Mobile support and accessibility

What does mobile-friendly mean? Implementation and testing for mobile websites

Mobile-friendly means more than fitting a small screen: people must be able to read, tap, enter information, and complete their goals without losing content. Starting with single-URL responsive design, check content fetched by Google, reflow at 320–390px, touch, forms, and actual-device browsers.

Published 2025.01.08 Last updated 2026.07.25 Reviewed and edited by Finite Field Editorial Team
A person showing a smartphone screen while explaining mobile display
An illustration of smartphone use retained from the older article; it is not evidence of a device passing a test or of search rankings.

Mobile-first indexing, retirement of the Mobile-Friendly Test, responsive design, and WCAG 2.2 reflow and target size were checked against official Google, Chrome, web.dev, and W3C sources on July 25, 2026. Mobile support alone does not guarantee improved rankings or sales.

What you can implement and check

Rather than relying on one automated verdict, separately check design, accessibility, device interaction, and search-visible content.

Assess current requirements

Explain mobile-first indexing and how it differs from the retired test.

Build a resilient structure

Design viewport, fluid layouts, images, tables, and embeds based on content.

Check operability

Check reflow, zoom, touch, keyboard, forms, and motion.

Check tasks on real devices

Distinguish simulation from real devices and keep records for Chrome, Edge, and Safari.

A practical mobile-friendly requirement: completing the same goals

Google primarily uses mobile content for indexing and ranking. This does not mean a site without a smartphone-specific page can never appear in search. Aim for users to reach the same information and functions rather than pursuing a search-only label.

Consider responsive design first

Using the same URL and HTML with CSS adapted to screen and input characteristics makes differences in content, metadata, and structured data less likely; Google also recommends it as easier to implement and maintain. Do not immediately consolidate an existing separate-URL site without a plan covering canonical, alternate, redirects, and measurement.

Align key content and functionality with desktop

Do not reduce main content, headings, images, video, links, structured data, title, description, or robots only on mobile. An accordion can organize the initial view, but Google may not obtain important content absent from HTML until a user interacts.

Allow Googlebot to fetch CSS, images, and JavaScript

If robots.txt or authentication blocks resources needed for display, Google cannot reproduce what users see. In Search Console URL Inspection, review fetched HTML, screenshot, crawl time, user agent, and blocked resources.

Do not entrust the verdict to one test

Google's Mobile-Friendly Test and Search Console Mobile Usability report were retired on December 1, 2023. This does not make mobile quality unnecessary. Combine implementation checks, browser simulation, device interaction, URL Inspection, and usage data.

Alternatives to the Mobile-Friendly Test

Use DevTools and actual devices for layout defects, URL Inspection for search-visible content, Core Web Vitals for speed and stability, and completion rates or error records for interaction failures. Each tool has a different scope; do not collapse them into one pass.

Implement responsive web design starting with small screens

Rearrange the layout where content becomes cramped rather than using fixed layouts for device names. Preserve content and controls between 320px and 390px, in landscape, on tablets, and when text is enlarged.

  1. 1

    Set the viewport correctly

    Set a viewport in the head with width=device-width and initial-scale=1. Do not prevent zoom with maximum-scale=1 or user-scalable=no. Disabling zoom may hide layout problems but does not improve readability.

  2. 2

    Replace fixed widths with fluid layouts

    In Grid and Flexbox, check child min-width, long URLs, code, images, and iframes. Use minmax(0, 1fr), max-width:100%, and overflow-wrap as appropriate. Do not hide the cause by applying overflow-x:hidden to the whole page.

  3. 3

    Choose breakpoints based on content

    Record widths at which navigation, cards, tables, or forms become hard to use rather than choosing breakpoints by iPhone or Android model. A mobile-first single column that gains columns as space permits is easier to maintain.

  4. 4

    Fit images and embeds within containers

    Set image width and height attributes to the intrinsic dimensions and generally use max-width:100% and height:auto for display. Use srcset and sizes when appropriate. Check that video, maps, ads, and social embeds fit their containers and reserve space before loading.

  5. 5

    Choose a table presentation that preserves meaning

    A two-column glossary can become cards or a definition list. Keep table semantics for multi-column comparisons and provide horizontal-scroll guidance, a first column, caption, and th scope. If hiding columns, ensure information needed for decisions is not lost.

  6. 6

    Do not rely only on hover or screen width

    Do not reveal menus, explanations, and row actions only on hover; support touch, keyboard, and focus. Small screens are not always touch-only and large screens may be touch-enabled, so treat pointer and hover characteristics as supplementary conditions.

