If the tool where you keep your clients disappeared tomorrow, would you lose the tool or would you lose the clients? That question separates two things modern platforms deliberately blur: what you rent and what you own. Losing the first costs you an afternoon. Losing the second is what some businesses never recover from, because a client relationship cannot be rebuilt from memory.
You do not need to be technical to answer it. You need half an hour and some honesty about where things actually are.
1. Nobody designed your operation. It accumulated.
The businesses that end up most tangled are not the disorganised ones. They are the ones that kept solving each problem well, one at a time, as it showed up.
You needed to schedule and found something free that worked. You needed to send quotes and somebody recommended something else. You needed somewhere to keep the photos of every finished job and used whatever was already open. A marina running two hundred slips today started on a spreadsheet built by an employee who no longer works there.
Every decision made sense at the time. The result, five years on, is an operation spread across seven places nobody chose as a set.
Our take
This takes the blame off, and it is worth saying plainly: it was not a failure of judgement, it was the sum of many small right calls that never got looked at together.
Looking at them together once is what is worth doing. It is the same logic as the map of the five disciplines: loose pieces feel fine and only make sense as a whole.
2. The distinction platforms work hard to keep you from seeing
You rent the tool. You own the asset.
The tool is the program where you write, the system where you schedule, the app where you keep your clients. All of them are rented, even at a price of zero. Not being charged does not make it yours: it means you are paying some other way, or you have not been charged yet.
The asset is what would be left if all of them vanished tomorrow: your client list, the history of what you discussed with each one, what you wrote, and the way you do your work.
Our take
The confusion is expensive and almost nobody sees it, because the tools are designed so you do not. Everything feels like yours: your dashboard, your contacts, your history. The word "your" is everywhere.
And there is a commercial reason behind it, which is not malice, it is design: a tool you can leave easily is a tool you will leave. Nobody builds that on purpose.
This week's example, and why it does not matter how it ends
A few days ago it was reported that Nvidia would acquire Hugging Face, the repository where much of open-source AI lives, for a figure in the billions.
Here is the caveat: by the time you read this it may be signed, it may have collapsed, or it may still be hanging. Between writing and publishing, the number already moved once.
And it does not matter how it ends. That is the point.
Because what did happen, and that part does not depend on the outcome, is that for days a lot of people who depend on that platform did not know whether their provider was about to change. Nobody told them. Nobody asked them. And there was nothing they could do about it.
That is what you are buying when you build your operation on a tool you did not choose: borrowed certainty. And it can end without warning, from a decision made in a room you were not in.
Watch out for the sibling problem, which is the same neglect through another door: hacking your business now costs $50, and what exposes a small business is almost never a sophisticated attack, it is having things scattered where nobody is watching them.
3. The moving test
One question, four places. No steps, no order, nothing to install.
The question: if this disappears tomorrow, do I lose the tool or do I lose the asset?
And these are the four places to ask it, with what you lose in each if the answer is the bad one:
| The place | What it is | If the answer is bad, you lose |
|---|---|---|
| Contacts | names, numbers, who is a client and who asked once | the whole list, and every future sale that depended on calling them back |
| Conversations | what you discussed with each one, what they asked for, what you promised | the context. You start over with people who already trusted you |
| Content | what you wrote, the photos of your work, your standard quotes | years of proof that you can do what you say |
| Process | how you do what you do, in what order, what gets checked before delivery | what walks out with the person who quits, not just with the tool that shuts down |
A boatyard sees it fastest in the third row: the before-and-after photos of every job are its strongest sales argument, and they usually live on the phone of whoever took them. They are also exactly what it needs when somebody outside is about to speak well of it, which is the work we cover in the format AI cites most is a list.
What the test does not ask of you: it does not ask you to switch tools, hire anyone, or migrate anything. It asks you to know where you stand. Switching platforms without doing this is moving house without packing boxes.
We ran it here and came out worst on the fourth row. A good part of our process lived in the head of whoever did it, and we found out the hard way: we discovered a module we use in every article since August that was not written down anywhere. We are fixing it, and it is the least glamorous and most useful work of the quarter.
4. Free tools do work, just not for everything
So this does not read as free tools being bad, because they are not and we use several.
A free tool is a perfectly sound decision when what you keep in it is not an asset. An image editor, a file converter, a document reader: if it shuts down tomorrow, you find another one and lost nothing.
It becomes a risk when it starts accumulating something you cannot rebuild. The classic case starts as "just to try it" and three years later holds the entire relationship with two hundred clients.
Our take
The rule we use, and you are welcome to it: free is for what you can replace in an afternoon. What you cannot replace in an afternoon gets paid for, or at least backed up.
It is not a budget rule. It is a rule about what you are willing to lose.
Todd and Naty's corner Todd: From the cold side: nobody is going to warn you. The company hosting your things will not consult you if it gets acquired, if the price goes up, or if the product shuts down. You control none of those three decisions. The only thing you control is how much of what is yours would leave with them. Have you ever put that number in writing? Naty: What makes me sad is watching someone lose twenty years of work over a folder. And it happens. I have seen excellent businesses left without the photos of everything they built, because those photos were on the phone of somebody who left on bad terms. Nobody mourns an app that shuts down. You mourn what was inside it, and by the time that hurts there is nothing to be done.
The one thing to do this week
The two-column sheet. Twenty minutes, once, and you need nothing but paper.
On the left, every tool your business uses. All of them, including the free ones and the ones only one person touches. There will be more than you think.
On the right, for each one, what stops if it disappears tomorrow.
When you finish you will have three kinds of row and each gets handled differently: the ones that stop nothing, which you can forget; the ones that stop something replaceable, where noting the replacement is enough; and the ones that stop something you cannot rebuild. Those last ones are usually two or three, and they are the only ones that matter.
That is the whole job. The sheet fixes nothing on its own, and it is still what separates someone who can react from someone who finds out when there is nothing left to do.
To close
Most of the decisions that could stop your operation are ones you will not see coming, will not be asked about, and cannot prevent. There is no fixing that and it is not worth fighting.
What is within reach is knowing, in advance, how much of what is yours would leave with them. And that gets answered with one question, four places, and a two-column sheet.
It is one of the least exciting things you will do this month. It is also one of the few that will still be working in five years, whatever happens to the tools you use today.
For you, who made it all the way down here: thank you. We know auditing tools is the last thing anyone wants on a Friday, so if you walk away with one row written down, it was worth it.
Keep reading
- The map of the five disciplines · the full frame for where each piece fits
- Hacking your business now costs $50 · the same neglect through another door
- AI doesn't guess, it skips you · what happens outward when your details do not match
- The format AI cites most is a list
- This week's news recap
- Coverage of the reported Nvidia and Hugging Face deal
Sources
- Initial report on Nvidia's agreement to acquire Hugging Face, published by The Information on August 26, 2026 and picked up by general outlets on the 27th · this piece deliberately does not assert the state of the deal as of publication: at the time of writing there was no confirmation from either party and the reporting was still moving · https://techcrunch.com/2026/08/26/nvidia-closes-in-on-hugging-face-acquisition/
