Blog / budgeting
What should a small business budget for IT?
Percentage-of-revenue benchmarks fall apart under 50 staff. Here is a simpler way to build a technology budget you can actually defend to your board.
Every few months someone asks me what percentage of revenue they should be spending on technology. There are published benchmarks for this, usually somewhere between two and six per cent, and I have never once found them useful for a company under fifty people.
The number moves too much. A twelve person accounting firm and a twelve person machine shop have almost nothing in common in what they need. One is buying software licences and good backups, the other is buying network gear that survives dust and a building full of equipment that has to keep running. Averaging them produces a number that describes neither.
So I build these from the bottom up instead. It takes an afternoon and the result is a number you can explain, which matters more than a number that matches a survey.
Start with what you already spend
Before deciding what you should spend, find out what you do spend. This sounds obvious and it is nearly always the part that gets skipped.
Pull twelve months of statements and mark every line that is technology. Internet and phone. Every software subscription, including the ones a department signed up for on a credit card. Hosting and domains. Your accountant's portal fee. The alarm monitoring. Anyone you pay for support, whether a contract or an hourly invoice. Hardware you bought when something died.
Most owners are surprised twice. First by the total, which is usually higher than they guessed. Second by how much of it is subscriptions nobody remembers approving. I have yet to review a list like this and not find at least one thing that was cancelled in spirit but never in practice.
That total is your real baseline. Everything else in the budget is a decision on top of it.
Four buckets, not one line
A single "IT" line in the budget hides the thing you need to see, which is that these costs behave very differently.
Running costs are the ones that continue whether or not you do anything: connectivity, licences, hosting, monitoring, the support arrangement. They are predictable and they are the easiest to trim, because most of the waste lives here.
Replacement is the money set aside for equipment reaching the end of its life. More on this below, because it is the line almost nobody has.
Projects are the discretionary work: the new phone system, the move to a different platform, the cameras for the second building. These come and go and they should be planned separately from the running costs, not smuggled in beside them.
Risk reduction is backups, testing those backups, security work and the boring maintenance that produces nothing visible. It is the first thing cut and the most expensive thing to have cut. If you want a way to sanity check what you are buying here, I wrote about auditing a backup strategy as an owner.
The line most budgets are missing
Hardware does not fail on your schedule. It fails when the machine is four or five years old and someone is in the middle of month end.
Business laptops and desktops typically run $1,200–2,500 CAD each and last three to five years in normal use. Network switches, access points and firewalls run longer, often five to eight years, though the firewall usually goes first because support and security updates for it stop before the hardware does. Cameras and recorders sit somewhere in between.
Count what you own, decide roughly how long each class of thing lasts, and divide. If you have twenty computers on a four year cycle, that is five machines a year, every year, whether or not you feel like buying them. Put that number in the budget as its own line and fund it monthly.
Companies that do this stop having emergency purchases. Companies that do not end up buying whatever is in stock at retail the afternoon something dies, at a bad price, with no time to check whether it fits.
When the total is too big
It usually is, the first time. That is fine, because now you can argue about the right things.
Sort the list by what breaks if you stop paying for it. Anything that stops the business cold stays. Anything that only stops one person from being mildly annoyed can wait a year. The middle is where the real conversation happens, and it should involve the people who use the thing, not just the person signing the cheque.
Be careful about deferring the replacement line to make the year look better. Deferring it does not remove the cost, it moves it into next year on top of next year's, and eventually into an unplanned month where it does the most damage. That is technical debt in its most literal form, and it accrues.
Revisit it once a year
Not more. A budget you reopen every quarter turns into a running negotiation and stops being a plan.
Once a year, take an hour: what did we actually spend against each bucket, what did we buy that we did not plan for, and what is now old enough to land in the replacement line next year. If the answers keep surprising you in the same direction, the budget is wrong in a way you can now fix.
If you are past the point where this fits in an afternoon, or you want someone to build the first version with you and pressure test the numbers before they reach your board, that is the kind of work I do on a fractional basis. Engagements are quoted as a fixed fee, hourly or a monthly retainer depending on what suits. Get in touch and tell me roughly how many staff and locations you have, and I will tell you whether this is an afternoon or a project.