A person checking responsive implementation on a laptop
An illustration of development work from the older article, not a screenshot of a particular CSS implementation or test result.
<meta name="viewport" content="width=device-width, initial-scale=1">

Test completion of tasks as well as text size

WCAG 2.2 reflow generally requires content and functionality without two-dimensional scrolling at a width equivalent to 320 CSS pixels. For exceptions such as tables and maps, make that area operable without making the entire page scroll horizontally.

A person considering how people use smartphones
An illustration from the older article, not a measurement of disability characteristics or accessibility conformance.

Zoom and reflow

At 200% browser zoom and a narrow viewport equivalent to 320 CSS pixels, check text, menus, dialogs, fixed headers, and errors. Clipped text, overlapping buttons, or an off-screen close control mean the work is incomplete.

Target size and spacing

WCAG 2.2 AA Target Size (Minimum) generally requires at least 24 by 24 CSS pixels, with exceptions such as spacing. Do not target the bare minimum: give primary buttons and icons ample padding and check for accidental taps on adjacent controls using actual devices.

Names, state, and focus

Give icon buttons accessible names and communicate open and closed states with attributes such as aria-expanded. Check keyboard focus order, visible focus, movement inside modals, and where focus returns after closing. Do not break reading order with visual-only reordering.

Forms and input

Provide labels, appropriate input types and autocomplete, examples, error causes, and correction instructions. Check that the on-screen keyboard does not cover submit controls or errors, and that date selection, choices, and file uploads can be completed in Safari and Chromium browsers.

Color, motion, and orientation

Do not communicate state by color alone; add text or icons. Reduce unnecessary animation with prefers-reduced-motion and avoid flashing and automatic movement. Do not lock orientation to portrait without a business need; preserve information and controls in landscape.

Six stages of mobile testing

Choose representative pages and important tasks, then move from static checks to device testing, search checks, and post-launch monitoring so that a pass in one environment does not hide problems elsewhere. Chrome Device Mode is useful, but its official documentation describes it as an approximation of a real device.

  1. Test 01

    Choose representative pages and critical tasks

    Choose screens with different templates and importance, such as home, articles, lists, tables, search, contact, login, and checkout. Test opening menus, entering information, correcting errors, and submitting, not just viewing.

  2. Test 02

    Inspect basic HTML and CSS conditions

    Check viewport, headings, landmarks, labels, alt text, image dimensions, table headers, and duplicate IDs. If documentElement.scrollWidth exceeds clientWidth, find the first overflowing element.

  3. Test 03

    Test widths around 320, 390, and 768px and check landscape orientation

    Continuously vary width in Chrome or Edge DevTools rather than checking fixed sizes only. Inspect navigation, tables, long Japanese and alphanumeric strings, zoom, and the software keyboard. Record CPU and network throttling as test conditions.

  4. Test 04

    Use Chrome, Edge, Safari, and real devices

    Do not stop at Chromium; test Safari with the same depth of interaction. Where possible, use iOS and Android devices for touch, scrolling, back, input, rotation, sharing, fixed elements, and safe areas. A DevTools pass is not a real-device pass.

  5. Test 05

    Check what Google fetched

    Use Search Console URL Inspection to check index status, crawling user agent, fetched HTML, and screenshots. Compare main content, metadata, robots, canonical, structured data, and image alt across mobile and desktop.

  6. Test 06

    Track usage and regressions after launch

    Track Core Web Vitals, JavaScript errors, form failures, completion rates, and exits by device and browser alongside release dates. Do not blame devices for regressions: record the URL and action and rerun the scenario after fixing.

Light trails representing mobile speed and interaction testing under varied conditions
A speed illustration from the older article, not actual-device, network, or Core Web Vitals measurements.

Do not hide defects; roll back small changes by cause

Hiding symptoms by suppressing page-wide horizontal scroll, hiding elements, or disabling zoom can remove information and controls. Identify the cause, roll back a small change, and redesign.

The page overflows horizontally

Check fixed widths, min-width, 100vw and scrollbars, long URLs, code, transforms, tables, and iframes. Record parent and child dimensions and use fluid widths or local scrolling for the affected element.

A menu or dialog cannot be closed

Check visibility and accessible name of the close button, focus movement, Escape, background scrolling, and overlap with fixed headers. Isolate changes so CSS can be rolled back, and do not publish an update that hides key navigation.

A form cannot be submitted

Check required fields, formats, keyboard use, error placement, submitting state, duplicate submissions, network failure, and back navigation. Preserve inputs for retry and do not omit server-side validation.

Content or SEO information is missing only on mobile

Distinguish CSS-hidden content from HTML not delivered, then compare robots, lazy loading, canonical, structured data, image alt, and internal links. For separate URLs or dynamic serving, also check cache and Vary; list differences before fixing.

