How Much Hosting Does a Small Business Website Actually Need?

The most common hosting mistake made by a small business is not choosing the wrong company. It is buying the plan for a problem the site does not have. Someone reads that WordPress needs a lot of resources, sees a tier labelled for growing businesses, and pays three times the entry price for headroom they will never use. The site runs fine. The invoice does not.

Sizing is arithmetic, and it gets easier once you stop estimating traffic from the feeling that the site is busy. Three numbers decide almost every small site: how much disk the installation occupies, how many concurrent requests the site has to survive, and whether the plan includes staging and email. None of them is visits, which is the number plan cards put in the largest type.

Measure your site instead of guessing at it

Storage first, because it is measurable today. A WordPress core install is small. Themes and plugins are usually a few hundred megabytes together unless you have installed several page builders. The database for a site with five hundred posts is typically under a hundred megabytes. The part that grows is the uploads directory, which is images, and only images.

On a stock install the arithmetic is so one-sided that the entry tier is rarely the constraint. SiteGround's StartUp tier includes 10 GB of storage. A five-page business site with a logo, a dozen stock photographs and a handful of plugins sits comfortably under 2 GB, and much of that is cached thumbnails a plugin generated rather than anything you uploaded. Ten gigabytes is years of headroom for a site like that, not a limit you will touch.

Transfer is equally modest. If a page weighs 1.5 MB including images and fonts, ten thousand page views in a month is about 15 GB. SiteGround lists unmetered data transfer on StartUp and describes the monthly visit figure as a guide rather than a cap, which is worth reading carefully: unmetered transfer is not the same promise as unlimited CPU, and the guide is the vendor telling you what the machine expects. For a brochure site, both numbers are far above what you will use.

If you want to check your own figures rather than trust a rule of thumb, two free tools do it without touching the live site. WordPress Playground runs a WordPress install in a browser tab, which is enough to see how a stock site plus your plugin list behaves. Local installs WordPress on your own machine and will import a copy of the real database, so you can look at the actual uploads size in the filesystem instead of adding up image files by hand. Neither costs anything and neither is a hosting recommendation.

The limit you actually hit is concurrency, not visitors

Visits are a proxy for load, and a poor one. What a server does per request matters more than how many requests arrive. A cached page is served from memory and costs almost nothing. An uncached page with twelve database queries costs a PHP worker for the length of those queries, and PHP workers are the resource shared plans meter quietly.

That is why two sites with identical traffic have different hosting needs. A blog with static posts and a page cache serves ten thousand visits a month from the entry tier without noticing. A store with a logged-in checkout, a dashboard widget and an uncached cart fragment does not. The second site is not bigger in storage; it is heavier per request, and the fix for that is either caching, a plugin removal, or a plan with more workers, roughly in that order of expense.

The embarrassing version of this problem is a plugin nobody uses any more. A gallery extension left active on a five-page site, or a related-posts module running a query on every page view, can consume more of a plan than genuine traffic ever will. Before you conclude that the site has outgrown the hosting, open the query monitor on a staging copy, or at minimum count the plugins and check when each was last updated. Unused plugins are free to delete and cost real money to keep.

Feature requirements, not traffic, decide the tier

When a small site does need to move up a tier, the reason is almost never bandwidth. It is a feature that starts on a higher plan. The clearest example on SiteGround is staging, which is listed on GrowBig and GoGeek but not on StartUp. A staging copy is how you test a plugin update, a PHP version change or a theme rewrite without risking the live site. If someone maintains the site professionally, that feature is not optional, and it costs the difference between $2.99 and $4.99 a month on the introductory rate. The renewal difference is $17.99 against $29.99, which is still less than one hour of a developer's time per year.

Email is the other quiet decider. SiteGround lists free email accounts on all three hosting tiers, so a small business can keep its addresses with the hosting. Kinsta does not provide mailboxes at all; its documentation covers pointing MX records at Google Workspace or Microsoft 365, which means the true cost of that platform includes a mail subscription. Neither approach is wrong, but a comparison that prices the hosting and forgets the mail is comparing half a product.

Backups belong in the same conversation. Daily automated backups are standard on SiteGround's plans, and elsewhere they vary from hourly to weekly. For a site where the content changes twice a month, daily is generous. For a store, ask how far back the retention goes, because the failure you are protecting against is often discovered a week late.

Where the small end of SiteGround is the answer, and where it runs out

For a small business site that needs WordPress, email on its own domain, daily backups, SSL and a control panel a non-developer can use, this is where to buy it: SiteGround's WordPress hosting lists StartUp at $2.99 a month for the first year, renewing at $17.99, and GrowBig at $4.99 then $29.99, both prepaid for twelve months and excluding VAT. Buy StartUp for a brochure or a blog. Buy GrowBig if you need staging or you are hosting more than one site, because StartUp covers one website and GrowBig does not cap the number.

Advertisement · SiteGround

SiteGround runs out of room in three specific places. A store with several hundred products and a checkout that runs uncached PHP needs dedicated workers, and no amount of shared hosting will make that comfortable; a managed plan with a published visit allowance and a stated overage rate is the honest answer there. A site whose traffic arrives in newsletters and seasonal bursts wants a plan where the peak is priced rather than tolerated, because on shared hosting the consequence of a spike is a suspended account rather than an invoice. And a site that needs root access, a specific compiled extension or a database larger than a shared plan allows has outgrown this class of hosting entirely.

There is also a cheaper answer, and it should be named. If the site is genuinely small and you are willing to commit to a long term, Hostinger's WordPress Premium plan is $2.99 a month on a 48-month term and renews at $10.99, with Unlimited at $3.99 renewing at $16.99. The catch is the commitment: four years is a long time to be wrong. If you would rather not commit, Kinsta's entry Business tier is $35 a month or $350 billed annually, about $30 a month, with no renewal increase and a published overage of $0.50 per 1,000 visits past the allowance. That is not a small-site plan, and for a site with fewer than ten thousand visits it is an expensive way to buy a guarantee.

And for the smallest sites there is a free option worth taking seriously. If the site does not need PHP or a database, a static export hosted on a free static host costs nothing and will outperform any shared plan. That is not most business sites, but it is more of them than the hosting industry would like to admit. The test is simple: if nothing on the site changes per visitor and you have no forms or logins, static hosting is enough.

Buy the tier you can justify, then watch one number

1. Measure the uploads directory and the database before you choose a storage tier. Most small sites are under 2 GB, which is one fifth of SiteGround's entry tier.
2. Decide whether you need staging and whether hosting carries your email. Those two answers pick the tier; the visit number does not.
3. If a page feels slow, check the plugin list and the page weight before upgrading. Removing one unused plugin is cheaper than moving up a plan.
4. If you do upgrade, write down the limit you expect to clear. A tier bought for a named constraint is a decision; a tier bought for a feeling is a subscription.
5. Track peak concurrent visitors once a month, not total visits. That is the number that tells you when the next tier is actually needed.

Everything above is arithmetic you can redo with your own numbers in ten minutes. The prices were on the vendors' own pages on 16 October 2026, introductory rates expire, and the renewal line is the one to read before you pay. Two hosts that a reader will encounter while shopping could not be checked from this machine that day, so they are named as categories rather than quoted at a price, which is the only honest way to write a comparison you cannot finish.