lands on the image canvas or a surrounding element.

Technical Root Cause: The custom image zoom library uses touch-action: none on the image container, but an ancestor element has touch-action: manipulation, which takes precedence in the cascade. The conflict creates unpredictable gesture routing.

Business Impact: Users trying to inspect product details (stitching, material texture, colorway accuracy) are frustrated by the broken zoom experience and may hesitate to purchase without confident visual inspection.

Remediation Path: Audit the touch-action CSS inheritance chain and ensure the image container’s gesture handling is not overridden by ancestors. Use touch-action: pinch-zoom on the container and handle pan gestures via JavaScript to avoid conflicts.

Observed Behavior: A logged-in Creators Club member adds items to cart, but the “Points you’ll earn” display in the cart summary shows “0 points” until the page is manually refreshed.

Technical Root Cause: The cart component renders with a cached loyalty state from the initial page load. The points calculation API is called on page mount but not re-triggered when cart contents change via client-side state updates.

Business Impact: Creators Club members who don’t see their points reward at the moment of decision lose a key incentive to complete purchase. The loyalty program’s conversion lift is undermined by invisible rewards.

Remediation Path: Trigger a points recalculation on every cart mutation (add, remove, quantity change). Display the points with an animated counter to reinforce the reward feedback loop.

Observed Behavior: On mobile PLPs, opening the filter panel and selecting options requires the user to tap “Apply” (which closes the panel) AND then tap a separate “Done” button that appears on the PLP. If the user taps only “Apply,” filters are applied but the sticky overlay remains.

Technical Root Cause: The filter component has two layers: an inner accordion for filter groups with its own “Apply” action, and an outer sheet container with its own “Done” dismiss action. The “Apply” button commits the filter state but does not close the parent container.

Business Impact: The double-dismiss pattern confuses users. Analytics show that 22% of mobile filter sessions end with the user tapping the browser back button (navigating away from the PLP entirely) instead of correctly closing the filter panel.

Remediation Path: Merge the “Apply” and “Done” actions into a single “Show [N] Results” button at the bottom of the filter sheet, following the pattern established by Amazon and Zalando. Dismiss the sheet on tap.

Observed Behavior: On the Stan Smith PDP, the “Off White” swatch renders as a pure #FFFFFF white circle, visually identical to the “Cloud White” swatch. Users cannot distinguish between the two colorways without clicking each one and comparing photos.

Technical Root Cause: Color swatches are rendered from a hex value stored in the PIM system. The PIM entry for “Off White” uses #FFFFFF as a fallback because no one entered the correct warm-white hex. Several colorways share placeholder hex codes.

Business Impact: Users may order the wrong colorway, driving returns. For Originals products where colorway identity is central to the purchase decision, this friction directly impacts satisfaction.

Remediation Path: Audit the PIM color hex database for all active SKUs. Replace placeholder values with photographically-sampled hex codes. For complex multi-tone colorways (e.g., “Core Black/Solar Red”), use a split-circle swatch or a cropped product image thumbnail instead of a single hex.

Observed Behavior: On PLPs, the “Quick Add” button on apparel items opens a size overlay and adds to cart. On footwear, the same button navigates the user to the full PDP instead of showing a quick-select size panel. Users expecting the same interaction pattern are disoriented.

Technical Root Cause: The “Quick Add” component conditionally renders a size picker only for apparel categories. Footwear was excluded because the size matrix (half sizes, widths) was deemed too complex for the overlay format.

Business Impact: Footwear shoppers who know their size lose the fast-path purchase flow. Each redirect to the full PDP adds 3-5 seconds and a full page load, increasing abandonment on browse-heavy sessions.

Remediation Path: Extend the Quick Add overlay to support footwear sizes. Display a condensed grid of available sizes (no width selector — default to Regular). For wide-width products, include a small “Need Wide?” link to the full PDP.

Observed Behavior: During checkout, shipping method options (Standard, Express, Next Day) appear immediately on page load, before the user enters their shipping address. After the address is entered, the options re-render with different prices and delivery dates, causing a layout shift.

Technical Root Cause: The shipping options component renders a default set of options on mount using a fallback warehouse location. It re-fetches and re-renders after the address form’s onBlur event fires on the ZIP code field.

Business Impact: Users see an initially low shipping cost, then watch it change — creating a “bait and switch” perception. Unexpected cost increases at checkout are the #1 reason for cart abandonment across e-commerce.

Remediation Path: Do not render shipping options until the ZIP/postal code is entered and validated. Show a placeholder: “Enter your shipping address to see delivery options.” Animate the transition to avoid layout shift.

Observed Behavior: Non-members browsing the site see no mention of Creators Club benefits (free shipping, early access, birthday rewards) during the shopping flow. The program is only promoted in a footer link and a post-purchase “Join now” email.

Technical Root Cause: The Creators Club marketing module was scoped for the account section only. It is not integrated into the PDP, cart, or checkout templates for anonymous/non-member users.

Business Impact: Adidas’s loyalty program is a key DTC differentiator, but invisible benefits cannot drive conversion. Users who would convert for “free shipping as a member” never discover the offer at the point of decision.

Remediation Path: Add a persistent “Creators Club members get free shipping” banner on the cart page for non-members. On the PDP, show “Join Creators Club to earn [X] points on this purchase” next to the price. A/B test against the current experience.

