A desktop reviewer has the payroll notice, language selector and help link in front of her. An employee reading the same page with enlarged text sees a little of the headline and a fixed menu that fills half the screen. A floating footer sits over the link to the alternate version. Each component worked in isolation. The layout put the pieces in competition for space exactly where the employee needed room to read.
Plan for two overlapping tests, rather than checking one narrow screenshot and marking “mobile” complete. Using a sample onboarding note, ask a tester to find the employer’s instructions, read the applicable date at their preferred size, and open the help version. The notice should remain useful before it points to a separate system, and completing a valid test does not require opening a real payroll account.
Start with a real small screen
Open the actual notice from a test email on a phone. Use the same link a new employee will find and distinguish between the mail client’s in-app view and the normal browser if they render differently. Read the page in portrait. Try the language controls without enlarging the page yet. If a button is visually clipped, determine whether it is still operable and whether the label explains where it will go. Then vary text size and repeat. Native text settings, browser scaling and browser zoom do not all change the same dimensions. Record what you changed so a colleague can recreate the result.
Don’t make “rotate the phone” a prerequisite for reading instructions. Some users cannot reposition a mounted device, and a worker may not know that horizontal text will appear only after turning it. Similarly, a chart may require scrolling inside the chart; a notice title, paragraph and help link do not normally need to extend beyond the viewport. Keep the essential steps and contact information in ordinary flowing text. Double check that any links stay identifiable when the layout rearranges itself and that a visual hierarchy is not doing the work that real headings should do.
Check what happens when zoom squeezes the available width
W3C’s reflow explanation describes testing vertically scrolling web content at a width equivalent to 320 CSS pixels, subject to exceptions for parts whose meaning depends on a two-dimensional layout. One setup starts with a 1280 CSS-pixel viewport and applies 400 percent browser zoom. Do not confuse physical screen pixels with CSS pixels. Use a proper viewport measurement or browser tools, and note adjustments like side panels that affect the space available for the page. Try zooming without a simulated device as well; a desktop browser and its sticky panels may expose different behavior.
Move through the route, not only down the clean headline. Expand the explanation for a printed version. Open and dismiss the privacy panel if the employer’s site has one. If menus stick to the top or bottom, check whether keyboard focus moves behind them. Check the same path with the provider name, label of the alternative format, and date wrapped onto multiple lines. Force a test layout where the translated text is longer than the English label. A two-line action should not be squeezed into a fixed single-line box. For interactivity, use a real keyboard path to see which controls survive rather than concluding from their appearance that they are reachable.
Try going back after opening a secondary resource. If the browser restores scroll but the sticky menu returns to its expanded state, the reader may lose the original place. Opening a help link should not leave the user hunting for the line they were just reading.
Differentiate a missing feature from a risky fallback
Suppose a scanned notice cannot be enlarged without turning the type into a blur. Allow a tested HTML version alongside it and repair the source if it is yours. If the only support button is hidden by a footer on the phone, move the link into the document flow rather than asking an employee to switch devices. Where you cannot fix a third-party destination immediately, the fallback needs an owner, comparable information and timely availability. An instruction to find a desktop computer at work is not an equivalent route for someone who gets the notice after the work site closes.
For each failure, record the starting link, viewport or device, text-size or zoom settings, label of the missing action, and number of steps needed to recover. W3C’s Easy Checks combine resizing and other observations in an introductory assessment. Track the actual action to completion after a fix. Reporting that a single width looks better does not tell you if a reader can finally request help.
Pair this with the keyboard pass for focus and alerts. Use the translation review if large-text problems appear only in a different language, and take remaining format requests through the contact handoff drill.