How to choose a web host
Most hosting advice is a list of features. Here is a way of deciding which of those features you actually need.
Choosing hosting is mostly an exercise in ignoring things that do not apply to you. Every host lists thirty features; perhaps five of them will change your experience.
Here is the order worth thinking in.
Step 1: What are you actually building?
This narrows the field more than anything else.
| What you are building | What that rules in |
|---|---|
| Brochure site, rarely changes | Any decent shared host |
| WordPress site with real traffic | WordPress-tuned shared, or managed WordPress |
| Shop taking meaningful revenue | Managed cloud, or a hosted platform |
| Web application | VPS, cloud, or a platform-as-a-service |
| Portfolio, minimal maintenance | A builder, or simple shared hosting |
| Many client sites | Reseller hosting, or managed with agency tooling |
Step 2: Where are your visitors?
Physical distance costs milliseconds that no amount of optimisation recovers. If your audience is in the UK, host in the UK or Western Europe. This is one of the few decisions that is both cheap to get right and annoying to fix later.
It matters more than most feature comparisons, and it is routinely ignored because every host's homepage looks the same regardless of where its servers are.
Step 3: Filter on the things that are expensive to change later
Some decisions are easy to reverse. Others are not. Weight your attention accordingly.
Hard to change later:
- Whether you can export. A hosted builder like Wix or Squarespace has no clean export path. That is fine if you never leave, and a serious problem if you do.
- Email arrangements. If your host does not do email, you need another provider — and moving mailboxes later is genuinely disruptive.
- Control panel. cPanel is portable knowledge and makes migration straightforward. A proprietary panel is often nicer to use and slightly harder to leave.
Easy to change later:
- Theme, plugins, most configuration
- Which plan tier you are on
- Adding a CDN
Step 4: Check the things people forget
- Renewal price, not the intro price — see how much hosting should cost
- Backup retention and whether restores are self-service
- Staging, if you will ever change anything on a live site
- Support channels and hours — chat-only at 2am is not the same as a phone number
- Whether SSL, email and migration are included or extra
Step 5: Ignore these
- "Unlimited" anything. It is not unlimited; it is governed by an acceptable use policy. Read the resource limits instead.
- Uptime percentages in adverts. Everyone claims 99.9%. It is unverifiable and the SLA credit is usually trivial.
- Awards and badges. Frequently paid placements.
- Review sites that rank purely by commission. Including, if we are honest, a lot of this genre — which is why our scoring methodology is published and commission is excluded from it by design.
A shortcut
If you want to skip the framework, the Hosting Finder asks six questions and applies the same logic — what you are building, expected traffic, what you care about, technical comfort and budget — then shows you why each provider matched, including its downsides.
Before you commit
- Check the money-back guarantee length — it is your escape hatch
- Confirm migration is included if you have an existing site
- Confirm what happens to email
- Diarise the renewal date the day you buy
- Take your own backup regardless of what the host provides
That last point is not cynicism about hosts. It is that backups you control are the only ones you can be certain about, and the cost of being wrong is your entire site.