Performance Brief
Hosting with redis cache — The object cache only helps where the page cache cannot
Your homepage is already instant and your checkout still takes two seconds, because the pages that matter to revenue are exactly the ones a page cache is required to skip.
The short answer
Run Redis as an object cache on a plan that provisions it properly: our WordPress tiers run Redis via AccelerateWP, and Turbo WordPress at $5.59 a month is where most stores and membership sites land.
Understand what it does before you judge it. A page cache stores finished HTML for logged-out visitors, so those requests never reach PHP. An object cache stores the results of database queries, so the requests that must be built fresh — basket, account, admin, anything logged-in — build faster. They solve different halves of the problem and you want both.
The test is specific: time an uncacheable page before and after. Five runs of `curl -s -o /dev/null -w '%{time_starttransfer}'` against your cart or your dashboard, median taken, before enabling and after. If that number does not move, the object cache was not your bottleneck and this page will tell you what probably was.
By the HostingFast team · Reviewed 24 August 2026
Free
Migration, run by us
24/7
Humans on the desk
Daily
Backup cadence
$0
Joining fee
Object caching is the most commonly misapplied optimisation in WordPress. It gets enabled on brochure sites where every page is already served from cache, and it gets skipped on stores where every important page bypasses cache by design.
The way to get this right is to know which of your requests actually reach PHP, and to measure those specifically. Everything below is aimed at that question.
Which of your requests actually reach PHP
Start there, because it decides everything else. `curl -sI https://yourdomain.example` and look at the LiteSpeed cache header: a hit means the page never woke PHP and no object cache can improve it. Repeat with a session cookie for a logged-in view and you will usually see the opposite.
On a brochure site, nearly every request is a hit and the object cache is close to irrelevant. On a store, the basket, the checkout, My Account and every admin screen are misses by design — you cannot serve one customer's basket to another.
Count the queries on one of those misses. Query Monitor will report the number of database queries and the time they consumed for a single page load. Two hundred queries taking 400 ms is an object-cache problem. Twelve queries taking 900 ms is not — that is one slow query, and caching it will hide it rather than fix it.
What Redis is actually storing
WordPress asks the database the same things constantly: options, term relationships, post metadata, transients. Without a persistent object cache, most of that is re-fetched on every request. With one, the results live in memory and the query never runs.
The gain therefore scales with how repetitive your query load is, which is why WooCommerce and membership plugins benefit most — they are query-heavy and their pages cannot be page-cached. It is also why wp-admin so often feels transformed: the dashboard is one of the most query-dense pages WordPress serves.
The connection matters too. Redis over a local socket rather than a network hop keeps the lookup cost negligible, and support that will actually engage when a connection drops is worth more than a tick on a feature grid.
Reading the hit ratio without fooling yourself
The Redis plugin's status panel reports connection state and hit rate. A hit ratio above roughly 80% means the cache is doing genuine work; below that, either the memory allocation is too small and keys are being evicted, or your traffic is too varied for the working set to stay resident.
Watch it over a normal day rather than for a minute after enabling it. A cache read immediately after a flush will show whatever you want it to show.
The allocation is the number to ask about before you buy. An object cache lives or dies on its eviction rate, and a memory limit nobody will state is a limit you will discover during your busiest hour.
What it will not fix, and what to do instead
A slow external API call. If checkout waits on a payment gateway or a shipping quote, the object cache is not in that path at all — the fix is an asynchronous call or a shorter timeout.
An N+1 query pattern in a plugin. Caching the results makes the symptom smaller and leaves the cause; profiling and replacing the plugin is the real repair. On Nitro at $7.08 and on the Nitro WordPress tier, PHP X-Ray on CloudLinux Pro will point at the offending function directly.
An unindexed table. `EXPLAIN` on the slow query will show a full scan; adding the index is a one-line fix that beats any amount of caching.
And images. If LCP is four seconds because of a hero photograph, no cache layer of any kind will help. WebP conversion is built into the platform and is the cheaper win. Turbo WordPress at $5.59 gives you 25 WordPress sites, 20 GB of NVMe, 2 cores and 4 GB of RAM, 400k monthly visits, AccelerateWP with Redis, Toolkit Deluxe and up to ten staging copies, renewing at $5.59.
