Shared vs VPS hosting: finding the fit
Roughly once a week someone opens a ticket that says the site is slow and asks whether they should move to a VPS. Sometimes the answer is yes. More often the slowness has nothing to do with the plan, and upgrading buys an expensive version of the same problem.
What shared hosting actually is
On a shared server your site is one account among many on a single machine. The CPU, memory and I/O are pooled, and the control panel enforces limits so that no single account can starve the rest. That pooling is precisely why shared hosting is cheap, and precisely where its ceiling comes from.
For a brochure site, a small WordPress blog, a portfolio or a business site doing a few thousand visits a month, that ceiling is nowhere near. Those sites are mostly cached HTML and images. They do not need a dedicated slice of anything.
The three signals that you have outgrown it
- You are hitting resource limits, not bandwidth. Check the CPU and entry-process graphs in cPanel. If you are being throttled during ordinary traffic — not during a backup or a crawl — the workload genuinely needs more headroom.
- You need software the platform does not offer. A specific PHP extension, a long-running Node process, a queue worker, a custom cron cadence, a particular Redis configuration. Shared hosting is a curated environment by design; if you need to install things, you need root.
- Your traffic is spiky and revenue-bearing. A store during a sale, a news site during an event. Shared plans are provisioned for averages. If an hour of downtime costs real money, buy the isolation.
The three signals that a VPS will not help
These are the cases we see most often, and they follow you onto any server you move to:
- An unoptimised database. Missing indexes and a bloated
wp_optionstable with hundreds of autoloaded rows will be slow on any hardware. More CPU just means the same bad query runs hot on a bigger machine. - Thirty active plugins. Each one adds queries and HTTP calls to every page load. Audit before you upgrade — a query profiler will usually name the culprit in ten minutes.
- Unoptimised images and no caching. A 4 MB hero image is 4 MB on a dedicated server too. Page caching, an object cache and properly sized images are still the highest-leverage fixes available, and they cost nothing.
The question is not "am I big enough for a VPS?" It is "is the thing that is slow something a bigger machine can fix?"
What changes when you move
A VPS gives you guaranteed CPU and RAM, root access and full control of the stack. It also gives you responsibility for the stack. Someone now has to apply kernel updates, configure the firewall, renew certificates, watch the logs and keep backups running. That is a real job.
Managed VPS plans exist for exactly this reason: you get the isolation and the root access, and the provider keeps the operating system patched and monitored. If nobody on your team wants to be on call for a server, that is the version to buy.
A sensible upgrade path
Fix caching, images and the plugin list first. Measure again. If you are still throttled at the platform limits, move to a VPS sized about 30% above your current peak, not five times it. Watch it for a month, then adjust. Servers are easy to grow and awkward to shrink, so start conservative.
And migrate with the old plan still running. Point a staging domain at the new server, verify the site properly, then move DNS. A migration you can walk back is a migration that cannot ruin your week.
Keep reading
Removing Chrome’s “Dangerous site” warning
What triggers Google Safe Browsing’s red interstitial, how to find what got injected, and how to get the flag lifted properly.
How to create your first n8n workflow
Build a real n8n automation from an empty canvas: trigger, nodes, data shaping, error handling, and the case for self-hosting it.