What an in-house IT helpdesk actually costs a 40-person company

Most businesses that hire their first IT person budget for one number: the salary. It’s the number that goes to the board, the number in the job posting, and the number people compare against a managed services quote when they’re deciding which way to go.

It’s also the smallest line in the calculation.

This is a full teardown of what an internal helpdesk costs a roughly 40-person Canadian business. Every figure below is an assumption you should replace with your own — payroll costs vary enormously by province, sector, and seniority. What matters isn’t our numbers. It’s the line items most people leave out entirely.

Start with the obvious line

One IT generalist, mid-level, supporting 40 users. Call the base salary $70,000.

That’s the number that gets budgeted. Now here’s everything that sits behind it.

Add employer costs

In Canada you’re paying more than salary before anyone touches a keyboard:

• CPP and EI employer contributions
• Provincial payroll tax, where applicable — Manitoba’s Health and Post-Secondary Education Tax Levy applies above a payroll threshold, and Ontario, Quebec, BC, and Newfoundland all have their own versions
• Workers’ compensation premiums
• Vacation accrual — two to three weeks that you’re paying for and not receiving coverage during
• Benefits — health, dental, sometimes an RRSP match

A common planning rule is to load salary by 15–25% for employer costs before benefits, and benefits can add another 8–15% on top. On a $70,000 salary that’s realistically $85,000–$95,000 in true annual cost.

Already the budgeted number is off by a fifth.

Add the tools they need to do the job

An IT person without a toolchain is an IT person with a notepad. At minimum:

• RMM platform — remote monitoring and management, priced per endpoint
• Ticketing system
• Endpoint protection, per seat
• Backup, per seat plus storage
• Password and secrets management
• Remote access tooling
• Documentation platform — because everything they know needs to live somewhere other than their head

Per-endpoint and per-seat pricing means a 40-person company pays 40 times over on most of these. Budget $8,000–$18,000 annually for a working stack, before any specialist licences.

Note that this cost doesn’t disappear when you outsource — but it usually gets absorbed into the provider’s per-seat rate, because they’re buying at volume across their whole client base. That’s one of the few genuine structural advantages of outsourcing, as opposed to the ones vendors invent.

Now the part nobody budgets: coverage

Here’s the real problem with one person.

One person covers roughly 40 hours of a 168-hour week. That’s 24% coverage. The other 76% of the time — evenings, weekends, statutory holidays, and every single night — nobody is watching anything.

And that’s the good case, because one person also:

• Takes two to three weeks of vacation
• Gets sick
• Attends training
• Eventually resigns

During any of those, coverage drops to zero. Not reduced. Zero.

For most businesses this is survivable right up until the moment it isn’t. The ransomware encryption that starts at 11pm Friday, the server that fails on the Saturday of a long weekend, the email outage on the first day of your one IT person’s vacation — these are the events that turn a modest IT budget into a catastrophic one, and they are precisely the events a single-person team is structurally unable to catch.

If you want genuine 24/7 coverage in-house, you are not hiring one person. You are hiring four or five, on a rotation, with overlap. For a 40-person company that’s not a budget decision. It’s simply not available.

Add turnover

IT staff move. When your one person leaves you get:

• Recruitment costs — agency fees commonly run 15–25% of first-year salary, and doing it yourself costs management time instead of money
• A vacancy gap — typically one to three months for a technical hire, during which you have no coverage at all
• Ramp time — three to six months before a new hire knows your environment
• Knowledge loss — this is the expensive one. Undocumented environments walk out the door. The new person rediscovers your infrastructure by breaking it.

Amortised across an average tenure of two to three years, turnover realistically adds $8,000–$15,000 per year to the true cost of the role.

The uncomfortable middle

Add it up and the honest annual cost of an internal helpdesk for 40 people lands somewhere in the range of $100,000 to $125,000 — for 24% coverage, one person’s knowledge, and no depth in any specialisation.

That last point deserves attention. Your generalist is simultaneously expected to be your network engineer, your security analyst, your cloud architect, your procurement lead, and the person who resets Debbie’s password. Nobody is good at all five. The gaps show up as things that never quite get done: the backup that was never tested, the firewall rules nobody has reviewed since 2023, the offboarding checklist that exists in principle.

When in-house is still the right answer

Outsourcing is not automatically correct, and any provider who tells you otherwise is selling rather than advising. Keep it internal when:

• You have genuinely unusual systems. Custom manufacturing controls, specialised lab equipment, legacy line-of-business software with no vendor support. A generalist provider will be slower than someone who lives in it daily.
• IT is your product. If you’re a software company, your infrastructure knowledge is core competency, not overhead.
• You’re above roughly 150 staff. At that size an internal team of three to four becomes both affordable and properly staffable, and the coverage maths changes completely.
• Compliance or contracts require it. Some procurement terms specify internal control of certain systems.

When it isn’t

Below roughly 100 staff, the coverage arithmetic is brutal and doesn’t improve. You cannot buy 24/7 with one salary, and you cannot buy specialist depth with a generalist. The question isn’t really “in-house or outsourced” — it’s whether you’re honest with yourself about what a single hire actually delivers.

The version that works for most companies this size is a hybrid: one internal person who knows the business, the relationships, and the priorities, backed by an external team providing after-hours coverage, specialist escalation, and vacation relief. Your person stops being a single point of failure. They also stop being on call every night of their life, which is the other reason IT staff leave.

Run your own numbers

Take our line items and substitute yours:

1. Base salary → your actual market rate
2. × 1.25 to 1.35 for employer costs and benefits
3. + your real tooling spend, per seat, annually
4. + turnover, amortised over expected tenure
5. ÷ your actual required coverage hours

Then compare against a per-seat managed services quote at the same coverage level. Not at 40 hours — at the coverage you actually need. Most comparisons are unfair because they’re comparing an in-house cost for 24% coverage against an outsourced cost for 100%.

If the numbers land close, keep it in-house. Proximity and business context are worth real money, and we’d rather tell you that than win a deal you’ll regret.

---

DNOTCH provides managed IT and helpdesk support to businesses across Manitoba and Canada, including after-hours and overflow coverage for companies with internal IT teams. If you’d like a straight comparison against your current costs, book a call — we’ll do the maths with you, including the case where staying in-house wins.