A client called me last spring, furious. Not at us — at his own website. His checkout page had been throwing an error for four hours on a Saturday, the busiest shopping day of his week, and the first he heard of it was a one-star Facebook review that said "tried to give you my money, couldn't."
Four hours. On a site doing maybe $600 an hour in sales during that window. Do the math and it stings more than the review did.
"It's probably fine" is not a monitoring strategy
Here's the thing that surprised him: he thought he had monitoring. His hosting provider had a status page. Green checkmark, server up, all good. And technically, it was — the server was fine. The database was fine. One payment library update had quietly broken the checkout flow, and the page still returned a normal response. It just didn't work.
This is the gap almost every small business owner falls into. "Is my website up" and "can a customer actually buy something" are two completely different questions, and most of the tools people use only answer the first one. A ping every thirty minutes telling you the lights are on doesn't tell you whether anyone can get in the door.
The real cost isn't the outage. It's the silence.
I've seen this pattern enough times now that I don't think of it as bad luck — it's a structural blind spot. Nobody budgets for monitoring because nothing about a working site tells you it's needed. Then something breaks — a plugin update, an expired SSL cert, a third-party script that starts timing out — and the business finds out the same way this client did: from a customer, hours later, in the worst possible tone.
The outage itself is rarely the expensive part. Four hours of downtime on a slow Tuesday might cost you fifty dollars. Four hours during a launch, a sale, or your peak traffic window can cost thousands, and you don't get to pick when things break. What actually determines the damage is how long the gap is between "it broke" and "someone who can fix it knows about it." That gap is entirely a function of what you're watching and how often.
What I'd actually check
You don't need anything fancy. You need something that checks the parts of your site people actually use — not just "does the homepage load" but "does the cart complete a checkout," "does the contact form actually send" — and checks often enough that an outage during your busy hours doesn't sit there unnoticed for a quarter of a day. And you need it to tell a person, immediately, not log quietly into a dashboard nobody opens until Monday.
We ended up building SiteBleed for exactly this reason, after watching enough clients discover outages the hard way. It checks the things that actually make you money, not just whether a server responds, and it translates downtime into a dollar figure instead of a green-or-red badge — because "your checkout has been down for 40 minutes" gets ignored, but "you've lost about $400 so far" gets someone's attention in a hurry.
The lesson isn't "buy monitoring software." It's smaller than that: if the first person to tell you something's broken is a customer, you've already paid the most expensive part of the bill. Everything after that is just cleanup.