Buy, rent or build your business software
If your process is the same one everyone in your industry runs, buy or rent something that already exists: rebuilding it wins you nothing. If the process is the reason customers pick you over the shop down the road, that part gets built. And the question that settles it isn't the monthly price, it's what it costs to pull your data out and walk away three years from now.
The process you share with your whole industry is not worth building
Invoicing, taking payment, booking appointments, sending reminders, keeping the books, getting an estimate signed. Thousands of businesses just like yours do all of that the exact same way. There's no edge hiding in there. Build your own invoicing system and the best case is that you arrive where any paid tool already sits, two years late, with one person on earth who knows how to fix it.
The test takes a minute. Take the process and ask: if the shop across town did this exactly the way you do, would you lose a single customer? If the answer is no, buy it. If the answer is that you'd lose half your book, now you've found something that's actually yours.
Eight years working inside businesses across seven industries left me with one constant: almost nobody loses money using a generic tool for a generic process. The money burns the other way around, building the boring part and renting the part that was the differentiator.
Renting monthly isn't cheap, it's reversible
Subscriptions get sold as the budget option and they aren't. What you're actually buying with a monthly fee is the right to be wrong without it hurting. Try it for three months, it doesn't fit, cancel, and you're out three months. When you still don't fully know how your own operation runs, that's worth a lot.
The trouble starts when the subscription stops being an experiment and becomes the backbone. That's when the price starts growing with you: per seat, per contact, per transaction, per volume. You grow, and the vendor charges more because you grew. Chain three or four tools together that way and the monthly number stops looking like the one you signed up for.
Working rule: if you're still figuring it out, rent, every time. If you've used the same thing daily for two years and the bill climbs purely because business is good, that's when running the numbers on building makes sense. Not before.
Building pays off when bending your business costs more than the build
Off-the-shelf never fits perfectly. The question isn't whether you'll have to adapt, it's how much the adapting hurts. If the compromise is renaming a field or accepting a different order of steps, adapt. If the compromise is dropping the thing that makes you different, that's not adapting, that's shrinking to fit the software.
There's one signal you can measure in hours: somebody spends hours every week retyping data from one place into another, or redoing by hand the same step the tool doesn't cover. Those hours have a price and you pay it monthly, same as a subscription, except it never shows up on an invoice. Once that number gets close to what it would cost to build the missing piece, the math has already flipped.
Build when the process is stable and you already understand it cold. Custom software built on a process that's still moving is the most expensive way to discover you didn't know what you wanted. I built a CRM for lifestyle and concierge agencies and a multi-currency finance platform for exactly that reason: nothing off the shelf handled that way of working without breaking it. The CRM side of it is laid out in when a business actually needs a CRM.
The cost nobody prices in is the cost of leaving
All three paths have an exit, and none of them puts it on the price tag. You can leave a subscription fast, but only if you can take the data with you. Before you sign, ask them to show you how a full export works, not whether one exists. Make them show you. If the export is a PDF or a file nothing else can read, you're locked in even though canceling is one click.
Leaving custom software is a different problem: the code is yours and nobody raises your price, but if the person who built it vanishes and left nothing written down, you own an asset only one human ever understood. So when you commission anything custom, insist from day one that the code lives in your own account, that every login is in your name, and that there's a written explanation of how it works. That's not distrust, it's the floor. There's a full list of those questions in what to ask before hiring someone to build for you.
The honest decision gets made on a three-year view, not a three-month one. Add the cost of leaving to each path and you'll see that today's cheap option is sometimes the expensive one, and that the thing that cost more upfront is sometimes the only piece still yours when the business changes shape.
Frequent questions
If my budget is tight, do I always rent?
Usually yes, but for the right reason: not because it's cheap, but because on a tight budget you can't afford a bet that goes wrong and can't be undone. Rent it, run it for a year, and write down every time the team has to do something by hand because of the tool. That list is what later justifies building, or what confirms you never needed to.
How do I know if my process is genuinely different or I'm just used to it?
Ask whether the customer notices. If the unusual step changes what the customer gets, how fast they get it, or what they pay, it's a real advantage. If the customer would notice nothing and the only argument is that it's always been done this way, that's habit, and changing a habit is far cheaper than building around one.
Can I mix all three?
That's the normal setup and the healthy one: bought for accounting and invoicing, rented for email and scheduling, custom only for the part that makes you different. The one condition is that the pieces have to talk to each other. Before adding any new tool, check that information can move in and out of it automatically; if it can't, you're buying an island.