Hosting is the one part of a website nobody thinks about until it fails, and by then the choice has already been made. Four things decide whether a plan is genuinely cheap or simply underpriced: whether the site stays up, whether the backups exist and can actually be restored, where in the world the server sits, and whether a human answers when something breaks at four in the afternoon on a Friday. Price tells you almost nothing on its own.
Uptime, and the number that sounds fine
Most hosts advertise 99.9 percent uptime. That sounds close enough to perfect that people stop reading.
It is about nine hours of downtime a year. If those hours land on a busy weekday afternoon, that is most of a trading day with the site unreachable.
The number worth asking for is not the promise but the record. A host that publishes a status page with history is telling you something. A host that only publishes the promise is telling you something too.
Backups that have actually been restored
Almost every plan says it includes backups. Far fewer can restore one quickly, and fewer still have ever been asked to.
Three questions settle it. How often is a backup taken. How far back do they go. And how long does a restore take, in hours, when you ask for one.
A backup you cannot restore inside a working day is a filing cabinet, not a safety net. Ask the host to restore one to a test location before you need it for real.
Where the server is matters more than it sounds
Distance costs time. A visitor in Toronto loading a site hosted in Singapore waits for every request to cross the planet and come back.
For a business selling locally, hosting inside the same country removes a delay that no amount of optimisation gets back.
There is a second reason that matters in Canada. Depending on the sector and the data being collected, keeping customer data in the country is either a preference or a requirement, and it is easier to start that way than to move later.
Support is the part you are really buying
When the site is down, the only question that matters is how long until somebody who can fix it is looking at it.
Test it before you commit. Send a question at an awkward hour and see what comes back and how fast. A reply in twenty minutes and a reply in two days are different products at the same price.
Find out whether you can reach a person by phone at all. Ticket only support is fine for a slow question and miserable during an outage.
The parts that are quietly missing
Read what the cheapest plan excludes rather than what it includes. The exclusions are where the price came from.
Common ones: backups charged separately, SSL charged after the first year, a renewal price two or three times the first year rate, hard limits on visits or on database size, and an email service that is either absent or barely usable.
Migration is worth asking about too. A host that will move your site for you is saving you the most fragile hour in the whole process.
Questions people ask
- Is shared hosting good enough?
- For most brochure sites and small stores, yes. Shared hosting becomes the problem when traffic grows, when the site does heavy work on every page, or when a neighbour on the same server misbehaves. The honest signal to move is the site slowing down at busy times, not a salesperson saying you have outgrown it.
- Should my email be with my hosting?
- It can be, and for a small business it often is simplest. The thing to check is deliverability: whether SPF, DKIM and DMARC are set up properly. Mail from a domain without them is far more likely to land in spam whoever hosts it, and the large mail providers now refuse some of it outright.
- How hard is it to move host?
- Less than people fear, if it is planned. Files and database move first, the site is tested on the new server before anything changes, and only then does DNS point across. Done in that order there is no gap. Done in the other order there is.
We do this
If you would rather have it done than do it, this is the part of what we offer that covers it.