Every holding company has to answer one structural question early, and most answer it by accident: what does the center decide, and what does each company decide for itself?
Get it wrong in one direction and the holding company becomes a bottleneck. Every hire, every price change, every vendor contract waits for a meeting with the founder. Get it wrong in the other direction and the holding company is just a name on a tax return, with no real reason to exist.
The rule I keep coming back to is simple to say and harder to live by. Centralize the capital. Decentralize the operating.
What the center should own
The center owns the decisions that are expensive to reverse and that affect more than one company.
Capital allocation. Where free cash goes, whether a company gets more investment, whether something new gets funded. That is the core job of a holding company, and it cannot be delegated to the companies competing for the money.
Leadership of each company. Who runs it, how they are measured, and when that needs to change. The center does not run the business, but it does decide who does.
Risk that could spill over. Legal structure, insurance, banking relationships, and anything where a mistake in one company could reach another.
Shared capabilities. In our case, that has meant engineering. Xsatori Labs builds software for the portfolio and for outside clients, and none of the individual companies could justify that capability on its own.
That list is short on purpose. Every item added to it is a decision someone closer to the customer no longer gets to make.
What the companies should own
Almost everything else.
Pricing, hiring below the top role, customer relationships, daily operations, vendor choices, scheduling, marketing, and the hundred small calls that make a business work or not. The people running a staffing operation know their clients and their field reps in a way nobody at the center ever will. The people building software know their codebase and their clients. They should be deciding.
The test I use is whether the decision requires context that only lives inside the company. If it does, the center should stay out of it, even when I have an opinion. Especially when I have an opinion.
Why founders struggle with this
Most holding companies are run by someone who started as an operator. Operators like to operate. It is what made them good, and it is what makes them a problem at the center.
The pull is constant. You see a pricing decision you would have made differently. You hear about a hire you are not sure about. You notice a process that could be tighter. Each time, stepping in feels like helping.
It usually is not. It teaches the leader of that company that their decisions are provisional, that the real authority sits somewhere else, and that waiting for approval is safer than acting. Do that enough and you end up with leaders who do not lead, and a founder who wonders why nobody takes initiative.
How to stay out without checking out
Decentralizing does not mean ignoring. It means changing what you pay attention to.
Instead of weighing in on decisions, the center sets expectations and reviews outcomes. We run the same quarterly scorecard against every company. Leaders know what they are accountable for and know they will be asked about it. Between those reviews, they run their business.
There are a few things that help:
- Write down the line. A short list of decisions that require the center, shared with every leader. Anything not on it is theirs.
- Ask questions instead of giving answers. When I disagree with something, I try to ask why it was done that way. Often the answer contains context I did not have.
- Intervene rarely and clearly. When the center does need to step in, it should be obvious that it is an exception, why it is happening, and when it ends.
- Let small mistakes stand. A leader who makes a recoverable mistake and fixes it learns more than one who is corrected before acting.
Where the line moves
The line is not fixed. A new venture still finding its footing needs more involvement than a mature business with a proven leader. A company going through a leadership change needs more attention for a period. A company that has missed its numbers for several quarters has earned closer review.
What matters is that the movement is deliberate and temporary. Involvement that was supposed to last a quarter and quietly becomes permanent is how holding companies turn into bottlenecks.
The point of the structure
A holding company exists to make each company better than it would be alone. Capital, shared capabilities, and long-term patience are real advantages. Second-guessing every operating decision is not one of them.
The best version of the center is one the companies barely notice day to day, and are very glad to have when it matters.