What an internal IT hire actually costs
The advertised salary is about three quarters of the real number, and the real number still buys you nineteen per cent of the calendar. Here is the arithmetic, with every assumption on the page so you can change it.
At some point every growing Australian business runs the same comparison: hire someone, or engage a provider. It is usually done on two numbers — an advertised salary against a monthly fee — and those two numbers are not comparable in any useful sense.
This works through the cost of managed IT services in Australia against the genuine cost of an internal hire and of paying by the hour. Every figure below is either sourced or labelled as an assumption you should replace with your own. We sell one of the three options, so the assumptions are stated rather than buried.
The salary is not the cost
SEEK's data for mid-2026 puts a systems administrator in Australia between $95,000 and $115,000, with network administrators a little below that. Take the midpoint — $105,000 — and add what an employer actually pays on top.
| Line | Basis | Annual |
|---|---|---|
| Base salary | SEEK midpoint, systems administrator, mid-2026 | $105,000 |
| Superannuation | 12% guarantee rate, from 1 July 2025 | $12,600 |
| Payroll tax | NSW 5.45%, only above the $1.2m wage bill threshold | $5,723 |
| Workers compensation | Assumed 1.5% — varies by state and classification | $1,575 |
| Recruitment | 17.5% of salary (RCSA 2025 survey average), spread over 3 years | $6,125 |
| Device, phone, training | Assumption — replace with your own | $4,000 |
| Total | Before a single tool | $135,023 |
Scroll the table sideways to see every column.
That is 1.29 times the advertised salary, and it is a conservative build. It excludes leave loading, any bonus, the cost of the manager's time supervising a function they do not understand, and the recruitment cost repeating sooner than three years — which in this market it often does.
Two lines are worth checking against your own situation. Payroll tax only applies above your state's threshold — roughly $1.2 million of annual wages in New South Wales, $900,000 in Victoria, $1.3 million in Queensland — but if you are already over it, every new hire is taxed at the margin, so the line belongs here. Workers compensation varies by state and industry classification; 1.5 per cent is a placeholder for an office-based role, not a quote.
Call it $135,000. Now look at what it buys.
The coverage problem, which is the real one
One full-time employee works 38 ordinary hours a week. Under the National Employment Standards they also get four weeks of annual leave, ten days of paid personal leave, and public holidays.
- 52 weeks × 38 hours = 1,976 hours
- less four weeks annual leave = −152
- less ten days personal leave = −76
- less around eleven public holidays = −84
- 1,664 hours actually at work
There are 8,760 hours in a year. One internal hire covers 19 per cent of the calendar — and none of the 19 per cent falls at 2am on a Sunday, which is when intrusions are deliberately run.
The question is not whether you can afford $135,000. It is whether $135,000 for one fifth of the year, with no cover during annual leave and no second opinion during an incident, is the best available use of it.
There is a second-order version of this that costs more than the gap itself. One person cannot be current on identity, endpoints, networking, backup, cloud and security simultaneously. They will be genuinely strong at two of those and improvising the rest, and you will not know which two until something goes wrong in one of the others.
And when they resign — the average tenure in this role is not long — the environment leaves with them unless someone insisted on documentation, which is not usually a thing an overloaded sole IT person has time for.
The stack does not arrive with the person
This is the line item almost every internal-hire business case omits. An employee is labour. They are not remote monitoring and management, ticketing, endpoint detection and response, patch automation, backup with immutable storage, documentation, or a security operations capability watching logs overnight.
Every one of those is licensed per user or per device, and most are priced for volume that a single business does not have. A managed provider spreads the same tooling across its whole client base; you would be buying it at list, for one site.
So the honest internal-hire number is $135,000 plus a per-seat tooling bill, and the honest comparison is against a managed fee that already includes it. If a provider cannot tell you which tools are inside the fee and which are extra, that is the first question in our seven-question list.
What paying by the hour really costs
Break-fix looks like the cheap option because you only pay when something breaks. Current Australian rates run roughly $150 to $250 an hour, lower in regional areas, often with a call-out fee of $80 to $150 on top.
Take a 60-person business. Assume — and this is your number to change — one support request per person per month at an average of 45 minutes:
- 60 tickets × 0.75 hours = 45 hours a month
- × $180 an hour = $8,100 a month, or $97,200 a year
Which is already more than the internal hire, for reactive work only. Halve the ticket assumption and it is still $48,600 — and it still buys nothing proactive.
The deeper problem with break-fix is not the rate. It is that nobody is being paid to prevent anything. Patching, backup verification, monitoring, access reviews and asset management are the work that stops incidents, and none of them generates a call-out. They either do not happen, or they get quoted as projects, at which point you are paying managed-service money without a managed service's accountability.
It also means your provider's revenue rises when your environment is unstable. That incentive is worth thinking about even when everyone involved is honest.
Comparing a managed quote properly
Managed services are priced per user or per device, monthly, and we are not going to quote you a per-seat figure here — it would be meaningless without knowing your environment, and any provider publishing one is either narrowing the scope until it fits or planning to reprice you later.
What you can do is make the comparison structurally fair. Take the annual managed fee, then put it next to this:
- Fully loaded internal cost. Salary × 1.3, using the table above with your own state's payroll tax position.