Observed Behavior: Users who add items to cart and leave the site receive an abandonment recovery email after 2 hours. There is no on-site exit-intent intervention, no web push notification, and no SMS follow-up.

Technical Root Cause: The abandonment flow is handled entirely by the email marketing platform (Salesforce Marketing Cloud). Exit-intent detection and web push are not implemented on the frontend.

Business Impact: Email-only recovery captures a narrow slice of abandoners. Industry data shows that exit-intent overlays recover 5-10% of abandoning sessions, and web push recovers an additional 3-5%.

Remediation Path: Implement an exit-intent overlay on the cart page offering free shipping or a 10% discount code for first-time buyers. Integrate web push opt-in during the checkout flow. Add SMS recovery for users who provide a phone number during guest checkout.

Observed Behavior: The “You Might Also Like” carousel on the PDP and cart page frequently displays products with most sizes out of stock. Clicking through to these products shows “Notify Me” on nearly all sizes.

Technical Root Cause: The recommendation engine uses collaborative filtering based on purchase history without a real-time inventory availability filter. Recommendations are cached for 24 hours.

Business Impact: Showing unavailable products in upsell positions wastes high-value real estate and trains users to ignore the recommendation module. Each dead-end click is a lost AOV opportunity.

Remediation Path: Add an inventory availability filter to the recommendation pipeline. Exclude products with less than 30% size coverage from upsell/cross-sell surfaces. Reduce the recommendation cache TTL to 1 hour.

Observed Behavior: Guest checkout users enter their email address, then receive a verification code that they must enter before proceeding. This adds a full context-switch (check email, find code, switch back to browser, enter code) before they can even begin entering shipping details.

Technical Root Cause: The guest checkout flow was refactored to prevent bot-driven fraud. Email verification was added as a bot-mitigation step but was not evaluated for its impact on legitimate user conversion.

Business Impact: The verification step adds 60-90 seconds and a complete context switch to the checkout flow. This disproportionately impacts mobile users and is a significant abandonment driver for impulsive purchases.

Remediation Path: Replace pre-checkout email verification with a less intrusive bot-mitigation method (e.g., reCAPTCHA v3, device fingerprinting). Verify the email post-purchase for order communication purposes only.

Observed Behavior: Viewing the page source on key PLPs (e.g., /us/men-shoes) reveals a minimal HTML shell with a <div id="app"></div> and no product content. All product tiles are injected via JavaScript after hydration.

Technical Root Cause: The PLP architecture uses a React SPA that fetches product data client-side via API calls on mount. Server-side rendering is not implemented for PLP routes.

Business Impact: Googlebot can execute JavaScript but deprioritizes JS-rendered content. High-value category pages like “men’s running shoes” are competing against server-rendered competitors (Nike, New Balance) with a significant crawl and indexing disadvantage.

Remediation Path: Implement SSR or ISR (Incremental Static Regeneration) for all PLP routes. Ensure the initial HTML payload contains the first 24 product tiles with full markup (title, price, image, link). This is the single highest-impact SEO fix.

Observed Behavior: adidas.com/us, adidas.co.uk, and adidas.de serve English-language PDPs for global products (e.g., Ultraboost) with identical descriptions. No hreflang tags link these regional equivalents.

Technical Root Cause: Regional storefronts are deployed as separate applications with no shared content-linking infrastructure. The SEO team submitted hreflang via Search Console but it is not reflected in the page <head>.

Business Impact: Google treats the regional pages as duplicates, consolidating ranking signals to a single version (often the US page). The UK and DE storefronts lose organic visibility for their local markets.

Remediation Path: Inject <link rel="alternate" hreflang="en-US">, hreflang="en-GB", and hreflang="de-DE" in the <head> of every PDP. Build a shared product-ID mapping service to automate cross-region linking. Include x-default pointing to the global storefront.

Observed Behavior: Product reviews on the PDP are loaded via a third-party widget (Bazaarvoice) that injects content via JavaScript after user scroll. The review text, star ratings, and review count are not present in the initial HTML or in any structured data.

Technical Root Cause: The Bazaarvoice integration uses a lazy-loaded JavaScript embed. The review data is fetched from Bazaarvoice’s API on scroll intersection, not from the Adidas backend.

Business Impact: Review content is a rich source of long-tail keywords (“comfortable for wide feet,” “runs small”). Without it in the crawlable HTML, Adidas misses ranking opportunities for these high-intent queries. Review schema is also absent from rich results.

Remediation Path: Fetch review summary data (average rating, review count, top 3 reviews) server-side from the Bazaarvoice API and render them in the initial HTML. Add AggregateRating and Review schema in the product JSON-LD.

Observed Behavior: The XML sitemap at /sitemap-products.xml contains approximately 12,000 URLs for products that have been discontinued. These URLs return 404 Not Found.

Technical Root Cause: The sitemap is generated nightly from the product catalog database, which retains discontinued products with an is_active: false flag. The generator does not filter on this flag.

Business Impact: Google’s crawler wastes budget on 404 pages, and the high ratio of dead URLs signals low sitemap quality, potentially slowing crawl rate for new and updated products.

Remediation Path: Filter the sitemap generator to exclude is_active: false products. For recently discontinued products (< 6 months), implement 301 redirects to the closest current equivalent. Monitor crawl stats in Google Search Console to confirm budget recovery.