When Your Website Breaks, Everyone You Pay Points Somewhere Else

Here is the uncomfortable truth about the website your business runs on. The day it breaks, there is a very real chance that nobody you are paying considers fixing it to be their job.

That is not cynicism. It is arithmetic. Most small business websites are not one thing bought from one company. They are five or six separate things, bought from five or six separate companies, and stitched together by someone who has long since moved on. Your domain sits with a registrar. Your site sits with a host. The contact form is one plugin, the booking calendar another, the backups a third. Somewhere there is an SSL certificate, an analytics tag, perhaps a security plugin bolted on after a scare. The person who wired it all together was a developer, a relative who was "good with computers", or a free weekend three years ago.

Every one of those companies wrote a contract. Each contract quietly ends at the edge of its own little box. When the whole thing falls over, the gaps between the boxes are exactly where your Saturday disappears.

I have spent twenty years building and running hosting infrastructure in Ireland, and the most common serious failure I see has nothing to do with hackers or hardware. It is the accountability gap. The empty space between vendors where a problem lives and no one will claim it.

You did not buy a website. You assembled a supply chain.

The web hides this well, because the result looks like a single thing. One address, one logo, one set of pages. Under the bonnet it is a supply chain, and a longer one than most owners ever realise.

The HTTP Archive's Web Almanac, which measures millions of real websites every year, found that more than nine in ten web pages now load at least one third party, and the typical page pulls in something in the region of eighty separate requests. A font loaded from one company, a map from another, a chat bubble, an analytics script, a payment button, each one phoning home to a different supplier you will never speak to.

Now add the parts you actually pay for: the registrar holding the domain, the host serving the pages, the plugin developers (often a different one for every feature), the certificate authority, and maybe a separate backup service, because the host's own backups turned out to be thin. None of that fragmentation is purely an accident of how you shopped, either. It partly reflects how the hosting market itself has consolidated and quietly narrowed what is actually on offer to smaller buyers.

In a large organisation, one role holds all of this together. Someone whose actual job is to understand how the pieces fit, and to own the whole when it misbehaves. Most small firms do not have that person, and the data is blunt about it. According to Eurostat, only around one in seven small enterprises across the EU employs an ICT specialist of any kind, roughly fourteen percent, and that share has barely moved in years. ICT specialists make up about five percent of the entire European workforce. Read it plainly: the person whose full-time job would be to own your technology stack almost certainly does not work for you.

So when the supply chain breaks, the integrator role lands on whoever happens to be closest. That usually means the owner, mid-service, with no idea which of six suppliers is at fault and no real pull with any of them.

The runaround is built in, not bad luck

Picture a craft brewery in Kilkenny that sells online. On the busiest weekend of the year, the checkout stops taking payment. The owner rings the host. The host checks the server, finds it healthy, and says the problem is a plugin. The owner hunts down the plugin vendor's support form, fills it in, and gets a reply two days later suggesting the host is running the wrong version of PHP. The developer who built the site answers a text on Sunday night to say they only did the build, hosting is not their department, and have you tried deactivating things one by one.

Nobody in that chain is lying. Each is correct about their own box. And the checkout is still dead, the weekend is gone, and the orders that did not happen are not coming back on Monday.

That is not a streak of bad luck. It is the predictable behaviour of a system with no single owner. Every supplier is quietly incentivised to prove the fault lives in someone else's box, because the moment one of them accepts it, the problem becomes their cost and their work.

I learned this the unglamorous way. Years ago I assumed that if every component in a setup worked on its own, the assembled whole would work too. Then I watched a perfectly healthy site go dark because two well-reviewed plugins disagreed about a single shared function, and not one of the three vendors involved thought the conflict was theirs to resolve. The lesson stuck. A system nobody owns end to end is a system that, sooner or later, fails with nobody there to catch it.

The technical name for what took that site down is a dependency conflict: two pieces of software, each fine alone, colliding over something they share. In a stack with one owner, that conflict gets caught in a staging environment before it ever reaches production, and rolled back within minutes if it slips through. In a stack with six owners, it gets caught by a customer, on a Saturday, ringing to tell you the checkout is broken.