- Plus the tooling that the hire does not come with, quoted per seat at list price for your headcount.
- Plus the coverage gap priced honestly. Either you accept 19 per cent coverage, or you buy after-hours support on top, or you hire a second person.
- Plus the concentration risk. What happens for the four weeks a year they are away, and for the two months after they resign.
Then check the managed quote for the things that make it comparable at all: what the pricing unit is, what is explicitly excluded, whether remediation of findings is inside the fee, and what leaving costs. Those four are where two superficially similar quotes diverge, and we work through them in how to choose managed IT services in Australia.
Be equally sceptical of us here. A managed fee has its own hidden costs — onboarding and transition, anything named in the exclusions schedule, project work, and exit assistance if it was not priced before you signed. A provider unwilling to write those down has not removed them.
Which model is actually right for you
There is no universal answer, and a provider telling you otherwise is selling. Roughly, by size and obligation:
| Situation | Usually right | Why |
|---|---|---|
| Under about 10 seats, no regulatory obligation, downtime tolerable | Break-fix or a light managed plan | Ticket volume is too low to justify a fixed fee, and the exposure is genuinely small |
| 10–250 seats, or any size holding client, health or student data | Managed services | You need the tooling and the coverage, and cannot buy either efficiently alone |
| 100+ seats, or daily on-site need, or a bespoke line-of-business system | Internal person plus a provider | Someone who knows the business, with a bench behind them for depth and cover |
| 250+ seats and able to fund three or more IT staff | Internal team, with specialists bought in | Three people is roughly where real coverage and a second opinion become possible |
Scroll the table sideways to see every column.
Note where the third row sits. Co-sourcing is frequently the right answer above about a hundred seats and it gets proposed least often, because it suits neither a provider selling a full contract nor a manager building a team. It is worth asking both parties about explicitly.
The number that usually decides it
When we run this with a business, the figure that changes the conversation is rarely the headline cost. It is the coverage percentage. Once somebody works out that $135,000 buys 19 per cent of the year and a single point of failure, the question stops being “can we afford a provider” and becomes “what do we actually need covered, and when”.
That is a much better question, and it is answerable with your own numbers.
Salary figures are from SEEK's published Australian salary data for mid-2026; recruitment fee ranges from the RCSA's 2025 fee survey; superannuation, payroll tax thresholds and leave entitlements from the relevant ATO, state revenue office and National Employment Standards positions current at the time of writing. Rates and thresholds change — check the current figures for your state before relying on the arithmetic, and this is general information rather than financial or employment advice.
Questions
Is a managed provider always cheaper than hiring?
No. Above roughly 250 seats, or where you need somebody physically present every day, an internal team usually wins on cost and on responsiveness. The crossover is not a fixed number — it depends on how much on-site presence you need and how much after-hours coverage you actually require.
What is close to universal is that a single internal hire, with no provider behind them, is the weakest of the available options at almost any size. It buys the cost of a specialist and the coverage of one person.
What about hiring a junior and training them up?
It lowers the salary line and raises everything else. A junior needs supervision from someone competent to give it, which you do not have, and the coverage arithmetic is unchanged — one person is still one person.
Where it does work well is alongside a provider: the junior handles the on-site and first-line work they can learn from, the provider carries the depth and the after-hours. That is the co-sourced model in row three of the table.
How do we compare quotes priced on different units?
Normalise them yourself before you look at anything else. Pick one unit — per user is usually cleanest — and ask every provider to requote on it. Some will resist, which is informative.
Then multiply by twelve and compare the annual figure against the fully loaded internal number, not against the salary.
Does moving to a provider mean making our IT person redundant?
It should not, and we would usually advise against it. The person who knows your business, your line-of-business software and your people is worth keeping; what they need is depth behind them and cover when they are away.
Where that goes wrong is when the boundary is assumed rather than written down. Agree it function by function, in a table, before anything starts.
What is the cheapest defensible option for a very small business?
Under about ten people with no regulatory obligation, a light managed plan covering identity, endpoint protection and tested backup, plus break-fix for everything else, is usually the sensible floor. Those three remove most of the probability and most of the consequence.
What we would not recommend is pure break-fix with no backup verification. That is not a cheaper model, it is a deferred cost.
What the fee covers that a hire cannot
Not a cheaper headcount — a different shape of cover. This is what sits inside one agreement instead of on top of a salary.
- The other 81 per cent. Rostered Australian engineers overnight and on public holidays, with the authority to contain rather than escalate.
- The tooling, at our volume. Monitoring, endpoint detection, patch automation, immutable backup and documentation, priced across a client base rather than one site.
- Depth, not one person. Identity, cloud, networking, endpoint and security specialists on the same account — and cover when somebody is on leave.
- Targets in the schedule. P1 acknowledged in 15 minutes against a four-hour resolution target, reported monthly whether or not we hit it.
- Named exclusions, priced exit. The hidden costs written down before you sign, with your tenants and licences in your name throughout.
- Co-sourcing if that fits better. If you should keep your internal person, we will say so and define the boundary function by function.
Get the comparison for your headcount
Tell us your seat count, your state and what you are obliged to protect, and we will build the same table with your numbers in it — including the case for not using us.