Black Friday Load Test: How Top Themes Carry Heavy Promo Homepages

Most merchants pick a Black Friday theme by watching its demo load once, on a good connection, with none of the countdown banners or extra product blocks a real sale adds. That single load tells you almost nothing about what happens when a collection page carries a sale badge on every card and a sticky bar at the top, because the market’s own Core Web Vitals numbers show a 52.5 point spread on mobile between the best and worst themes, all measured from real visitor traffic in July 2026, not a lab run.

The market’s performance floor is lower than most merchants assume

Shopify sets its Theme Store bar at a Lighthouse performance score of 60 and an accessibility score of 90, averaged across a theme’s product, collection, and home pages, on both desktop and mobile. That is a floor every listed theme has to clear, and it is a low one compared to what the strongest themes actually deliver in the field. Real-world data across 247 themes with sufficient mobile samples puts the median share of stores passing all Core Web Vitals at 85.5%, with the best theme in the market reaching 98.6% and the worst sitting at 46.1%. Desktop tells a similar story across 199 themes: a median of 86.2%, a ceiling of 99.2%, and a floor of 30.3%. A theme that clears Shopify’s minimum bar to get listed can still land anywhere in that wide range once real shoppers start loading it.

What Horizon’s collection settings actually let you touch

Horizon’s sections/main-collection.liquid section, named t:names.collection_container in the theme’s own schema, carries 14 settings as measured in commit 8b42ace from August 31, 2026. The ones that matter for a heavy sale page are products_per_page, a range from 8 to 36 in steps of 4, enable_infinite_scroll, a checkbox, product_card_size, a select with 4 options, and mobile_product_card_size, a select with 2 options. The search results section, sections/search-results.liquid, mirrors most of this with 15 settings total, including the same products_per_page range and the same enable_infinite_scroll checkbox, plus a paragraph field for merchant-facing notes.

None of these settings compress images or defer offscreen media on their own. Raising products_per_page to 36 during a sale, with enable_infinite_scroll off so shoppers see everything at once, adds real weight to the page that the setting panel does not flag. The range simply lets you choose the number, it does not warn you what that number costs.

Two block types that change how heavy a homepage gets

Horizon’s blocks/_accordion-row.liquid, named t:names.accordion_row, has 5 settings across 2 block types: a heading text field, an open_by_default checkbox, an icon select with 59 options, an image_upload image picker, and a width range from 12 to 200 pixels in steps of 2. It is a light block. There is no video field anywhere in it.

blocks/_layered-slide.liquid, named t:names.slide, is a different scale entirely: 20 settings across 10 block types. It includes media_type_1, a select choosing between image and video, an image_1 image picker, a video_1 video field, content_direction as a select with 2 options, vertical_on_mobile as a checkbox, and horizontal_alignment as a select with 4 options. A homepage that stacks several layered slides, each set to video, is carrying a fundamentally heavier load than one built from accordion rows, and nothing in the settings panel stops a merchant from doing exactly that during a sale.

A word search in Dawn’s code is not a performance feature

Searching Dawn’s codebase for the word “load” returns 96 hits out of 351 files, and searching for “top” returns 150 hits out of 351 files, as measured in commit 258f00f from August 10, 2026. Neither search corresponds to a dedicated section or setting. Both words live inside asset files, assets/animations.js, assets/base.css, assets/cart-disclosure-modal.js, assets/collage.css, assets/cart-drawer.js, and assets/cart-notification.js, which are internal implementation code, not anything a merchant configures. This absence is worth stating plainly because merchants searching for a “load speed” toggle or a “top of page” setting in Dawn will not find one under that name, and no amount of digging through the settings panel will surface it, because it was never built as a merchant-facing control.

Price and reviews say less than the traffic data does

The median theme price across the market is $300, inside a $100 to $500 range, and the median number of reviews behind the 4.53 average rating across 308 themes is only 23. That is a thin sample to lean on when picking a theme meant to carry a Black Friday homepage. Somewhere in the broader 329-theme ranking sits UTD’s catalog of presets (browse UTD’s themes), priced and rated alongside everything else in the market, with no measured claim here that any UTD preset leads on Core Web Vitals or Lighthouse scores, because the facts gathered do not isolate that vendor’s numbers from the aggregate.

If a Black Friday homepage matters more than a demo reel, check where a candidate theme sits on the CrUX pass rate before checking its star rating, because 23 reviews will not tell you what real mobile traffic will. Then open the actual section files for products-per-page limits and video-capable slide blocks before the sale starts, not after the first spike hits.