Performance Buying Guide
Every shared host says fast. Here is how you check
Two things decide shared hosting speed, and only one of them appears on the pricing page.
The short answer
Test the claim rather than reading it. Choose shared hosting where every layer is the quick option — NVMe, LiteSpeed and current PHP compound into response times that do not feel shared at all — and confirm it with four measurements you take yourself inside the refund window.
Below: the four readings, what each one exposes, the benchmark that flatters a provider without helping a visitor, and the tier we would put a demanding site on.
By the HostingFast team · Reviewed 15 August 2026
24/7
Engineers on duty
1-click
Application installs
Free
SSL certificates
Daily
Backups, all plans
For buyers who want shared-tier pricing to behave like a performance tier, and who would rather run a test than read a superlative.
Two things determine shared hosting speed: the software stack the provider chose and how many accounts are packed onto each machine. Identical hardware loaded to different densities produces wildly different results, which is why tier names tell you almost nothing.
Four readings that settle it
First, uncached time to first byte on a real page — not the homepage of a demo site, and not a static file. Append a cache-busting parameter so PHP and the database are genuinely doing work. Under 600 milliseconds is the working target.
Second, the cache headers. On a LiteSpeed-backed account the response reports whether it was a hit or a miss, so you can tell the difference between a fast server and a warm cache. A provider that cannot show you this is asking you to take the fast part on faith.
Third, variance. Take the first reading every hour across a working day. A well-packed machine produces a tight cluster; an oversold one produces a spread, and the spread is the number that predicts how the site feels at your busiest hour.
Fourth, behaviour under concurrency. Twenty simultaneous uncached requests for thirty seconds is enough to see whether response times degrade gracefully or fall off a cliff. Keep it modest and within the terms — you are sampling, not attacking.
What each reading actually exposes
Uncached TTFB exposes the PHP and database path: interpreter version, opcode cache, disk latency. It is the closest thing to a specification you can measure from outside.
Cache headers expose the architecture. Server-level caching skips the application entirely, and no amount of hardware beats not doing the work — which is why a LiteSpeed hit and a PHP miss are two different products wearing one price.
Variance exposes density: the accounts-per-server figure nobody publishes. You cannot ask for it and get an honest answer, but you can measure its consequence in an afternoon.
Concurrency exposes the resource limits on your account, which is where a plan's real ceiling lives — process counts and entry processes rather than the advertised gigabytes.
The benchmark that flatters a host
A cached homepage timed once from a nearby test node is the number every provider would like you to use. It measures the cache and the network and nothing else, and it will look excellent on hardware that cannot carry your actual site.
The related trap is accepting adjectives in place of an architecture list. Every sluggish provider in this market has the word fast on its homepage, and almost none will name their density figures.
So ask for the architecture — storage type, cache layer, PHP release, whether resources are isolated per account — and then measure the consequences yourself rather than debating the answers.
Our pick, and the invitation to test it
Nitro is where this page lands: 200 GB of NVMe, 2 cores and 4 GB of RAM, backups every 6 hours with Imunify360 in front, and PHP X-Ray on CloudLinux Pro so a slow request can be traced to the line of code responsible rather than guessed at.
That last item is the reason we would choose it for a demanding site. Most shared hosting leaves you inferring where time went; X-Ray tells you, which turns a hosting complaint into a code fix.
Run all four readings during the 30-day window on hosting plans. Bring a real site across — migration is free — and if the numbers do not justify the move, leaving takes one request rather than a negotiation.
Why this page has no affiliate table
Most 'best hosting' content is ranking-for-hire. We sell the thing being measured and say so, which is why the page hands you the method instead of a verdict you would have to trust.
The rate you sign up at is the rate you renew at, so year two costs exactly what year one did — no ambush waiting in the invoice.
- Four readings you take on your own account
- Cache hit or miss, readable in the headers
- Free migration, so the test costs nothing
- No retention call to escape
Why HostingFast
Standard on every plan
Tracing, not guessing
PHP X-Ray on CloudLinux Pro attributes a slow request to the code that caused it, on the Nitro tier.
NVMe on every tier
The storage layer is the baseline build here rather than a feature gated behind the most expensive plan.
Cache status you can read
A hit and a miss look different in the response headers, so the fast part is verifiable rather than asserted.
Nitro, in one line
200 GB of NVMe, 2 cores, 4 GB of RAM and backups every 6 hours at one published monthly rate.
A 30-day window to test in
Hosting plans carry 30 days, which is more than enough to sample variance across several working days.
