Do layout A/B tests work on homepages?

Better here than elsewhere. Layout tests win 27% of the time on homepages, against 19% across all page types — 8 points better. If you are going to test layout anywhere, this is the page for it.

It is also the strongest change type we can measure on homepages — nothing else we track wins more often there.

The numbers

	homepages	all page types
Tests analysed	100+	250+
Beat control	27%	19%
No measurable difference	64%	—
Lost to control	9%	—
Median lift when it won	+15%	—

Median traffic per variant was 2837 visitors.

What else is worth testing on homepages

change type	win rate on homepages	tests
Layout (this page)	27%	100+
Hero image	18%	50+
Price framing	16%	49 tests
Social proof	15%	50+
Styling	15%	500+
Body copy	14%	100+
Split URL	14%	100+
CTA copy	11%	100+
Headline	10%	250+

Only change types with at least 30 tests on homepages appear here.

What counts as a layout test

Blocks move, reorder or change position. No copy is rewritten and no new content is added. If you moved a section *and* rewrote it, that is two variables and the result will not tell you which one worked.

Control and variant wireframe for a layout A/B test — Control on the left, variant on the right. For how this differs from adjacent categories, see the layout benchmark.

Layout tests that won on homepages

hides the existing carousel and replaces it with a new grid of product/promo cards (award banner, stairlift categories) with images, titles, and CTA buttons
replaced the three-audience card selector layout with a single personal-notary focused hero featuring a new headline, process steps, price framing box, and trust stats/rating
Replaced the generic three-audience selector page with a single-focus, direct-response layout: new headline targeting individuals, product screenshot image, process steps, pricing badge, and trust stats replace the segmented card grid
Hides native pricing/header/footer/documents sections and injects a custom two-card offer block with default-selected add option, custom CTA button, and proxy click handling.

Layout tests that did not

Hides native family-plan banner, header, image, pricing box, features, and footer, replacing them with a custom injected two-card add/skip family-plan selector with default-selected add option, dynamic CTA text/discount copy, and a loading spinner state.
Injects a new sticky header with logo, nav links, and mobile burger menu, and hides the existing collections selector element
Moves the color swatch picker to the top of the product details container and restyles spacing/alignment

These are individual tests, not rules. Each ran on one site, with one audience, against one page we are not showing you. They are picked to be illustrative rather than sampled at random, and a change that won here can lose on your page for reasons none of this captures. Read them as prompts for what to test, not as findings to copy.

Who ran these tests

This is every Mida account that ran a readable test — in-house marketers, founders, product teams, and agencies working on client sites. Nothing here is filtered by who ran the experiment or how experienced they are.

Low win rates are normal in experimentation, including at the top end. Microsoft's experimentation team, reporting on its own platform, found that only about one third of ideas improve the metric they were designed to improve — and that roughly another third actively hurt it. That is a dedicated experimentation organisation with research, prioritisation and review behind every test.

A mixed population like this one runs below that. The gap is roughly what disciplined practice buys you: ideas grounded in research rather than opinion, one variable at a time, and tests built so the result can actually be read.

So treat these as a general base rate, not a target. If you would rather not close that gap the slow way, find a CRO agency in the United States, the United Kingdom, the Netherlands or Australia — or browse every region.

A win is a variant that beat its control on that test's primary goal with a statistically significant result. Tests that never got enough traffic to say anything either way are excluded. The methodology has the full detail, including what these numbers cannot tell you.