Every founder conversation eventually arrives at the same question: what's the exit?
It has become the assumed goal. You build a thing in order to sell the thing. The building is a means and the sale is the point, and if you say you intend to keep something, people wait for you to explain the real plan.
I've sold businesses. I don't regret it. But somewhere along the way selling stopped being one option and became the only respectable one, and I think that costs founders more than they realize.
What selling actually is
A sale converts a stream of future cash flows into a single number today, at a discount, in exchange for certainty.
That is genuinely valuable. Certainty is worth paying for. If the business depends entirely on you, if you are exhausted, if the market is turning, if the number offered is more than the business will realistically produce — take it. Those are real situations and I have been in some of them.
But notice what you are giving up. A business that throws off cash every year, that you understand, that you have already paid the tuition to learn — you are trading all of its future years for a multiple of a few of them. Sometimes that trade is good. It is not automatically good, and the culture around building treats it as though it is.
What holding gets you
Compounding. A business you keep for fifteen years produces fifteen years of cash and fifteen years of learning. Every sale resets the second part to zero. You start over as a beginner in something new, having handed the expertise you paid for to someone else.
Optionality that has to be earned once. A profitable business you own is a permanent option to do the next thing without asking anyone's permission. That is worth more than most founders account for, because it is invisible until you need it.
Better decisions. A company being groomed for sale gets managed for the sale — near-term metrics, clean books, nothing that looks risky to a buyer in eighteen months. A company you intend to keep gets managed for the decade, and those are genuinely different decisions. One of them builds something durable.
What holding costs
I want to be honest, because the hold-forever position gets romanticized too.
Holding means you never get the number. Your capital stays illiquid, sometimes for a very long time. You stay responsible — for the people, the customers, the problems — with no date on which that ends. A sale is also permission to stop, and there are seasons when stopping is the right and healthy thing.
Holding also means concentration risk. Everything you own is exposed to the same operator, the same region, often the same economy. Selling diversifies. That is not nothing.
And the hardest part: a business you hold has to actually be good. You cannot dress it up for eighteen months and hand it to someone else. Every shortcut stays yours.
How this shapes T47
We build, operate, and back companies across technology, staffing, and consumer products, and the default is to keep them. That default changes how we work.
We are willing to let something take longer if the longer version is more durable. We build shared capability — design, engineering, operating systems — across the portfolio, because we expect to be using it in ten years, and that investment makes no sense if you plan to sell each piece separately.
We also say no more often. A business that looks good on paper for three years and becomes someone else's problem in year four is a fine thing to flip and a bad thing to own.
The question worth asking
Not "what's my exit?" but: would I be happy owning this in ten years?
If yes, you probably do not need to sell it, and you have a lot more freedom than the standard script suggests. If no, that is worth understanding now — because it usually means the business has a problem you would rather find early than hand to a buyer with a representation and warranty attached.
I have five companies in the T47 portfolio. I intend to own all of them for a long time. That is not a strategy I would prescribe to everyone. But I would like the question asked more often than it is.