An uptime guarantee is a refund policy rather than a performance promise. It defines a percentage, excludes most of the events that actually cause outages, and compensates in service credit that you generally have to notice and claim yourself. Useful as a signal of how seriously a provider takes availability, and close to worthless as protection against the cost of being offline.

What the percentages mean in real time

The figures are quoted in a way that makes them sound close to identical. Converted into time, they are not.

GuaranteePermitted downtime per yearPer month
99%About 3 days 15 hoursAbout 7 hours
99.9%About 8 hours 45 minutesAbout 43 minutes
99.95%About 4 hours 22 minutesAbout 22 minutes
99.99%About 52 minutesAbout 4 minutes

The common offer is 99.9 percent, which permits most of a working day offline across a year without breaching anything. Whether that matters depends entirely on when it happens. Eight hours spread across a year in short overnight increments is genuinely unimportant. The same eight hours in one block during your busiest week is a serious revenue event, and the guarantee treats those two identically.

What the guarantee excludes

This is where most of the substance sits, and it is in the terms rather than the marketing.

Read together, these exclusions remove a large proportion of the events that actually take small business sites offline. The guarantee covers the narrow case of the provider's own hardware failing unexpectedly, which is real and is not the whole picture.

How outages are measured

Almost always by the provider, using their own monitoring, on their own definition of unavailable.

That definition usually means the server not responding. A site returning a database error, serving blank pages, running so slowly that visitors leave, or presenting an expired certificate is frequently not counted as downtime at all, because the server responded.

The practical consequence is that a period your customers experienced as the site being broken may not register as downtime for guarantee purposes. If you want an independent view, you need your own monitoring, which is worth having anyway for the reasons set out in when your host goes down and nobody tells you.

What you actually receive

Service credit, not compensation. A breach typically returns a percentage of that month's hosting fee, applied to a future invoice.

Put concretely: a business on modest hosting suffers a four-hour outage on a Tuesday morning, loses a day of enquiries, and is entitled to a credit worth a few dollars against next month. The compensation and the loss are not in the same order of magnitude and were never intended to be.

Claims also usually have to be made by you, within a stated window, with evidence. Credits are rarely applied automatically. A guarantee nobody claims costs the provider nothing, which is part of why they are offered so freely.

What it is genuinely useful for

Not as protection, but as a signal. How a provider talks about availability tells you how they operate.

The question that replaces it

Rather than asking what the guarantee is, ask what happens during an outage. Who notices, how quickly, who acts, and how you are told.

A provider whose answer is that they are alerted, they investigate immediately, and they contact affected customers is describing something with actual value. A provider whose answer is a percentage and a credit policy is describing a refund process.

The distinction matters because the cost of being offline is almost entirely in lost enquiries and customers who went elsewhere, and no service credit has ever addressed either. What reduces that cost is a shorter outage, which comes from monitoring and a response, not from a clause.


Frequently asked questions

What does 99.9 percent uptime actually mean?

About eight hours and forty-five minutes offline per year, or roughly forty-three minutes a month, without the guarantee being breached. Whether that matters depends entirely on when it falls.

Is a 99.9 percent uptime guarantee good?

It is the common industry offer rather than a distinguishing feature. What matters more is what the guarantee excludes, how downtime is measured, and what the provider actually does when something goes wrong.

What is excluded from an uptime guarantee?

Typically scheduled maintenance, anything attributed to your own software or configuration, third-party and network failures, denial of service attacks, and a broad clause for events outside the provider's control.

What compensation do I get for downtime?

Service credit against a future invoice, usually a percentage of that month's fee, and usually only if you claim it within a stated window. It is not intended to cover business losses and does not come close to doing so.

How is downtime measured?

By the provider, using their own monitoring and their own definition. That definition usually means the server not responding, so a site serving errors, blank pages, or an expired certificate may not register as downtime at all.

Should I have my own uptime monitoring?

Yes. It gives you an independent record, it catches failures a provider's definition misses, and it means you find out before a customer does, which is the part that actually reduces the cost.

West Coast Media Solutions Inc. provides web design, web development, hosting, digital marketing, and business consulting to organisations across Canada, drawing on more than twenty-five years in the field.

Want hosting where somebody notices before you do?

We monitor every site we host and respond ourselves, because a shorter outage is worth considerably more than a service credit.

Start a Conversation