How to Choose Web Hosting (and Dodge the Renewal Trap)

Nearly every "best web hosting" list sells you on a headline price — "$2.99 a month!" — that is one of the most misleading numbers in tech. It's an introductory teaser that renews at three to four times the price once your site is live and moving feels like a hassle. The two decisions that actually matter aren't captured by that number at all: what type of hosting your site needs, and what it'll truly cost over a few years. Get those right and hosting is cheap and boring; chase the teaser price and you'll overpay for years. This guide covers both.

Our full terms are on our disclaimer page.

Who this is for, and how we chose what to cover

This is for someone buying hosting for a small site or blog — a first site, or a move away from a host that has become expensive. It assumes no server administration experience.

How we chose what to put in this guide:

• We organise around the type and the renewal price, because those two decisions account for nearly all of what a reader will pay and nearly all of whether they are happy in two years.
• We do not rank hosts or test them. We have run no uptime measurements, and the ones you find in "best hosting" lists are almost universally affiliate-monetised.
• We name hosts with links, because "pick a shared host" is not something you can act on without knowing which ones exist.
• We give the measurement methods — how to check uptime and speed yourself — because every host claims excellent numbers and only your own measurement is evidence.

What you need before you buy

The four types (and who each is for)

Most people starting a blog or small business site do not need VPS, cloud, or dedicated hosting. Shared hosting (or managed WordPress if you're on WordPress) is the right, cheap starting point — you can upgrade later when traffic actually demands it. If you're setting one up from scratch, our guide to starting a blog that can actually pay covers the decision that comes before this one.

Why shared hosting slows down under load — the part the sales page won't explain. On a shared plan you're allocated a small number of PHP workers: the processes that build each page before it's sent. If your plan has two workers and three visitors request pages at once, the third waits in a queue. At low traffic nobody notices. As traffic grows, that queue is what people experience as "the site got slow" — and it's the real signal that you've outgrown the plan, far more useful than any raw visitor-count limit. It's also why "unlimited bandwidth" can be true and irrelevant at the same time.

The point: buy for the site you have, not the traffic you dream of. Start on shared or managed WordPress and upgrade when the queue tells you to, not preemptively.

The flagship: match the type, then compute the true cost

Step 1 — pick the type (above): shared/managed-WP for most new sites; VPS/cloud only when you have real traffic or control needs.

Step 2 — compute the true multi-year cost, not the teaser. This is where hosts make their money. That "$2.99/month" almost always requires paying 1–3 years upfront, and then renews far higher. Run it (illustrative, computed): a plan at $2.99/month for year one that renews at $11.99/month costs about $36 the first year, then ~$288 over the next two — a 3-year total of ~$324, a true average of ~$9/month, or 3× the sticker price. The advertised number is real for exactly one term; the renewal is what you'll actually live with.

Bottom line: decide the type first, then look up the renewal price and multiply it out over 2–3 years before you buy. The host with the lowest intro price is frequently not the cheapest once it renews — compare true multi-year costs, not teasers.

A worked example: the cheaper host is the more expensive one

All figures below are hypothetical model inputs, not any host's prices. They exist to show the arithmetic, which is the part you will reuse.

Suppose you are choosing between two shared plans for the same small site:

Host B costs $60 less over three years while advertising a price twice as high. That is the entire thesis of this article in one table, and it is why the comparison sites — which sort by advertised price — systematically point readers at the more expensive option.

What this assumes, and what would change it. It assumes you stay three years, which most people do because migrating is a chore; if you genuinely intend to leave after twelve months, Host A wins and you must actually leave. It ignores the first-year "free" domain, which typically renews at $15–$20 and shifts both columns equally. It assumes no add-ons for SSL or backups — if Host A charges for either and Host B includes them, the gap widens further.

Run this yourself before buying. Two numbers per host — the intro rate and the renewal rate — and one multiplication. It takes five minutes and it is the highest-value five minutes in the purchase.

