8 June 2026
Why Most Internal Software Should Be Built, Not Bought
I'd build the software that runs your business before I'd rent it. The argument on the merits — why owning beats renting, and why the build is smaller than you think.
The software that runs your business internally — the thing only your company does the way your company does it — should be built and owned, not rented from someone else. That’s my position, and it’s the opposite of the default advice, which is to find the right SaaS tool and pay for it forever. I think the default is wrong about where the value is, wrong about what the build actually costs, and wrong about what you’re really buying when you buy.
I want to argue that, not just assert it, because the assertion sounds like a builder talking his own book. It isn’t. I’ll tell you exactly where buying is the right call. But for the core systems that carry how your business actually operates, renting is a worse deal than its monthly price makes it look, and the reasons are structural, not a matter of taste.
The default thinking, stated fairly
The case for buying goes like this: building software is expensive and risky, off-the-shelf tools are mature and cheap by comparison, someone else maintains them, and you get to focus on your actual business instead of becoming a software company by accident. Every one of those points is true in isolation. That’s why the advice is so common — it’s not stupid, it’s just incomplete in a way that costs people enormous amounts of money over time.
What it leaves out is that the comparison is rigged at the moment you make it. You compare the price of buying today against the price of building today, and buying wins, because a finished product always looks cheaper than an unbuilt one. But you’re not buying today. You’re buying for the life of your company, and on that timeline the numbers invert.
Renting never ends, and that’s the whole problem
When you buy software, you don’t buy software. You rent access, monthly, with no terminal date. The price only moves one direction — up per seat as you grow, up per tier as the feature you need migrates behind the next paywall, up whenever they decide your plan is now the legacy plan.
A build is the opposite shape. It’s a cost concentrated up front that then stops. You pay to create the thing, and after that you pay only to maintain and extend it — work you control, on your schedule, often a fraction of what you were renting. The rented tool’s cost line goes out flat-or-rising to the horizon. The owned tool’s cost line spikes once and then drops and stays low. Anywhere past the first year or two, the owned line is underneath, and it stays underneath for as long as the business exists. You’re not comparing a big number to a small number. You’re comparing a number that ends to a number that doesn’t.
People feel this and discount it anyway, because the monthly figure is small and the build figure is lumpy, and a small recurring pain reads as cheaper than a single visible one. It isn’t. It’s just spread out so you stop noticing it.
You’re paying for a product built for the average of everyone
The second structural problem: an off-the-shelf tool is built for the average of every company that might buy it, which means it is built for no company in particular, and certainly not for yours.
Most of the product exists to serve use cases that aren’t yours. You pay for all of it and use a slice. Worse, the slice you use is shaped to the vendor’s idea of how the work should go, not yours — so you bend your actual process to fit their software, and call the friction “best practice” because the alternative is admitting the expensive tool doesn’t fit. The thing that’s genuinely yours, the workflow that is your business, is exactly the thing no general tool will ever model well, because modelling it well for you would mean modelling it badly for everyone else they’re selling to.
A built tool is the inverse. It does the 20% you actually do, shaped the way you actually do it, and nothing else. It’s smaller, it fits, and it doesn’t drag a thousand features you’re funding for strangers.
The build is smaller than the word “build” makes it sound
The objection I get here is always cost and risk: building is a big scary project, buying is a checkout. But “build” is doing a lot of unfair work in that sentence. You’re not building the SaaS product, with its ten thousand customers’ worth of edge cases and configuration and multi-tenancy. You’re building the part you use — the lean core that does your job and stops.
That’s usually weeks of focused work, not a moonshot, because the thing you genuinely need is far smaller than the thing you were renting. And here’s the part that makes it click: the money you were about to spend on the first year of subscriptions very often covers that build outright, so the first year costs you about what renting would have anyway. The difference shows up afterwards. With SaaS, every subsequent year you start paying again from zero with nothing banked; with the owned system you already hold the asset, and you spend that year extending it instead of renting it. The build doesn’t just stop costing — it starts compounding, because every system you own can feed the next one. That’s why “expensive to build” is usually a failure of scoping, not a fact about building.
Where I’d buy, and mean it
I am not telling you to build everything. That’s how you end up writing your own payroll engine to save thirty pounds a month and inheriting a legal liability you never needed. There’s a clean line.
Buy the commodity: the mature, undifferentiated, often regulated things that everyone needs and nobody competes on — accounting ledgers, payroll, card processing, email. Off-the-shelf is cheap relative to the build, hardened by millions of users, and getting it wrong has consequences a custom build only adds risk to. Building those is ego, not leverage. Buy them, wire them in, forget them.
Build the thing that’s actually yours: the workflow only your business runs, the way you serve your clients that nobody else serves them. That’s the thing no tool will fit and the thing worth owning, because it’s where your leverage lives. The rule is that simple — buy the commodity, build the edge — and most people get it backwards, renting the edge and agonising over the commodity.
What this means for you
Next time the answer to a problem looks like “find the right tool and subscribe,” stop one step earlier and ask whether this should be a thing you rent at all. Is it your edge or a commodity? Is the part you actually use small enough to own? More often than the industry wants you to believe, the honest answer is that you needed a lean system you control, not a subscription you’ll feed forever — and the build was smaller and cheaper over any real timeframe than the renting you were about to sign up for.
Underneath all of it is the thing I keep coming back to: own your data and systems and every later move gets cheaper, because you can connect them, automate across them, and build on top of them on your terms. Rent them and you’ve capped what your business can ever do with its own operation.
This is one of my essays — the hidden cost of renting your operations takes the same conviction and follows the data instead of the money, and what “MVP” gets wrong is about getting the build small enough that this argument actually holds. If you want the decision turned into a procedure — when to build, when to buy, with the math — that’s my build-or-buy playbook.
If you’re weighing this for a specific system and want a straight read on which side of the line it sits, talk to me. And if it turns out you need a whole system built and run by a team rather than one lean thing you own, that’s exactly what my agency, Digital Royalty, is for.