How to Run a Portfolio of Companies Solo
The operating system for one person running many ventures at once. What you centralize, what you template, what you refuse, and where the shared foundation pays the rent.
Running many companies alone is not a productivity trick. It is a set of decisions about what gets shared, what gets copied, and what never gets built at all. Get those wrong and one person cannot hold it. Get them right and the portfolio mostly runs on rails you laid once.
This is the operating system underneath nineteen ventures with no partners and no staff. It is honest about where it works and where it strains.
Centralize the things that must be identical
The first move is deciding what lives in one place for everyone. The rule: anything that must be identical across ventures gets centralized, because letting it vary is how a solo operation turns into nineteen part-time jobs.
So identity, the data layer, billing, deploy, and the governance guardrails all live in one shared foundation. They are not rebuilt per company. The promise that the software will not lie and will not do damage is enforced once, underneath, not reinterpreted nineteen times. I broke down how that base layer is structured in How to ship SaaS faster with a shared foundation, and centralizing it is what makes one person enough.
Template the things that repeat but vary
The second move is for everything that recurs but is not identical. A new venture needs a site, a launch, a content rhythm, a support surface. These are the same shape every time and different in the details.
Those get templated, not centralized. There is a known starting point I copy and adjust, so I am never deciding the structure from scratch and never forced into one rigid version either. The foundation handles what must not vary. Templates handle what should vary a little. The difference between those two buckets is most of the job.
Refuse more than you accept
The third move is the one people skip, and it is the most important for a solo operator. You hold the portfolio together by refusing things, not by doing them.
I refuse work that cannot sit on the shared foundation, because a one-off stack is a second business I now have to maintain. I refuse ventures that need a team to be honest, because I cannot hold a governance standard I am not personally enforcing. I refuse capability that I cannot also govern, every time, because shipping the ungoverned version is how the one asset I cannot rebuild gets spent. The refusals are the operating system as much as the foundation is. You can see what survived the filter at my work.
Where the foundation pays the rent
All of this only works because of where the leverage sits. The foundation is what makes venture number two cheap, because most of it already existed. The model is what makes the build work small enough for one person, because it does the labor a department used to. Together they are why a solo operator can run a portfolio that used to require a company.
The failure modes, stated plainly
I will not pretend this is free. The ceiling is real: only so much can be in active motion at once, and I am the single point of failure at the center of all of it. If I am wrong about the foundation, I am wrong everywhere it reaches. If I am unavailable, the portfolio waits.
Those are not bugs I am going to fix. They are the cost of the structure, and I think it is the right cost for what this is. A portfolio whose value is one consistent standard is better held by one person who refuses to dilute it than by a team that slowly would. You can see the result at Girard Media.