A refund without a fight
If the readings do not justify the move, leaving is one request rather than a retention conversation.
Prices Side by Side
Our numbers next to the big names
Typical intro and renewal pricing across the market, lined up against ours — including the renewal figure most comparison charts skip.
| Feature | HostingFastMost popular | Typical household-name host | Typical bargain host | Typical loss-leader deal |
|---|---|---|---|---|
| Starting price / mo* | $2.42/mo | $4–$6 | $2–$4 | $1–$3 |
| Renewal price / mo | $2.42/mo | $10–$15 | $8–$12 | $4–$6 |
| Cheapest plan renews at its sign-up price | ||||
| SSL as standard | ||||
| Migration done for you | ||||
| NVMe drives on the cheapest tier | ||||
| Daily backups from the first tier | ||||
| Real humans on support, 24/7 |
*Our number is the lowest-priced plan we sell on an annual term, pulled live from the same catalogue that powers our pricing page — it physically cannot fall out of date. The other three columns show the ranges shared hosts in each bracket typically advertise: introductory offers that mostly demand a one-to-four-year commitment, then jump once it expires. We no longer name rivals or print their prices, because a figure we cannot verify on the day you read it has no business on this page. Line us up against whichever host you are really weighing — and give the renewal row the hardest look.
Quick Start
From order to online
- 1
Bust the cache before you time anything
A cached response measures the cache. Add a unique query parameter so PHP and the database are genuinely doing the work you are trying to price.
- 2
Sample hourly, not once
One reading is an anecdote. A day of hourly readings shows you the variance, and variance is what an oversold machine cannot hide.
- 3
Work out the two-year total
Renewal rate times the term, plus every extra you would genuinely enable. Introductory pricing never survives that arithmetic.
- 4
Test concurrency politely
Twenty simultaneous requests for thirty seconds tells you how the plan degrades. Stay modest and inside the terms — you are sampling, not stress-testing.
Built In
Loaded onto every plan
- NVMe SSD on every tier, not only the expensive ones
- LiteSpeed caching built into the server rather than patched in by plugin
- Cache hit or miss reported back in the response headers
- A current PHP release, selectable per site
- Resource isolation between accounts on the same machine
- PHP X-Ray available for tracing a slow request to its cause
- Free migration, so the before-and-after test costs nothing
- Money-back cover: 30 days on hosting plans, 7 on reseller
- Renewal charged at the rate you ordered at
- In-place upgrades, so growth is not a migration
Frequently Asked
What people ask us most often
How do I stop a speed test measuring the cache instead of the server?
Add a unique query string to the URL on every request, so each one is a cache miss and PHP has to run. Compare that figure with the cached one: the gap is what the cache is saving you, and the uncached figure is what the machine can actually do when it has to do something.
What does variance across a day tell me that an average does not?
It tells you about density. A machine with room produces a tight cluster of response times; an oversold one produces outliers that cluster around the working day. Averages hide exactly that, which is why one flattering reading at midnight is the most common way buyers talk themselves into a slow account.
Can I load-test a shared account, and how hard?
Modestly, and within the acceptable-use terms. Twenty concurrent requests for half a minute is a sample, not an attack, and it tells you whether your plan's process limits degrade gracefully. Anything heavier is both unrepresentative and a good way to have your account suspended by an automated rule.
What does PHP X-Ray give me that a plugin profiler does not?
It runs below your application, so it captures requests you were not watching — including the slow ones that happen at three in the morning to a crawler. A plugin profiler tells you about the page you loaded on purpose; X-Ray tells you where the account's time actually went, which is a different and more useful question.
Keep reading
Squarespace vs WordPress
A closed builder against a stack you can instrument, compared on what you can change.
Cheap Web Hosting That Actually Performs
Cheap hosting that still holds a response time, and the readings that prove it.
Shared Hosting (Glossary)
What the shared model actually shares, and which parts of it are isolated.
Web Hosting
cPanel hosting on NVMe with SSL, migration and a year-one domain in the price.
Domain Names
Register or transfer a name, with the first year free alongside annual hosting.
Changing hosts? Run through our checklist first.
A straightforward sequence for a switch your visitors never feel: which files move first, how to shift email across without losing a single message, the right moment to repoint DNS, and the two mistakes behind almost all the downtime we get asked to rescue.
Run the four readings on us.
NVMe, LiteSpeed and cache headers you can read, inside a 30-day window with free migration in front of it.
View Web Hosting plans