The hosts, named and unranked

Seven a US buyer will encounter, alphabetically, not ranked, not tested by us, with no price, uptime or performance claim asserted here. Each link goes to that company's own site, where its current pricing — and, crucially, its renewal terms — live:

• Bluehost — shared and WordPress-oriented
• Cloudways — managed cloud hosting on third-party infrastructure
• DreamHost — shared, VPS and managed WordPress
• Hostinger — shared and VPS, aggressive introductory pricing
• Kinsta — managed WordPress, premium end
• SiteGround — shared and managed WordPress
• WP Engine — managed WordPress, premium end

What to compare across them: the renewal price (not the intro); whether SSL, daily backups and email are included or charged; the term length required for the advertised rate; whether migration is free; and where the servers are relative to your readers, since distance is a direct input to the speed section below. Everything else on a hosting sales page is noise.

The traps hiding in the fine print

• The renewal hike. Intro rates commonly renew 200–400%+ higher. Always find the renewal price (it's usually buried) before committing.
• "Unlimited" isn't. "Unlimited" bandwidth/storage plans have real caps in the terms; a host can label ordinary growth as "abuse" to push you to upgrade.
• Hidden add-ons. SSL certificates, daily backups, email accounts, and site migration are sometimes extra — and a "free domain" is often free for the first year only, then $15–$20/year.
• Long lock-ins. The cheapest rate usually needs a 36-month prepay. That's fine if you've verified the renewal and the host is good — risky if you haven't.

What to check: the renewal price, what's genuinely included (SSL, backups, email), the real bandwidth terms, and the commitment length — before the intro price seduces you.

What actually matters in a host

Beyond price, three things decide whether you'll be happy in two years: uptime, speed, and support you can actually reach when the site is down. The first two are measurable, and you don't have to take anyone's word for either — including ours.

Uptime, in hours rather than percentages. A year is 8,760 hours, which makes the marketing percentages concrete:

That's why the gap between 99.9% and 99% matters so much more than the single digit suggests: it's the difference between an afternoon offline and the better part of a working week. Aim for a 99.9% track record or better — and note that a published SLA is a promise, not a measurement.

How to check it yourself. Independent uptime monitors track real availability for major hosts, and free tiers will monitor your own site once it's live. This is worth doing precisely because every host claims excellent uptime — the number that matters is the one you or a third party measured, not the one on the sales page.

Speed, and why hosting is most of it. The metric to know is Time to First Byte — how long the server takes to start responding, before any of your page has loaded. web.dev's explainer covers what it measures; your host is the single largest input to it, which is why a cheap overloaded server can't be fixed with a faster theme. It matters commercially too: Google treats page experience, measured through Core Web Vitals, as a ranking-relevant signal. You can measure your own site's TTFB free in any browser's developer tools or a public speed test.

Also weigh: support you can reach out-of-hours, whether SSL and backups are genuinely included, whether email accounts are bundled or extra (if email matters to your business, a dedicated email platform is a separate decision), and how painless the upgrade path is.

Our take: a reliable host at a fair renewal price beats a cheap teaser from a host with poor uptime and support — you're buying years of service, not one promotional month.

How to check it's actually good, once you're live

Every host claims excellent uptime and speed. These are the three checks that replace the claim with a measurement, and all three are free.

1. Put your own site on an uptime monitor. Free tiers — UptimeRobot is the commonly used one — check your site every few minutes and email you when it stops responding. Do this in week one. Without it, you find out your site was down because someone tells you, or because traffic quietly fell.
2. Measure your Time to First Byte. Open your site in a browser, then the developer tools' network panel, and look at the waiting or TTFB figure for the main document. Do it from a normal connection, not on the same network as anything unusual. This is the number your host controls; a theme change will not fix a slow one.
3. Run PageSpeed Insights on your own URL. It reports Core Web Vitals — the same page-experience signals Google describes — and separates what the server contributes from what your page does. If server response time is flagged, the finding is about your host; if layout shift is flagged, it is about your theme.

