Take a family office run by three partners. A deal arrives on a Thursday: a growing logistics business, a confident founder, a tight deadline. By Sunday the money is committed. Nobody wrote anything down, because the most senior partner liked the founder and the other two did not want to be the ones who slowed things down.
Eighteen months later, the same three people are trying to reconstruct why they said yes.
That is the gap an investment committee closes. Not a monthly meeting where partners share opinions, but a written filter every deal must pass before capital moves. Most private investors and family offices believe they have one. Very few do.
An investment committee is a filter, not a forum
The word "committee" makes people think of a room. The room is the least important part. What matters is what a deal has to survive before it reaches the room, and what the room is allowed to decide once it is there.
A working investment committee answers four questions in writing. What do we invest in, and what do we refuse? Who decides, and who only advises? What evidence does a proposal need before we will discuss it? And what happens after we say yes?
If those answers live in one person's head, you do not have a committee. You have a founder with an audience.
The five parts of an investment committee charter
The charter is a short document, three or four pages at most. It holds five things:
A mandate: the sectors, ticket sizes, stages and geographies you invest in, and the ones you will not touch regardless of how good the pitch is.
Membership: named voting members, at least one of whom has no personal relationship with the deal or the sponsor, and a secretary who records decisions.
Rules of decision: quorum, voting thresholds, and how a member with a conflict of interest steps out of the discussion, not just the vote.
The investment memo standard: the same template for every deal, covering the thesis, the downside case, the key risks, and the conditions that would make you exit.
Post-investment review: when the committee looks at the deal again, what numbers it expects to see, and who reports them.
The last part is the one most charters skip, and it is where most losses begin. We have written separately about why post-investment governance decides more outcomes than the entry price does.
A committee that never says no is worse than no committee
Here is the uncomfortable part. A committee that approves everything it sees is not neutral. It is harmful, because it adds a layer of false confidence to decisions that were made elsewhere.
Look at your last twelve months. How many proposals reached the committee, and how many did it reject or send back with conditions? If the answer is zero, the committee is not filtering anything. It is signing.
The fix is structural, not personal. Give one member the explicit job of arguing the case against each deal. Require the memo to state what would have to be true for the investment to fail. Record dissent in the minutes, with names. None of this is hostile. It is what makes a yes worth something.
How to set up an investment committee before the next deal
You do not need a new entity or a large budget. You need a sequence:
Write the mandate on one page and have every partner sign it.
Name the voting members and the independent voice. An experienced outsider for two meetings a quarter costs far less than one bad deal.
Adopt a single memo template and refuse to discuss any deal that arrives without it.
For the first meeting, review a deal you have already made. Run it through the new process and see what you would have asked.
That last step tends to be the most useful hour a group of investors spends all year. It turns the committee from a theory into a habit, using a decision everyone already has an opinion about.
This is the discipline we teach in our Investment Governance and Risk Management course, and it applies whether you manage a family office, an angel syndicate or a corporate investment arm.
When your next deal arrives on a Thursday, which document will it have to pass before Sunday?
