We moved a 1,600-post WordPress site off Cloudways onto a self-hosted box last year. People ask whether it was worth it, so here are the actual numbers and the honest trade-offs. Your answer might differ from ours, and that’s the point.
The site ran on managed WordPress on Cloudways, sitting on a small VPS at around $40 a month. That’s fine until traffic moves. A busy month, a post that did well, a crawl from a few aggressive bots, and the site needed a bigger plan. Under load it scaled to roughly $100 a month for the headroom to stay responsive.
Cloudways earned that money. Automatic backups, SSL handled for us, server-level updates applied, and support available around the clock. For a team without infrastructure capability, that’s a reasonable deal. We just happened to have the capability.
A Hetzner CCX13 dedicated-vCPU cloud server at around $15 a month. For that price the box has dedicated CPU rather than shared, more RAM than the comparable Cloudways tier, more disk, and we control all of it.
The headline comparison is stark. Roughly $15 against $40 to $100. But the headline isn’t the whole story, because the price difference is partly us taking on work Cloudways used to do.
More machine for less money is the obvious one. Dedicated CPU matters for WordPress under load, where a noisy neighbour on a shared plan can drag your response times even when your own traffic is flat. We got predictable performance and room to grow without the next pricing tier looming.
Full control is the other gain. We tune the stack, choose the PHP version, configure caching the way we want, and run whatever else we like on the same box. On managed hosting you live inside the provider’s choices.
Honesty matters here, because the savings come with responsibility transferred to us.
We lost automatic backups, so we had to build our own. We lost managed SSL, so we had to provision certificates ourselves. We lost server-level updates applied for us, so patching is now our job. And we lost 24/7 support, so when something breaks at an awkward hour, the person fixing it is us.
That’s a real loss. For a team that can’t carry it, it’s the whole reason managed hosting exists.
Each thing Cloudways did, we rebuilt with tooling we already trust.
For SSL and edge, a cloudflared tunnel. Cloudflare terminates TLS, the origin isn’t directly exposed, and certificate management stops being something we think about.
For backups, restic to object storage, scheduled and encrypted, with a restore we actually tested rather than assumed.
For configuration, Ansible, so the whole box is reproducible from a playbook instead of remembered steps. If the server dies, we rebuild it from code.
For monitoring, our own stack, so we see resource use, uptime, and the early signs of trouble before they become an outage.
Standing that up took time. Once it existed, it cost almost nothing to run and it’s reusable across every other box we operate.
We took on more responsibility and gave up the comfort of someone else owning the platform. We pay for that in attention, in the playbooks we maintain, and in being the support line. In exchange we pay a fraction of the hosting cost and we control the environment completely.
That trade is good for us because we have the engineering capability and the tooling already. It would be a bad trade for a team that doesn’t, where the time spent operating the box outweighs the money saved and the first failed restore is a disaster.
At certain scales and team capabilities, self-hosted is clearly the right answer. You have people who can run a server, you have or can build the tooling, and the savings compound across your fleet.
At others, managed PaaS still wins. A small team with no infrastructure capability, or a business where engineering attention is better spent on the product than on patching servers, should pay Cloudways or its equivalents and get on with building.
For most small-to-mid SaaS products with real engineering capability, the maths favours self-hosting. Hetzner or DigitalOcean delivers equivalent capacity at roughly three to five times less than managed hosting, once you account for the work you take on. The break-even is your team’s ability to own the platform, not the price tag.
If you’re weighing a move like this, our cloud team has run both sides and can tell you which way the numbers fall for your specific setup. Book a consultation and we’ll work through your hosting costs and your team’s capacity honestly, including the case for staying managed if that’s the right call.