What a fail looks like, and what to do about it. Repeated short outages across weeks, or a TTFB that is consistently poor from an ordinary connection, is a host problem and neither a plugin nor a caching layer will durably fix it. A slow site under load specifically — fast when quiet, slow when busy — is the PHP-worker queue described above, and the fix is a bigger plan or a better host, in that order.

The test at six months: can you state your uptime for the last month and your TTFB from a number rather than an impression? If yes, you can tell whether a renewal is worth paying. If no, you will renew on inertia, which is what the pricing model is designed for.

Moving hosts later

You will probably do this once, and knowing the cost in advance changes what you buy now.

What migration involves: copying the site files and database, pointing the domain's DNS at the new host, re-issuing the SSL certificate, and re-checking email if it was hosted alongside. Many hosts perform the whole migration free as an acquisition incentive — ask before you pay anyone, and confirm it is free rather than "free with a plan".

What it costs you in practice: an afternoon, plus a DNS propagation window during which some visitors reach the old server and some the new. Keep the old plan running for a few days after switching, and do not cancel it until you have confirmed the new one is serving everything.

The thing to do now, before any of this: register your domain somewhere other than your host. A domain held at the host you want to leave is the single most effective lock-in in this industry, and separating them costs nothing at registration time.

If shared hosting isn't the answer

Bottom line: for a new small site the correct answer is nearly always shared or managed WordPress, and the interesting question is not which tier but which host's renewal price you can live with.

Common mistakes

• Buying on the intro price. It's the trap; compute the renewal-inclusive cost.
• Over-buying hosting you don't need. Most new sites don't need VPS or cloud — start small.
• Ignoring "free for the first year" domains and SSL that quietly add to the renewal bill.
• Committing 36 months without checking the renewal price or the host's uptime reputation.
• Assuming "unlimited" is truly unlimited — read the caps.

Putting it together

Hosting is a simple decision made confusing by teaser pricing. Choose the type that fits your site — shared or managed WordPress for almost every new site, upgrading to VPS/cloud only when real traffic demands — then ignore the headline price and compute what you'll pay over two or three years including the renewal. Check uptime, support, and what's actually included, and don't commit to a long term until you've seen the renewal number. Do that and you'll pay a fair price for reliable hosting instead of a cheap month followed by expensive years.

Your next three moves, in order: (1) find the renewal price for two hosts and run the three-year table above — five minutes, and it frequently reverses the obvious choice; (2) register the domain somewhere other than the host you pick; (3) put the site on a free uptime monitor in week one, so your next renewal decision rests on a measurement rather than a memory.

Where to go from here

• If you are setting the site up from scratch, starting a blog that can actually pay covers the own-versus-rent decision that comes before this one.
• If email matters to your business, it is a separate purchase from hosting — choosing an email platform covers it, including the DNS records you will need access to.
• To measure rather than trust: PageSpeed Insights for your own site, and web.dev's TTFB explainer for what the number means.

Our full terms are on our disclaimer page.

FAQ

(Only questions the body doesn't fully answer.)

• Can I switch hosts later? Yes — most sites can be migrated, and some hosts do it free. But it's a hassle, which is exactly why hosts count on you staying through the pricey renewal. Picking a fair-renewal host up front saves the trouble.
• Do I really need managed WordPress? If you run WordPress and don't want to handle updates, backups, and security yourself, it's worth the premium; if you're comfortable managing those, plain shared hosting is cheaper and fine to start.
• What uptime should I expect? Look for a 99.9%+ track record. "99.9%" is about 8–9 hours of downtime a year; dropping to "99%" is over three days a year — the extra nine matters.
• Is the cheapest plan ever the right call? Only after you've checked the renewal price and the host's reliability. The lowest intro price from an unreliable host with a steep renewal is usually the most expensive choice over time.