"Managed" is meant to mean somebody owns the whole thing

What does the alternative actually look like? Not a single magic product. A standard. Before you judge any platform, ours included, here is the bar a genuinely accountable setup has to clear.

One company runs the infrastructure, the application layer, the backups and the security, rather than just one slice of it. A single support line that can resolve a problem instead of routing it sideways. The provider that hosts the site is also the one that can restore it, patch it, and tell you with confidence whether the fault is theirs or yours. And, crucially, they employ people who genuinely do operations, rather than reselling someone else's server and forwarding your ticket when it starts to smoke. If a host sells you "managed" but cannot say who patches the server or where your backups physically live, it is selling you a word.

That standard is the entire reason we built Web60 the way we did. The WordPress runs on our own Irish infrastructure, on a tuned stack of Nginx, PHP-FPM, Redis and FastCGI caching, which in plain terms means a customer on a phone is not watching a spinner while your homepage decides whether to load. Nightly backups run automatically with one-click restore, and a safety snapshot is taken before any update, so a bad change costs you a day at worst rather than the business.

SSL certificates are issued and renewed for you, so your site never greets a customer with a red "not secure" warning on the morning a certificate quietly expired. Hardening, fail2ban and malware scanning sit at the server level, run by the same team that answers the phone. And when something does go wrong, that phone is answered by an Irish team of real people who can see the entire stack, not an outsourced queue reading from a script. It is one team standing behind the whole thing, for a flat sixty euro a year with nothing held back for "premium" tiers.

Notice what actually changes. There is no longer an edge between boxes for a problem to hide in, because there is one box. When the checkout breaks, there is a single number to ring, and the person who answers it can fix all of it.

None of this means surrendering the open platform, by the way. WordPress is not a walled garden you get sealed inside. W3Techs puts it at roughly 41.5% of all websites, more than four in ten, and close to sixty percent of every site running on a recognised content management system, though I would treat any tidy single number about the entire web with a degree of caution.

The point holds regardless. It is the most widely used website platform on earth, your content stays yours, and consolidating who runs it is not remotely the same thing as being trapped by who runs it. Plenty of owners are quietly folding their patchwork back into one managed platform for exactly this reason, and the switch tends to be far less painful than the lock-in horror stories suggest.

The honest limits of one throat to choke

Let me be straight about what consolidation does not do, because any platform claiming to own absolutely everything is overselling, and you should distrust it on principle.

One accountable provider cannot save you from yourself. Install a pirated "nulled" plugin off a forum, hand your admin password to a stranger on a freelance marketplace, or paste in custom code you do not understand, and you have reached well past the edge of what any host can reasonably own. The accountability is real, but it covers the platform, not every decision made on top of it. Verify who you hand access to. That part stays your job, on any platform, forever.

There is also a genuine case where the patchwork is perfectly fine. If you run a single static page that never takes a payment or a booking, and you happen to have a capable technical person who owns the whole thing end to end and answers the phone when it matters, a self-assembled setup on cheap hosting will serve you well. The patchwork is not dangerous because it has many parts. It becomes dangerous when it has many parts and no owner. If you already have that owner, you have solved the problem this article is about. Most small firms have not, and pretending otherwise is where the trouble tends to start.

The question to ask before anything breaks

You never find out whether your website has an owner on a calm Tuesday. The truth surfaces at the worst possible moment, with something down and money walking out the door, when you start working the phones and discover how many of them there are. By then it is far too late to change the answer.

So ask it now, while nothing is on fire. If my site goes down this weekend, who do I call, and can that one person actually fix all of it, or will they reach for someone else's number? If the honest answer involves three companies and a hope that they choose to cooperate, you do not have a website problem waiting to happen. You have a supplier problem you simply have not been billed for yet.

The fix is not more vendors, and it is not a better spreadsheet to keep track of them. It is fewer places where responsibility can fall through the cracks: one stack, one team that owns it, and one number that ends the call instead of forwarding it. Whatever platform you land on, that is the standard worth holding it to.

Sources

W3Techs, Usage Statistics and Market Share of WordPress

Eurostat, Rising Share of ICT Specialists Among Employed People

HTTP Archive, Web Almanac 2025: Third Parties