Why we tell you when not to bother
Selling an object cache to someone whose whole site is page-cached would be easy and would generate a refund request three weeks later. Describing when it does nothing is cheaper for both of us.
This is our own WordPress plan being recommended, stated plainly. The measurement method works on any host, and you can run all of it before you spend anything.
- Measure the uncacheable pages, not the homepage
- Hit ratio above roughly 80% means real work
- Redis over a local socket, provisioned on the plan
- The causes it hides, named rather than glossed
Why HostingFast
Standard on every plan
Redis provisioned, not improvised
Object caching via AccelerateWP on the WordPress plans, connected over a socket and supported by the desk rather than left to a plugin.
Both cache layers together
LiteSpeed in the web server for logged-out HTML, Redis behind it for the queries that logged-in pages have to run anyway.
A measurable cache header
Every response says whether it was a cache hit, so you can tell which of your pages the object cache can even affect.
X-Ray on the top tiers
PHP X-Ray on CloudLinux Pro identifies the slow function on Nitro, which is how you find the cause a cache would have hidden.
Ten staging copies to test on
Enable it on a clone, measure, then decide. Changing cache configuration on production without a baseline is guessing.
WebP conversion built in
Because on plenty of sites the real bottleneck is a photograph, and no cache layer will ever fix that one.
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
Find out which pages miss the cache
`curl -sI` for the LiteSpeed cache header, logged out and logged in. Only the misses are candidates for an object cache.
- 2
Count queries on a miss
Query Monitor on a cart or dashboard page. Many cheap queries means Redis will help; few expensive ones means it will not.
- 3
Baseline, enable, re-measure
Five curl timing runs before and after on the same uncacheable page, median taken, same time of day.
- 4
Check the hit ratio over a full day
Above roughly 80% is genuine work. Lower usually means eviction, which means the memory allocation, not the idea, is wrong.
Built In
Loaded onto every plan
- Redis object caching via AccelerateWP on the WordPress plans
- LiteSpeed caching in the web server, with a verifiable response header
- NVMe on every tier, Sprint upwards, not just the top of the range
- Up to ten staging copies for testing cache changes safely
- Support staffed through the night rather than queued until Monday
- Year one of the domain free when you order annually
- WebP image conversion built in at no extra cost
- Daily backups with self-service restores from the panel
- In-place upgrades between tiers, with no migration to arrange
- Your existing site moved across by our engineers, free of charge
- Webmail plus IMAP, POP and SMTP for any mail client
Frequently Asked
What people ask us most often
I already have a page cache. What does Redis add?
Speed on the requests the page cache is obliged to skip. Basket, checkout, My Account and wp-admin cannot be served from a shared HTML cache, so they run PHP and the database every time. The object cache keeps repeated query results in memory so those pages build faster. Brochure sites see almost nothing; stores and membership sites see the most.
What hit ratio should I expect, and how do I read it?
The Redis plugin's status panel reports it. Above roughly 80% means the cache is doing real work. Below that, keys are usually being evicted because the memory allocation is too small for your working set. Read it over a normal trading day rather than in the minute after you enabled it.
Redis is connected but nothing got faster. Why?
Almost certainly you measured a page that was already served from the page cache, in which case PHP never ran. Or the load is a few expensive queries rather than many cheap ones, which caching hides rather than fixes. Count queries with Query Monitor on an uncacheable page — that number tells you which case you are in.
What happens when the object cache runs out of memory?
It evicts, and your hit ratio falls. That is why the memory allocation is the specification worth asking about before you buy: an object cache is only as good as its ability to keep the working set resident, and a limit nobody will state is one you find during your busiest hour.
Keep reading
Hosting for Small Business Websites
Where the page cache does almost all the work, and an object cache would be effort spent on nothing.
Hosting for Students
The same measurement discipline on a smaller budget, with the commands to produce a defensible number.
How to Set Up WordPress Caching
Both layers configured step by step, with how to confirm each one is genuinely serving requests.
WordPress Hosting
Sprint, Turbo and Nitro WordPress with AccelerateWP, Redis, Toolkit Deluxe and ten staging copies.
Mail Filtering
Transactional mail from a busy store is a separate path from the page load, with its own failure modes.
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.
Cache the pages the cache cannot.
Turbo WordPress at $5.59 a month with Redis via AccelerateWP, LiteSpeed in front and NVMe underneath all of it.
View WordPress Hosting plans