A minor fix could block a menu, break a form, or move a button. This short test, repeated after each upload, identifies defects before visitors.
Sometimes a change seems too small to justify a recipe: change a sentence, add a tracking script, correct a form, update an extension or adjust the menu.
However, these interventions often affect elements shared by several pages. A modified style rule for a button may offset other blocks. A consent banner can hide a call to action. An update can leave the form visually intact while stopping its notifications from being sent. The cache can also serve an old version to some visitors.
Automated checks are useful, but they don’t answer a very simple question: Can someone who discovers the site on their phone understand the offer, reach the main action, and complete their journey?
A five-minute test can quickly address this issue. It does not replace a complete recipe, an accessibility audit, or a security check. It serves as a safety net after each change.
Prepare a test that resembles a real visit
The test must be carried out on a real phone, with a current browser and without being connected to the site administration. A private window prevents an old session, a cookie or an overly favorable cache from masking a problem.
You have to start from a page through which a visitor can actually arrive: the home page, an important service page or an article visible in Google. The goal is not to browse the entire site. It is about verifying a short path from discovery to main action.
Before you begin, simply choose:
- an entry page;
- information or an offer to find;
- an action to complete, for example sending a request;
- a receiving address where to check the result.
The timer can then start.
First minute: look at the first screen without explaining it
The first control concerns what appears before any scrolling.
The name of the company, its activity and the main action must be understandable without zooming. You must also check that no banner, menu, chatbot or floating element hides the text or the main button.
At this stage, the most common anomalies are very concrete:
- a title cut over three or four lines;
- a button pushed back under a visual that is too high;
- a consent window that cannot be closed;
- text that is too small or has too little contrast;
- horizontal scrolling caused by an image or block that is too large;
- an item that moves while loading.
This first minute does not require any technical interpretation. If the visitor has to search what the company does or move the page horizontally, the defect is already observable.
Second minute: open the menu and reach an important page
The mobile menu concentrates several risks because it combines display, animation, links and scrolling management.
You have to open it, select an important section, go back and then close it. The button must remain accessible. The page should not get stuck after closing, nor should it continue to scroll behind the menu when it is open.
On the page reached, the selected link must lead to the correct content. This seems obvious, but a tree change, an old anchor, or a poorly prepared redirect can send the visitor to an empty section or an error.
The test should also check the practical size of the tactile areas. Two links that are too close together are difficult to use with your thumb, even if their display looks clean in a screenshot.
Third minute: complete the main action
For many showcase sites, the main action is a contact form, a quote request, a call or an appointment.
When a form is present, the test must go to the end:
- attempt a submission with an empty required field;
- verify that the error message clearly indicates what to correct;
- complete the form with a request identified as a test;
- send the request;
- control the message displayed after sending;
- check the actual receipt in the mailbox or the planned CRM.
The last point is essential. A “sent” message displayed by the browser does not guarantee that the request has reached the person who should process it. The mail server, a spam filter, a forwarding rule, or an external integration may interrupt the chain after the visual confirmation.
If the main action is a call, clicking on the number should open the dialer with the correct number. For an e-mail, the address and the prepared subject must be checked. To make an appointment, you must reach at least the slot selection screen without loops or errors.
Fourth minute: observe loading and movement
The page should then be reloaded and then navigated up and down at a normal pace.
This passage allows you to identify images that arrive too late, blocks that jump, animations that interfere with reading, buttons that respond with a delay and sections that remain empty. It is useful to redo this check once without Wi-Fi when conditions permit: a comfortable site on the office connection can become painful on an average mobile network.
A performance score helps to diagnose, but the observation retains its own value. Core Web Vitals measure loading, responsiveness and visual stability. They don’t say whether an important button is hidden by a banner or whether the reader understands what to do.
When a WordPress slows down especially on phone, the causes are often divided between hosting, theme, plugins, images and third-party scripts. These families of causes are detailed in this guide to mobile slowness of a WordPress site.
Minute Five: Keep Evidence and Decide
The test ends with a short trace:
- date and time;
- entry page;
- phone and browser used;
- action tested;
- receipt confirmed or not;
- anomalies observed;
- decision made.
A screenshot is often enough to make a defect understandable. For a form, the most useful proof combines the confirmation visible on the site and the receipt of the message. For a navigation problem, a short description of the path allows you to reproduce it.
Each anomaly must receive a simple level:
- blocking: the main action fails, the content becomes inaccessible or the site presents incorrect information;
- to be corrected quickly: the route remains possible but creates significant friction;
- to monitor: the fault is minor or intermittent and requires further observation.
This classification avoids two opposing reactions: putting online despite a serious commercial defect, or blocking a publication for a detail without immediate impact.
What the tools should additionally check
The five-minute human test benefits from being accompanied by automatic checks. They can in particular detect:
- HTTP errors and broken links;
- JavaScript messages in the console;
- overflows to multiple screen widths;
- visual regressions between two versions;
- performance variations;
- certain accessibility errors;
- the absence of an expected follow-up event.
Automation allows the same checks to be repeated. The manual visit provides context: it shows whether the path remains understandable and whether the action really produces its result.
A keyboard check on a computer must also be provided regularly. The menu, links and fields must be navigable in a logical order, with visible focus and without blocking. WCAG 2.2 explicitly addresses keyboard usage, focus order, and minimum size of certain interactive targets. The rapid mobile test alone does not allow a conclusion of compliance.
When to run this test again
The protocol must be repeated after any intervention likely to affect a shared element or an important path:
- updating the CMS, theme or extension;
- modification of the menu, the first screen or a call to action;
- adding a third-party measurement tool, banner or script;
- change of form, antispam or messaging;
- publication of a new service page;
- deployment of a new version;
- change of cache, CDN or hosting.
For a purely editorial modification, the test can remain centered on the affected page. For a technical update or shared component, the full path must be replayed.
A little routine is better than a long forgotten recipe
Most small business showcase sites don’t have a dedicated revenue team. This constraint makes short, reproducible checks even more useful.
Five minutes is not enough to certify a site. They are often enough to intercept a blocked menu, a hidden button, a page that is too wide, a form that no longer delivers its messages or a modification that has moved the problem elsewhere.
The value of the protocol comes from its repetition. A simple test carried out after each change protects the course better than a complete audit carried out once and then forgotten for a year.
