Rent vs. Own: The Five-Year Math

Most build-vs-buy decisions are settled in year one, where off-the-shelf can have a lower upfront cost. A subscription starts Monday for the price of a dinner; a build is months of cost before it returns a thing. If the comparison ended there, nobody sane would build.

The comparison doesn’t end there. Software is a multi-year commitment whichever way you go. Whether the curves cross depends on licensing, operating costs and the cost of the build.

The comparison everyone gets wrong

The instinctive framing is build cost vs. zero: the new system costs money, the status quo is free. It isn’t free. The status quo is renting a near-fit forever, plus the near-fit tax you pay to keep it usable: the workaround headcount, the reconciliation hours, the integration glue.

So the real comparison is:

Renting (licences plus escalation plus workaround labour, every year, forever) versus owning (a one-time build plus modest maintenance, on an asset you keep).

Put both on a five-year horizon and the question stops being “which is cheaper to start?” and becomes “which is cheaper to live with?”

A worked example

Numbers are illustrative. Plug in your own. Take an operations platform for an 80-person team.

Rent. 80 seats at $120/seat/month is about $115k in year one. Assume a 10% annual increase in licence costs; workaround labour stays at $90k a year in this example. Add the workaround layer the near-fit requires (call it 1.5 full-time-equivalents at $60k, so $90k a year), because that cost belongs to the rented tool even though it hides in payroll.

• Year 1: about $205k
• Five-year total: roughly $1.15M. Year six keeps running.

Own. A fit-built system for that operation: a one-time build of, say, $300k, plus 15% of that per year in maintenance and iteration (about $45k), starting in year one.

• Year 1: about $345k, more than renting, as expected.
• Five-year total: $525k ($300k plus five years at $45k). You own the asset.

Under these assumptions, owning pulls ahead by the end of year two: cumulative rent is about $422k versus $390k to own. After it, every year widens the gap. The owned system is also doing something the rental never does: appreciating in fit as you extend it, instead of escalating in price as you renew it.

This simplified comparison includes rental licences and workaround labour, versus the build and annual maintenance. It assumes the owned system removes the modeled workaround labour. Hosting, migration, security, training and any residual operational labour are not separately costed; add them to either option where they are not already included. The figures are undiscounted, with no inflation applied to labour or maintenance.

When renting is the right answer

The math doesn’t always favour building. Pretending it does would be the same dishonesty in reverse. Rent when the fit gap is small (the tool genuinely matches how you work), when the function is a true commodity (email, payroll runs, accounting ledgers), when scale is low enough that licences stay trivial or when the domain is changing so fast that you would be rebuilding anyway. For these standard functions, off-the-shelf may be the better fit.

Build when the misfit is structural and permanent: when the software runs a core operation, the workaround layer has become a department and the customisation quotes have started arriving with commas in them.

Run it before you renew

Before the next multi-year term, build the table yourself. Three inputs decide it: your real seat count, the renewal escalator in your contract and the headcount that exists only to bridge the gaps. Most teams have never added the third line. It is the one that moves the answer.

The question, once the curves are in front of you, is the one the thesis ends on: not “can we afford to build?” but “can we afford not to own?”

An Advisory assessment builds this model on your actual numbers and returns the call (stay, extend or build) with the five-year figure attached.