Why your WordPress site broke on a Sunday evening

The site that goes down on a Sunday evening

If you have been running a WordPress site for 2 or 3 years, you have probably seen this screen before. One Sunday evening, a customer writes to tell you that your page stays blank, or shows “There has been a critical error on this website”. You have not touched anything all day. On Friday, you had simply clicked “Update”.

What follows is always the same. You call your provider, who does not answer on a Sunday, and you wait until Monday morning. The site stays broken for 12 to 16 hours, during which your prospects turn to a competitor and your shop makes no sales at all. And you have no idea what happened, or how to stop it from happening again.

That moment is almost always the mechanical result of the way WordPress is put together: a core, a theme and a stack of plugins written by dozens of different authors, all of which have to get along at every update. It is worth understanding why that stack eventually gives way, and how a site built differently holds up on a Sunday evening just as it does the rest of the week.

Why WordPress ends up breaking

Four causes come up in almost every case.

Updates that collide

A WordPress site in active use rarely runs with fewer than 15 to 25 plugins. Each one follows its own update schedule, is written by a different author and is tested in isolation. When you click “Update”, you are combining code that nobody has ever checked together.

One plugin expects a particular version of PHP, another a particular version of the core, a third modifies a function that the fourth was still using. The day two schedules clash, a page starts returning a fatal error. WPScan recorded more than 8,000 vulnerabilities and incompatibilities reported across the WordPress ecosystem in 2025 alone. Most of them come from this friction between versions.

Fragile or abandoned plugins

The WordPress directory lists around 60,000 plugins. A significant share of them is maintained sporadically, or not at all. A Sucuri study from June 2026 identifies 1,250 active CVE vulnerabilities in plugins still installed on live sites.

An abandoned plugin keeps working until the day a core update changes a rule it no longer follows. It then breaks without warning. You discover the problem when your contact form stops sending, or when your basket refuses to add an item. Nobody is there to fix it, because the author dropped the project two years earlier.

Bloated themes

Many sites rely on a theme bought on an online marketplace and assembled with a visual builder such as Elementor or Divi. These themes pile up hundreds of features you will never use, and each one adds code that has to be loaded and maintained.

The more code the theme carries, the more plugins it crosses paths with, and the more breaking points it exposes. An update to the builder can break a layout. An update to a plugin can break a theme feature. The site becomes a tangle in which nobody knows any more which layer produces which behaviour.

Technical debt that piles up

WordPress carries 22 years of history. Backward compatibility is maintained, which is remarkable, but the price is an accumulation of layers. The PHP engine has evolved from version 4 to version 8.3. A REST API was added on top of the old XML-RPC interface. The block editor coexists with the classic editor. The visual builders each reinvent their own system.

Each layer adds a chance of conflict. On a site that is 3 or 4 years old, nobody knows all the dependencies any more. The Sunday update becomes a gamble that either goes through or breaks, and one day it breaks.

The real cost of a site that breaks over a weekend

A 2025 Gartner study put the average cost of one hour of website downtime at USD 5,600 for an SME. On the Swiss market in 2026, the figures observed are more like CHF 1,000 to 5,000 per hour for an SME, depending on size and sector.

For a typical outage from 8 pm on Sunday to noon on Monday, that is 12 to 16 hours, the direct cost reaches CHF 12,000 to 80,000. On top of that come:

• Indirect cost: prospects who went elsewhere and were never won back.
• Reputational cost: if the outage is visible, your image suffers.
• Technical cost: the emergency call-out billed at a premium rate, often CHF 200 to 400 per hour.
• Emotional cost: the stress, the internal tension, the doubts about your technical choices.

These amounts are systematically underestimated. A site that is “not seriously” broken for one night easily costs CHF 20,000 in total. Over 5 years, two or three episodes of this kind exceed CHF 60,000.

Reliability by design

A site does not break on a Sunday evening when nothing can collide. That is the principle behind a site built without WordPress: you remove the source of the breakages instead of monitoring it.

No third-party plugins that have to coexist

In an Astro or Next.js site, the features that WordPress would hand over to 20 plugins written by 20 authors are written in the project’s own code, in a single language, following a single logic. There are no third-party update schedules to line up, so no collision between versions is possible. The main cause of Sunday-evening outages disappears.

Pages fixed at the moment of publication

On a modern architecture, pages are generated once, when the site goes live, then served as they are. Nothing runs on each visit that could fail under the load of a Sunday evening. What was checked before going live is exactly what your visitors see.

Updates checked before going live

Every change goes through a test version before it reaches the public. You see the site working on this preview, you check it, and only then do you switch over. An error is spotted before your customers can discover it on a Sunday evening, and the Friday deployment stops being a gamble.

A surface reduced to what you actually use

Without a bloated theme or dormant plugins, the site only carries the code for its own features. Less code means fewer breaking points, and a site that a single person can understand in full. When something needs to change, you know exactly which part to touch, and why.

The test to run this weekend

Before the next outage, an SME in French-speaking Switzerland can carry out a simple check. Count your signs of fragility:

1. How many plugins are installed on your site? Beyond 10, the risk of collision rises quickly.
2. How long has it been since they were all updated? A plugin that has not moved for a year is one to keep an eye on.
3. Who tested the last update before it went live? If the answer is “nobody”, every click on “Update” is a gamble.
4. How many times has your site shown an error or a blank page this year? Each episode is a warning.
5. Who answers on a Sunday evening when a page breaks? If nobody does, the next outage will last until Monday.

If you add up three signs or more, your site is holding up out of habit rather than by design, and the next Sunday evening is only a matter of time.

What we do at Voren

The sites Voren delivers are built on a modern stack, without WordPress, so without third-party plugins that have to coexist or a bloated theme: technical debt does not pile up year after year. Pages are generated at publication and checked on a preview before every deployment.

In practice, there is nothing to bring down on a Sunday evening: no combination of plugins can collide, no function runs live that could fail, and the reliability comes from what we removed from the site.

Going further

If your site broke on a Sunday evening recently, or if you want to stop that from happening, three options:

See our business website service: sites built without WordPress, so without the mechanism that breaks.

See our maintenance plan: updates checked before they are applied, and a response within 24 hours.

Request a free diagnostic: an analysis of your current site, a count of your signs of fragility, prioritised recommendations, with no obligation.