A speed change breaks interaction or display

Disable lazy loading, JavaScript bundling, fonts, cache, and third-party tags one at a time to isolate the cause. Test consent, analytics, menus, forms, purchasing, and accessibility as well as speed.

Historical 390px and browser checks of the Japanese article

On July 25, 2026, the generated Japanese article was checked in a Chromium-based browser and Microsoft Edge. The image records a 500px-wide Google Chrome view. A separate Safari WebDriver check compared viewport and horizontal scroll at 390px. The simulated image is not an actual device and was not used as a substitute for interaction testing. These historical checks do not establish results for this English translation.

A recorded 500px-wide Google Chrome view of the Japanese article
A 500px-wide view of the Japanese article recorded July 25, 2026. Its 390px overflow check was performed separately by comparing clientWidth and scrollWidth in Safari WebDriver; this screenshot does not establish that result.

Display

For the Japanese article: no page-wide horizontal scroll, one H1, no broken images, and contents before the body.

Interaction

For the Japanese article: mobile menu, contents links, FAQ expansion, contact link, and keyboard focus were checked.

Search and semantics

For the Japanese article: canonical, Japanese-only hreflang, Article, BreadcrumbList, FAQPage, and image alt were checked.

Pre-launch checklist

If any important task cannot be completed, stop publication, fix the cause, and repeat the same test conditions.

Layout and content

  • Key content remains available at 320–390px, in landscape, and at 200% zoom.
  • The page has no unintended horizontal scrolling.
  • Main content, metadata, and structured data match desktop.
  • Images, tables, video, ads, and embeds do not exceed their containers.

Interaction and accessibility

  • Menus, dialogs, and forms can be completed by touch and keyboard.
  • Control size, spacing, names, states, and focus are clear.
  • Error causes and corrections are understandable visually and through a screen reader.
  • Reduced motion, non-color state indicators, and orientation changes have been checked.

Browsers and search

  • The same important tasks were checked in Chrome, Edge, and Safari.
  • Differences from simulation were checked on actual devices or through remote debugging.
  • Fetched HTML, resources, and crawl information were checked with URL Inspection.
  • Publication date, test conditions, screenshots, and known limitations were recorded.

Official sources

Specifications and tools change; check current descriptions and update dates at the linked sources when implementing.

Frequently asked questions about mobile-friendly websites

Is mobile support unnecessary now that the Mobile-Friendly Test is retired?

No. The retired items were a specific Google test and Search Console report. Check display and interaction with DevTools and actual devices, search-visible content with URL Inspection, and speed and stability with tools such as Core Web Vitals.

Does responsive web design automatically pass?

The method alone does not determine a pass. Fixed widths, long strings, tables, menus, forms, zoom, keyboards, images, and JavaScript can all cause problems. Test representative pages and important tasks in real browsers.

Can desktop DevTools replace actual-device testing?

No. Chrome describes Device Mode as an approximation. Differences include mobile CPU, browser UI, on-screen keyboard, touch, and safe areas, so also check important tasks on actual devices or through remote debugging.

Does hiding horizontal scroll with CSS solve overflow?

No. Content and controls may remain clipped but hidden. Find the overflowing element and fix its cause, such as fixed width, min-width, long strings, tables, or iframes.

Does mobile support improve search rankings?

Higher rankings are not guaranteed. Aim for mobile users to read content and complete tasks and for Google to fetch the main content. Search also evaluates relevance, content quality, and other factors.

Related mobile and SEO guides

Check content, interaction, devices, and search in sequence, not with one pass/fail label

Mobile-friendly design starts with preserving key content and functions in a single-URL responsive design so users can complete tasks at 320–390px and with zoom. Implement structure and fluid layout first, then check accessibility, DevTools, Chrome, Edge, Safari, actual devices, URL Inspection, and post-launch data. Retirement of the Mobile-Friendly Test does not mean testing is over.

Verification policy

Finite Field Editorial Team

We edit articles about web development, servers, and business systems in an order readers can reproduce and use to make decisions. A passing automated result is not sufficient: we check Chrome, Edge, Safari, narrow screens, keyboards, and semantic structure.

For teams improving mobile display and interaction

Discuss mobile and accessibility improvements with Finite Field

Share affected URLs, devices and browsers, reproduction steps, screenshots, and recent changes if available. We can help isolate overflow, menus, forms, speed, and search display, and discuss responsive implementation and regression checks in Chrome, Edge, and Safari. First try the self-service steps in this article.

Discuss mobile and accessibility improvements

A consultation does not guarantee particular rankings, sales, conformance certification, delivery dates, or a quoted price. We will review the scope and test conditions before advising.