Every company reports numbers. The problem with numbers is that they tell you what happened and not whether the business is getting healthier, and those diverge more often than founders like to admit.
We run five questions against every company in the portfolio, the same way, every quarter. They are deliberately not the standard dashboard.
1. Is this business better off inside the portfolio than outside it?
The first question is whether we should own this at all.
Not "is it profitable" — would this company be growing faster, operating better, or making more money as an independent business with its own investors and its own team? If the honest answer is that it would do better on its own, we are holding it for our benefit rather than its own, and that is not a defensible reason to keep something.
Most quarters the answer is clearly yes, usually because of shared engineering capability none of them could afford alone. The quarter it turns is the quarter we should be talking about letting it go.
2. What would break if the person running it left tomorrow?
Every business has key-person risk. The question is whether we know where it is.
If the answer is "the client relationships," that is a documented and transferable problem. If the answer is "nobody else knows how the pricing works," that is an undocumented liability sitting inside one person's head, and it is the kind of thing that becomes a crisis with no warning.
I am not looking to eliminate key-person risk — in a business of any size that is not achievable. I am looking to know precisely what it is, so we are choosing to carry it rather than discovering it.
3. Is the unit economic improving, and do we know why?
One number per business that actually indicates health. Margin per placement. Contribution per customer. Whatever the real unit is for that company.
Both halves required. Improving without knowing why is luck, and luck reverses without warning. Knowing why without improving means we have a good theory and no results.
The failure mode this catches is growth that makes the business worse — revenue climbing while the unit economic quietly degrades. That looks like success on the top line for a surprisingly long time.
4. What did we learn that changes what we do next quarter?
If a company operated for ninety days and learned nothing that alters its behavior, either nobody was paying attention or it is running on autopilot.
Autopilot is acceptable for a stable, mature business, and I want that said out loud rather than assumed. For anything still finding its shape, a quarter with no learning is a quarter wasted, and the cost is invisible unless someone asks.
5. Would I be happy owning this in ten years?
The same filter we apply before starting anything, asked again after it exists.
Businesses drift. A company that was clean can end up in an adversarial relationship with its customers, or dependent on something you would not want to explain, without any single decision having been obviously wrong. Asking annually catches the drift while it is still correctable.
What we deliberately leave off
Revenue growth as a headline. It is an output, and it can be bought at the cost of margin, quality, or the customer relationship. We look at it, but not first.
Comparisons between companies. A staffing business and a software studio have completely different shapes. Ranking them produces bad decisions — usually starving the slower-growing business that happens to be funding everything else.
Anything a competitor is doing. Occasionally useful, mostly a distraction, and almost always a way of importing someone else's strategy without their constraints.
Why the same questions every time
The consistency is the point.
Different questions each quarter produce answers you cannot compare. The same five, asked identically, produce a trend — and the trend is the actual signal. A company that has answered question two the same worrying way for four quarters running has told you something no single quarter's answer would.
It also removes the temptation to ask easier questions of companies that are struggling, which is a very human thing to do and exactly backwards.
Five questions, four times a year, five companies. It takes an afternoon a quarter and it is the highest-leverage afternoon I spend.