ABSee / Guides / Homepage design
GuideThis guide covers what ecommerce brands test on their homepage, the page many shoppers open first, apart from the menu and search. ABSee has recorded 100+ brands testing here over May to October 2026.
Updated October 2026 · Source: ABSee record · Methodology →
What changed The homepage no longer shows the 'All categories' block of round shortcut tiles or its heading, so shoppers scroll straight from the hero into the featured brand campaign.
Seen 22 Aug to 30 Aug 2026
Each tile is one kind of change brands tested on the homepage. The chip is a size word: how many brands in the record tested it.
Photo swaps, still against moving footage, carousel against a single image, a shorter hero.
Category tiles removed, reordered or replaced by pills, a category row shown by default.
Picked-for-you, most-loved and recently viewed rows, ordered by the shopper's interests.
Newsletter pop-ups shown straight away or delayed, offer modals, loyalty and account banners.
Free-shipping messages removed, rotating against fixed banners, sale wording by country.
A review-score strip, benefits cut to two claims.
A welcome-back recently-viewed pop-up, a cart-reminder pop-up.
Homepage visitors sent straight to a deals hub or to the all-products grid.
A quiz or diagnostic offered from the homepage.
Four of the themes, with what brands changed. Menu and search changes are covered in the search and navigation guide.
The first screen is tested in many ways: a new photo, still or moving footage, a carousel or one image, a compact hero with side-by-side buttons, text moved below a shorter image, and a hero chosen by audience.
Brands tested whether shoppers need shortcuts into the range from the homepage: tiles removed, reordered or replaced by pills with products, and swipeable tile rows. Galeries Lafayette tested removing its round category shortcut tiles, shown at the top of this page.
Rows below the hero picked for the shopper, most-loved and recently viewed products, rows ordered by the shopper's interests, and recommendations switched to discounted products.
When a sign-up pop-up appears, what it offers and how a visitor declines. Timberland tested adding a member-benefits list and a clearly named opt-out to its pop-up; the record is shown in full below.
Every theme above is made of records like this one, with every field shown: before and tested, what changed, the likely hypothesis, the takeaway, the dates and what the brand did next.
TimberlandHomepageSeen 6 Aug to 4 Oct 2026
Three hypotheses to bring to your assistant about the homepage, each written as if, then, because:
Our hypothesis: if we remove the category tiles under the hero, then more shoppers will reach the featured products, because nothing sends them into the menu first.
Comes back with comparable experiments, for and against.
Our hypothesis: if we delay the sign-up pop-up, then more shoppers will browse past the homepage, because nothing blocks the first screen.
Comes back with comparable experiments, for and against.
Our hypothesis: if we replace the hero carousel with one image, then more shoppers will click through, because there is one clear message.
Comes back with comparable experiments, for and against.
Start from where your own shoppers leave or hesitate.
Find brands like yours, what they tested and what they did next.
Only your own experiment tells you what works on your site.
No outcome rates. Whether each brand kept a change is part of the Intelligence plan.
First seen is when ABSee first recorded a test, not when the brand began it.
One experiment is one brand's decision, with its own shoppers.
ABSee publishes no rate for any theme here. The record shows, per experiment, whether the brand kept the change; that is part of the Intelligence plan.
In the ABSee record, many brands tested the hero, category entry points and personalised product rows. Several tested sign-up pop-ups and announcement bars.
No. It describes what brands put to a test on the homepage, with two real records. The before and after for every experiment is in the record.
Galeries Lafayette and Timberland, as public examples. The other brands are named in the record.
What do brands test on their homepage?
Works in Claude, ChatGPT or Gemini once ABSee is connected.
Connect your AI →
