The distinction between managed and unmanaged hosting comes down to one question: who is responsible for keeping the server environment running, secure and up to date? With unmanaged hosting, the provider supplies the raw server — virtual or physical — and your team handles operating-system patches, web-server configuration, database tuning, backups and security hardening. With managed hosting, the provider takes on some or all of those operational tasks.

The complication is that "managed" is not a fixed standard. One provider might patch the operating system and restart services, whilst another handles application-level monitoring, automated backups, SSL certificate renewal and scaling decisions. The only reliable way to understand what you are buying is to read the service description line by line and confirm in writing which responsibilities remain with you.

For a business web application — a CRM, a customer portal, an internal admin system — the hosting decision is not purely technical. It affects how quickly your team can respond to incidents, what skills you need to recruit or retain, and how much control you have over the environment your software runs on. It also shapes your exit options if you later need to move to a different provider or bring the infrastructure in-house.

Hosting decisionBusiness consequenceEvidence to request
Where the responsibility boundary typically fallsWith unmanaged hosting, the provider is responsible for the physical hardware, the network connection and the hypervisor if the server is virtual.Everything from the operating system upwards is your concern.
Ownership and controlHosting type does not change who owns your data or your application code.But it does affect how easily you can access both.
When managed hosting tends to be the stronger optionManaged hosting suits organisations that do not have, and do not want to recruit, dedicated infrastructure or DevOps staff.If your team is primarily focused on building and improving the application — adding features to a SaaS product, refining a CRM workflow, or extending a customer portal — then spending time on server patching and certificate renewal is a distraction.
When unmanaged hosting tends to be the stronger optionUnmanaged hosting can suit a business with capable in-house engineers and requirements that need unusual runtime, caching, database or network configuration. The business must still assign patching, monitoring, backup and incident ownership.Name the people responsible for hardening, patching, monitoring, backups, recovery and incidents, then test that they have the access and time to perform those duties.

Where the responsibility boundary typically falls

With unmanaged hosting, the provider is responsible for the physical hardware, the network connection and the hypervisor if the server is virtual. Everything from the operating system upwards is your concern. That includes user accounts, firewall rules, log rotation, disk monitoring and failover procedures.

With managed hosting, the provider usually covers operating-system updates, web-server and database-server maintenance, basic security configuration and often backup routines. However, application-level concerns — deploying new code releases, managing application secrets, handling application-specific scaling — almost always remain with your development team, regardless of the hosting label.

Ownership and control

Hosting type does not change who owns your data or your application code. But it does affect how easily you can access both. An unmanaged server gives your team direct shell access and full control over the environment, which simplifies migration if you need to move. A managed service may restrict access to certain configuration layers, which can make extraction more involved. Before committing, check what access your team will have, what formats backups are supplied in, and whether there are any contractual barriers to retrieving your data.

When managed hosting tends to be the stronger option

Managed hosting suits organisations that do not have, and do not want to recruit, dedicated infrastructure or DevOps staff. If your team is primarily focused on building and improving the application — adding features to a SaaS product, refining a CRM workflow, or extending a customer portal — then spending time on server patching and certificate renewal is a distraction.

It also fits well in regulated environments where you need to demonstrate that infrastructure is being maintained by qualified personnel. A managed provider can supply evidence of patching schedules, backup test results and incident response procedures that would otherwise take your team considerable effort to document.

For early-stage SaaS products where the founding team is small and speed to market matters more than infrastructure optimisation, managed hosting reduces the operational surface area so the team can concentrate on product-market fit.

When unmanaged hosting tends to be the stronger option

If you already have in-house engineers who understand server administration, unmanaged hosting avoids paying a premium for tasks your team can perform themselves. It also gives you complete freedom to configure the environment exactly as your application requires — specific versions of language runtimes, custom caching layers, unusual database configurations or particular networking setups that managed providers may not support.

Unmanaged hosting becomes more attractive at scale. As your application grows, the markup that managed providers apply to compute, storage and data transfer can become a significant line item. A team that can automate its own provisioning, monitoring and scaling through infrastructure-as-code practices may achieve lower variable costs with unmanaged infrastructure.

Organisations with strict compliance requirements sometimes prefer unmanaged hosting because it removes any ambiguity about who configured a particular security setting. When your own engineer applies every firewall rule and every hardening step, there is no gap between what you assume has been done and what has actually been done.

Questions to put to a provider

  • Exactly which tasks are included in the managed service, and which are excluded?
  • What is the response time for infrastructure issues, and how is that measured?
  • Are backups stored in a separate region or account, and can you restore them independently?
  • What access will your team have — full root, restricted admin panel, or something in between?
  • What happens if you need to migrate away — what formats and access are provided?
  • How are operating-system and middleware patches applied, and what notice is given?
  • Is scaling manual or automatic, and what are the cost implications at each tier?

Assuming "managed" means fully hands-off

The most frequent error is treating a managed hosting contract as a complete outsourcing of operational risk. In most cases, the provider manages the infrastructure layer, not the application. If your application has a memory leak, serves incorrect data, or fails to handle a surge in concurrent users, that is your team's problem regardless of the hosting label. Managed hosting reduces operational burden; it does not eliminate it.

Underestimating the true cost of unmanaged hosting

An unmanaged server may look considerably cheaper on a per-month basis, but that figure excludes the cost of the people who keep it running. If you need to hire or contract a DevOps engineer, pay for monitoring tooling, and allocate on-call time for incident response, the total cost of unmanaged hosting can meet or exceed a managed alternative. When comparing options, include staff time, tooling subscriptions and the cost of covering absence or turnover in your calculations.

Not verifying backup independence

Some managed hosting services include backups that are stored on the same physical infrastructure as the live environment. If that infrastructure fails, the backups fail with it. Check whether backups are geographically separated, whether they are stored in a different provider account, and whether you can initiate a restore without contacting the provider's support team. The ability to restore independently is a basic safeguard that is worth confirming before you commit.

Ignoring the exit path

Hosting decisions are easy to make but can be expensive to reverse. If a managed provider uses proprietary tooling or non-standard configurations, extracting your application and data later may require significant engineering work. Before signing, ask what a migration out would involve, what documentation is available, and whether there are any contractual restrictions on moving. This is not about expecting the relationship to fail; it is about retaining control over your own operational choices.

Overlooking security responsibility gaps

With managed hosting, there is a risk that both sides assume the other is handling a particular security task. The provider may assume your team is managing application-level access controls, whilst your team assumes the provider is handling web-application firewall rules. The only way to close these gaps is to map every security responsibility explicitly — ideally in a shared document that both parties review and agree on. Security and privacy arrangements should be reviewed by a current specialist, as requirements change and provider capabilities vary.

Key checks before deciding

  • Map your team's actual infrastructure skills honestly, not aspirationally.
  • List every operational task your application requires and check which ones the managed service covers.
  • Calculate the full cost of unmanaged hosting including staff, tools and on-call cover.
  • Confirm backup independence and restore procedures in writing.
  • Verify the exit path — what you would need to do to move, and how long it would take.
  • Ensure security responsibilities are explicitly assigned with no assumed handoffs.
  • Check whether the hosting model affects your ability to meet specific compliance or audit requirements.

The right choice depends on your team's current capabilities, your application's specific operational demands, and how much control you are willing to trade for convenience. Neither option is universally superior — but choosing without understanding where the responsibility boundary actually sits is a reliable way to create problems later.