Google's field data from real Chrome visitors puts layout shift (CLS) on the harpercollins.com homepage at 0.35, against a limit of 0.1, and the main image takes 3.1 seconds against a 2.5-second target. Across the whole site, CLS is 0.26. The largest shift is the shelf with the NEW RELEASES / BESTSELLERS / COMING SOON tabs, which is where shoppers tap a cover. The slowest element is the rotating hero slider. When covers move under a thumb, taps land on the wrong book or tab, and the Direct to Consumer team buys paid social for about 40 titles a season that sends people to pages like this.
First action
Set a fixed minimum height on the tabbed shelf in the theme and re-run PageSpeed.
How to measure it
Field CLS (Chrome UX Report)
- Now
- –
- Target
- under 0.1, with LCP under 2.5 s
- By
- 6 weeks (one full 28-day field window after release)
Steps
Reproduce the shift
Record the homepage on a mid-range phone in the Performance panel and confirm that the shelf and the hero are what moves.
E-commerce Developer (Shopify) 1 day
Reserve the space
Fixed aspect-ratio cover slots, a minimum height for the shelf and tab bar, and no late banners or app widgets above them.
E-commerce Developer (Shopify) 2 days
Server-render the first tab
Print the New Releases covers in the theme HTML so the shelf doesn't wait for JavaScript.
E-commerce Developer (Shopify) 2 days
Replace the slider
Show one static, preloaded hero image with set dimensions, and move the other promotions lower on the page.
Head of E-commerce 2 days
Roll out to other templates
Collection and product pages that use the same shelf component.
E-commerce Developer (Shopify) 1 week
Test the change
A/B test the new homepage against the old one in the existing testing tool, on cover click-through and add-to-cart.
E-commerce Analyst 2 weeks
Watch field data
Check the Search Console Core Web Vitals report and the Chrome UX Report weekly until a full 28-day window has passed.
E-commerce Developer (Shopify) 6 weeks
Why it works
Reserve the shelf's space before its books load: fixed aspect-ratio cover slots, a fixed-height tab bar, and nothing inserted above the shelf after load. Put the first tab's covers in the theme's server HTML so they don't wait on scripts. Replace the rotating hero slider with one static image that has a width and height, is preloaded and uses fetchpriority=high; move the other slides further down the page. Then fix the collection and product templates that use the same shelf, and measure in the Chrome UX Report and in store conversion.
Core Web Vitals are measured on real visitors and Google uses them in search ranking. Reserving space is a CSS and theme change rather than a redesign. A steady page means more taps reach the intended cover and more of them become add-to-carts, which matters most on mobile, where paid social lands.
Ready to use
ChecklistHomepage stability fix list
[ ] Cover slots with a fixed aspect ratio (width and height on every img) [ ] Minimum height on the NEW RELEASES / BESTSELLERS / COMING SOON shelf and tab bar [ ] First tab's covers in the server HTML [ ] Static hero with preload and fetchpriority=high; slider removed [ ] No banners or app widgets inserted above the fold after load [ ] Same fix on collection and product templates [ ] A/B test on cover clicks and add-to-cart [ ] Field CLS under 0.1 after 28 days
Risks and how to handle them
Merchandisers lose the rotating hero slots
Schedule the static hero to change by day or campaign, and keep the other promotions further down the page.
Shopify apps inject content above the shelf
List every app block on the homepage and give each a reserved slot or move it below the fold.
Evidence
- Google PageSpeedMobile test of the homepage (November 3, 2026): real Chrome visitors fail Core Web Vitals on layout shift and loading. CLS is 0.35 at the 75th percentile against Google's 0.1 limit, and LCP is 3.1 s against 2.5 s; INP (193 ms) passes. Across the whole site CLS is 0.26 and LCP 2.7 s. The biggest shift is the homepage shelf with the "NEW RELEASES / BESTSELLERS / COMING SOON" tabs, and the LCP element is the hero slider image. Lab score 32/100 on mobile (57 on desktop), lab LCP 11.7 s.Source
- Google PageSpeedHomepage weight (November 3, 2026): 7.7 MB in 359 requests, 3.5 MB of it JavaScript and 4.3 MB from 40 third parties. Google Tag Manager blocks the main thread for about 1 s. Two testing and heatmap tools load side by side (VWO and Crazy Egg), plus Facebook, TikTok, Snapchat, Pinterest, Reddit, Bing and The Trade Desk tags, Attentive and Sailthru. Google estimates 1.3 MB of unused JavaScript, 0.9 s of render-blocking requests and 0.9 MB of image savings.Source
- HiringSenior Associate, Advertising (NYC Hybrid), Direct to Consumer team: "executes media buying across HarperCollins adult and children's titles (roughly 40 titles/season)" on Meta, TikTok, Snapchat, Reddit and Pinterest, and will "Set up and launch programmatic campaigns via The Trade Desk".Source