In-House IT Team or Managed Hosting? Calculating the Real Cost

The short answer: for a mid-size company running one or two custom systems, managed hosting paired with a single in-house developer or engineer costs less and carries less risk than building a full IT team that runs its own servers. The rough numbers make this clear. A mid-level Saudi systems engineer costs the business between 240,000 and 360,000 SAR a year once you add allowances, social insurance (about 12% employer share for Saudi nationals), medical insurance and training. That engineer works around 1,900 hours a year, while a year has 8,760 hours, so true round-the-clock coverage needs at least four people on rotation, which means more than one million riyals a year. By comparison, a cloud server on the Enterprise plan (8 vCPU and 16GB RAM) costs 18,000 SAR a year, reaching 21,200 SAR with 30-day daily backups and DDoS protection. An in-house team stays essential for knowing the system and its business logic, but infrastructure such as the data center, power, network, hardware and backups is not your competitive edge, and handing it to a specialist frees your team to work on what actually sets your company apart.

The Real Cost of an In-House Team

Salary is not the whole cost. When you price a systems engineer, add every annual line item and then what they need to do the job:

Item	Annual estimate (SAR)
Base salary and allowances, mid-level engineer	200,000 to 300,000
Employer social insurance (about 12%)	20,000 to 30,000
Medical insurance	8,000 to 15,000
Training and certifications	10,000 to 20,000
Monitoring tools and licenses	5,000 to 15,000
Total for one person	243,000 to 380,000

These are approximations that vary by city and experience, but the point holds: one person costs many times the infrastructure they manage. If you also keep owned servers in an office room, add cooling, backup power and replacing disks and hardware every 3 to 5 years.

What Stays with Your Team and What Moves to the Provider

Managed hosting does not make your team disappear. It makes the boundaries clear. This is the usual split on a cloud server:

Task	Provider	Your team or developer
Data center, power and cooling	Yes	No
Hardware and failed disk replacement	Yes	No
Network and DDoS protection	Yes	No
Virtualization layer and server activation	Yes	No
Server-level backups	Yes, when enabled	Periodic restore tests
OS and package updates	Depends on service	Usually
Application code and database	No	Yes
Business logic and government integrations	No	Yes

With managed cloud hosting the provider carries a larger share, since the server, its updates and daily and weekly backups are its responsibility, and you get a ready cPanel account priced from 2,900 to 5,750 SAR a year. A cloud server gives you full root access instead, which suits systems that need custom settings or Docker containers.

The Risk of the One Person Who Knows Everything

The most dangerous scenario in mid-size companies is not a breach. It is the resignation of the only person who knows the passwords and how the server runs. We see it often: an ERP running for 6 years, the engineer who built it has left, and nobody knows where the backup lives or why the server reboots every night. Managed hosting reduces this risk in three ways: the infrastructure is documented by the provider instead of living in one employee's head, access goes through a control panel with transferable accounts, and support is available regardless of vacations and resignations.

Even with managed hosting, document what is yours. A simple file in the code repository is enough:

# ops/RUNBOOK.md
- Server: Riyadh, Business plan, IP: x.x.x.x
- Access: provider panel + team SSH key (no passwords)
- Backups: daily, 30-day retention, last restore test: 2026-09-14
- Restart: docker compose up -d inside /srv/erp
- Contacts: in-house developer, provider support

Office Hours vs Round-the-Clock Monitoring

Your in-house team usually works from 8 a.m. to 4 p.m., which is 40 of the week's 168 hours, covering only about 24% of the time. Failures ignore office hours, whether it is a full disk at 3 a.m. or a network attack on a Friday. A provider managing the infrastructure watches hardware and network across all 168 hours and swaps a failing disk before you notice. You still need to watch the application itself, and a free tool like Uptime Kuma on a small server alerts you within a minute of a page going down:

docker run -d --restart=always -p 3001:3001 \
  -v uptime-kuma:/app/data --name uptime-kuma louislam/uptime-kuma:1

With this split you get full coverage without hiring four people for night rotations.

When the Hybrid Model Fits Best

For most of our customers with custom systems, the hybrid model wins: one in-house developer or engineer who understands the system and its business logic, plus a provider running the infrastructure. The rough math:

Model	Approximate annual cost	Coverage
Full four-person in-house team with owned servers	Over 1,000,000 SAR	Round the clock
One in-house engineer with owned servers	270,000 to 400,000 SAR	Office hours only
One engineer with a provider-managed cloud server	265,000 to 400,000 SAR	Infrastructure round the clock, application in office hours with alerts

The cost gap between the second and third rows is small, yet the risk gap is very large, because you have removed hardware, power and the server room, and you now have a provider accountable for infrastructure under a contract with a 99.99% SLA.

When you really need a large in-house team

• Technology is your core product, such as a software company serving thousands of users.
• You run more than 20 interconnected systems that need daily coordination.
• A regulator requires you to operate every layer internally.

Outside these cases, a large team running hardware usually consumes budget that would be better spent developing the system itself.

The Hidden Cost: Downtime and Firefighting Hours

The payroll sheet hides an important line, which is the time lost when an engineer switches from development to firefighting. In a small team running its own servers, often 30% to 50% of an engineer's time goes to infrastructure work: security patches, a full disk, an expired SSL certificate, a backup that failed silently. If the engineer costs 300,000 SAR a year, somewhere between 90,000 and 150,000 SAR is spent on work that adds not a single feature to your system.

Add the cost of downtime itself. A system used by 50 employees at an average hourly cost of 80 SAR loses 4,000 SAR for every hour it is down. Three incidents a year lasting 4 hours each cost 48,000 SAR, more than double the price of an Enterprise cloud server with backups and protection. Once you count this line, the question shifts from "can we afford a provider" to "can we afford running the infrastructure ourselves."

Questions to ask the provider before handing over

• Who replaces a failed disk, and within how many hours?
• Are backups stored separately from the server, and for how long are they kept?
• Does the provider offer guidance sessions for your developer during migration and incidents?
• Where exactly is the data located, and does it stay inside the Kingdom in every case?

At CLOUD, for example, servers sit in Riyadh at around 8ms latency and we run free guidance sessions for customer developers, while migration and application code remain your team's responsibility, a clear split where each side knows its limits from day one.

How to Decide Within a Week

1. Calculate your current team's full cost using the table above, not salary alone.
2. List every technical task from the last 3 months and tag each one as infrastructure or application.
3. If infrastructure tasks exceed 40% of the team's time, that is time you can free up.
4. Trial a right-sized cloud server, such as the Business plan at 11,500 SAR a year, and move a test copy of the system.
5. Write the RUNBOOK before any migration so nothing stays in one person's head.

The decision is not really in-house team versus provider. It is where your team spends its hours, on hardware or on the system that runs your business.