Hosting and uptime: what to look for as a business owner

Hosting is the part of a website you think about least, until something goes wrong. Then it is suddenly the only thing that counts: the site is unreachable, forms do not arrive and customers drop off. In this article we explain what you can expect from a hosting provider, how to read an uptime promise, what an outage does to your findability and which questions to ask before you sign. We name no providers or prices; it is about the questions.

What hosting actually is

Hosting is the computer (or network of computers) your website runs on and where visitors fetch it. Your domain name is the address, the hosting is the building. A hosting package can be just space, but can also include updates, backups, security and support. That difference determines almost everything: two packages can look the same and still deliver something very different.

What “99.9% uptime” means

Uptime is the share of time your site is reachable. The percentages all sound high, but the difference in hours is bigger than you think. Work it out yourself: a year has 8,760 hours.

• 99% allows the site to be down for about 87.6 hours a year, almost four days.
• 99.9% allows about 8.76 hours a year, roughly 43 minutes a month.
• 99.99% allows about 53 minutes a year, about 4 minutes a month.

Note what such a percentage does not say. An uptime promise usually counts only the outages the hosting provider causes itself, often with exceptions for planned maintenance. Who measures the percentage and how is in the terms. So ask: how is uptime measured, what does not count and what do you get if the promise is not met? A promise without a consequence is a marketing sentence.

What an outage does to your search results

In its documentation on HTTP status codes, Google writes that server errors (5xx) and 429 lead Google's crawlers to temporarily crawl less. For pages already in the index, they are initially kept, but they are eventually dropped if they keep failing; URLs that persistently return a server error are removed from the index. As soon as the server responds with a 2xx code again, Google gradually raises the crawl rate again. The message for a business owner: a short outage usually costs little search traffic, but an outage that lasts days, or a structurally unstable site, can certainly cost visibility. There is also the direct damage: a visitor who lands on an error page often does not come back.

For planned maintenance: let the server return a server-error code (such as 503) instead of a page that suggests “all is well” but is empty. Google describes that empty pages or error messages returned with a 200 code are flagged in Search Console as a “soft 404”. Check this with your developer or hosting provider.

Speed starts at the server

The first step in every page load is the response of the server. web.dev calls this Time to First Byte (TTFB): the time between the start of the navigation and the first byte of the response. TTFB consists of, among other things, DNS, setting up the connection, TLS and processing on the server. As a rule of thumb, web.dev says a TTFB of 0.8 seconds or less is good and that values above 1.8 seconds are poor. TTFB is not a Core Web Vital, but it precedes all the other loading metrics: a slow server makes it hard to achieve a good Largest Contentful Paint. More about those metrics in Core Web Vitals explained. A good host helps here in two ways: fast servers close to your visitors and caching, so pages are not rebuilt at every visit.

The requirements your software imposes

If you use WordPress or another system, check that the host supports the current versions. WordPress.org lists as recommended requirements, among others, a current PHP version, a recent MariaDB or MySQL version and HTTPS support. A host that keeps running old versions for years is a risk for your security and eventually also for the working of your site. This is a simple check you can do before you sign: ask which versions are running and who updates them.

HTTPS is a basic, not an extra

Almost every serious host now offers certificates. web.dev explains that HTTPS protects three things: confidentiality, integrity and authenticity. Check that the certificate is renewed automatically and that the host takes responsibility for that. An expired certificate causes a warning screen for your visitors, which feels to them like an unsafe site. More about the basic measures: Website security for business owners.

Backups: they exist, off the server, tested

A backup is only worth something if you can restore it. Ask: how often is a backup made, how long are they kept, are they stored somewhere other than the server itself and has a restore ever been tested? A backup on the same server that fails is gone in a serious outage. Who wins the discussion when a mistake is made and data is lost? That is in the terms.

Privacy, location and data processing agreement

If your site processes personal data (for example through a contact form), a hosting provider is often a processor in the sense of the GDPR. Article 28 GDPR requires processing by a processor to be governed by a contract, and Article 32 asks for appropriate security measures. In practice that means: ask for a data processing agreement, where the servers are located and which security measures the host takes. More on the basics in cookies and GDPR for your website. This is not legal advice; have your situation checked if you are unsure.

Questions to ask before you sign

1. What is included? Updates, security, backups, support, monitoring.
2. How is uptime measured and what is the compensation? And what does not count?
3. How quickly does support respond, and on which channels? And is anyone reachable when it goes wrong?
4. Which software versions are running and who updates them?
5. What do the backups look like and when was a restore last tested?
6. Where are the servers and is there a data processing agreement?
7. How do you move away? Can you take your files and database with you, and what notice period applies?
8. What does it cost on an ongoing basis, and what is the rate after the first period? Low starting prices sometimes get more expensive later.

Monitor it yourself

You do not have to wait for a customer to call you. A simple monitoring service checks your site regularly and sends an alert at an outage. Also switch on Search Console: it tells you when Google has trouble fetching your pages. With us, monitoring is part of hosting and maintenance, so you do not have to look yourself.

Common mistakes

• Choosing on price alone. Anyone who chooses cheap without looking at support and backups often pays the bill at the first outage.
• Not holding the hosting yourself. If everything is in the name of a previous agency or freelancer, you are at risk when they are gone. Make sure the account and your domain are in your name.
• No plan for when it goes wrong. Who calls whom, how is it restored, how long may that take?

In summary

Good hosting is invisible. You only notice it when it is missing, and then you want someone to be responsible. Want to know what suits your site, or have an existing setup checked? Book a free intro call. Also read What does a website cost? for the ongoing costs.

Sources

Reviewed on 28 October 2026 against the sources below. Rules and guidelines change; this article is general information, not legal or tax advice.

• How HTTP status codes affect Google's crawlers (Google Search Central)
• Time to First Byte (TTFB) (web.dev)
• Optimize Largest Contentful Paint (web.dev)
• WordPress requirements (WordPress.org)
• Why HTTPS matters (web.dev)
• Verordening (EU) 2016/679, Algemene verordening gegevensbescherming, artikel 28 en 32 (EUR-Lex)

Book an intro call