Build vs. Rent

You have one objection left, and it is a fair one. Everything so far points toward owning the tools your practice runs on. But owning sounds like work, and there is a tool you can rent this afternoon that does the job tonight. Why build when you can rent?

This chapter answers that, and it is the last thing standing between you and the part where you start making things. The answer is not that renting is always wrong. It is that renting carries a price most firms never see, and the thing that once made renting the only sensible choice has quietly stopped being true.

The bill you stopped reading

Start with the cost you can see, then the one you cannot.

Software you rent arrives as a subscription. Thirty dollars a month here, two hundred there, a per-seat fee that climbs every time you hire. Each one is small enough to approve without thinking, which is exactly the trouble. They do not arrive all at once. They accumulate, one reasonable yes at a time, until a firm that would never write a single large check is quietly writing a very large one every month, spread across a dozen line items nobody has reviewed in a year.

Sit down sometime and add them up. Almost everyone who does is startled by the number. It is common for a small business to find it is spending many thousands of dollars a month on software it stopped evaluating long ago, some barely used, some forgotten. That is the visible cost, and it is larger than you think.

The invisible cost is worse, and you already know it from Chapter 2. When you rent, you own nothing. The day you stop paying, you are left with nothing: no tool, and often no clean way to get your own data back out. The price only ever rises. The vendor decides what the tool does next, not you. And the whole time, your information has been living on their machine, making their product more valuable. You are not building an asset. You are renting one, forever, and feeding it your edge as you go.

Why renting used to make sense

None of this is new, so why did renting ever win? Because the alternative was out of reach.

Building software used to require engineers, real money, and months of time. For a law firm, that was not a close call. Renting was not laziness; it was the only rational choice, because building your own anything meant a team, a budget, and a project that might never ship. So firms rented, and they were right to.

That world is gone. Three things changed at once, and together they flipped the math. The tools got simple enough that a determined non-engineer can assemble real, working software, which is the whole premise of the back half of this book. The hardware got cheap enough that a single machine in your office can run capable AI, the kind of thing that would have filled a server room a few years ago. And strangest of all, the AI now helps you build: the thing you are making is also your assistant in making it. The wall that justified renting did not get lower. It fell over.

What owning buys you

When you build, the cost moves from forever to once. You spend some effort up front, perhaps some money on a machine, and then the thing is yours. Nobody takes it away. The price does not climb. You use it as much as you like with no meter running, because there is no per-seat tax on a tool you own.

And it fits you. A rented tool is the average of a thousand firms, shaped for none of them. A tool you build is shaped for exactly how your practice works, because you built it. Best of all, and this is the thread running through the whole book, your data stays yours and compounds. The longer you run your own tools on your own work, the better they get at the things only your practice knows, and the further ahead you pull.

The honest trade-offs

This book is worthless to you if it only sells. So here is the other side, plainly. Building costs time and attention up front. It is front-loaded effort in exchange for not paying rent forever, which is a good trade, but not a free one. A model you run yourself is a step below the very best cloud models, fine for the great majority of work and occasionally outclassed on the hardest tasks. Someone has to keep the machine running, which is closer to maintaining a computer than running a data center, but it is not nothing. And you will not replace everything you rent in a weekend, nor should you try.

Which points to the rule that keeps all of this sane: rent the commodity, own the core. Some software is generic, touches none of your data and none of your edge, and renting it is perfectly fine. Keep doing it. The things to own are the ones that are your data, your client information, your differentiated way of working, the assets that make your firm your firm. Those you never want to be renting, because renting them means handing your edge to a landlord one month at a time. Own the core. Rent the rest. That is the whole policy.

The end of the why

That closes the case. You know why your data is the thing that matters, where it really goes, when you are safe and when you are not, and now why owning the core beats renting it. There is nothing left to be convinced of.

The only thing left is to do it, and it is far easier than the noise made it sound. Close this half of the book, find a desk, and turn the page. In the next chapter you set up your own model, and all of this stops being talk and becomes a machine humming quietly in your office, answering you with the internet switched off.