I have sat in more than one board meeting that was, functionally, lunch with a lawyer present. Four people around a table, a founder narrating the quarter, general agreement, a signature block circulated three weeks later. Everyone in the room was competent and honest. And a year on, not one of them could have told you why a particular decision was made, who had objected, or what the board had actually approved as opposed to merely heard about.
That is the real risk in a small business board of directors. It is not fraud. It is ambiguity — decisions with no author, approvals with no record, and relationships close enough that nobody wants to be the person who asks for something in writing. Ambiguity is cheap right up until the moment a bank, a buyer, a co-founder's spouse or a disgruntled minority holder asks a question nobody prepared for.
Decide what the board is for before you decide how it runs
Small boards go wrong because they never settle their own purpose. A board is not a management team with better chairs, and it is not a monthly progress review. In an early or small organisation it does three things well and everything else badly.
It hires, evaluates and pays the chief executive. It approves the handful of decisions that cannot be cheaply reversed. And it is the institutional memory of what the organisation has committed to — to investors, to funders, to employees, to itself. Everything else the board touches, it touches as advice, and advice does not need a resolution.
Write that down in a paragraph and read it aloud at the first meeting. If your directors think they are there to help sell, you will get a sales meeting. If they think they are there to protect the founder from bad news, you will get that too. Founder board management starts with the founder saying plainly what he wants the board to be accountable for — and then not resenting it when they do the job.
Document one: a one-page schedule of who decides what
The first of the two documents that prevent most disputes is a schedule of reserved matters. One page. On the left, the decisions that require board approval. On the right — and this half matters more than people expect — the decisions management makes alone and simply reports.
Build the reserved list by asking a single question of each item: if this goes wrong, can we undo it cheaply? Anything you cannot undo cheaply belongs to the board. In practice that usually means:
- Issuing equity, options, or anything convertible into either
- Borrowing, guaranteeing, or pledging assets
- Hiring, firing, or changing the pay of an officer
- Any contract longer than a term you pick, or larger than a number you pick
- Any transaction with a director, an owner, or a relative of either
- Starting or settling litigation, and changing auditors or legal counsel
The thresholds should be uncomfortable to set. If picking the number is easy, it is probably too high to matter. And the right-hand column is what makes the document survive contact with a working business: a founder who can point to a page proving that hiring, pricing and vendor selection are his to make will accept the constraints on the left without treating each one as a loss of trust.
This is the habit that transfers most directly from institutional investing to a small company. Disciplined investors do not argue about authority in the middle of a decision; they agreed on the boundary in advance. I have written before about what private equity discipline teaches an owner-operator, and this is the cheapest piece of it to adopt.
Document two: a conflicts policy that actually gets used
The second document is a conflict-of-interest policy with a live register behind it. Small organisations resist this because it feels like an accusation. It is the opposite. In a four-person board, everyone is conflicted, and that is by design — you recruited people who are close to the work, who know your customers, who might one day sell you something or hire your employee.
The mistake is treating conflicts as things to avoid rather than things to declare. A workable policy says three things: what counts as an interest, that it must be disclosed in writing when it arises rather than annually in arrears, and what happens next. What happens next is almost never expulsion. It is usually that the interested director speaks, answers questions, and then leaves the room while the others decide — and that the minutes record all three steps.
Small nonprofit governance lives or dies on this. As founder and chairman of Berkeley Florida, the Cal alumni community here in the state, I take it as given that our directors will be entangled with the people and businesses in our own network, because a community board made of strangers to the community would be useless. The discipline is not distance. It is that nothing goes unsaid. A board that declares routinely finds the difficult declaration easy when it finally arrives. A board that has never declared anything finds it impossible.
Minutes written for the person who was not there
Most small-board minutes are either a transcript or a shrug. Neither helps. The right test is simple: could a competent outsider, reading this alone in two years, understand what was decided and be satisfied the decision was taken with care?
That means recording four things per item. What was decided, in the actual words of the resolution. What the board relied on — the memo, the forecast, the counsel's advice, the fact that a director spoke to two customers. What alternatives were considered and set aside. And who recused, and for what. It does not mean recording who said what, how heated it got, or a director's speculative legal theory. Debate belongs in the room; conclusions belong on the page.
Draft them within a week, while the reasoning is still recoverable, and approve them at the next meeting rather than by silence. And keep the pre-read attached. A set of minutes that references a board memo nobody kept is a set of minutes that cannot be defended.
Cadence, quorum, and ten minutes without management
Four people can meet informally about anything, which is precisely why the formal meeting must be unmistakably formal. Set a calendar for the year at the start of the year. Circulate materials a fixed number of days ahead and hold the line on it — the single most effective governance reform available to a small board is refusing to discuss anything that arrived that morning. Note the quorum. Note who attended. If you act by written consent between meetings, ratify it at the next one rather than letting it float.
Then reserve the last ten minutes for directors alone, without management, at every meeting including the ones where there is nothing to discuss. If executive session is an event, calling one becomes an accusation. If it is routine, it becomes a valve. Where the founder is also the chair, that valve is the only structural protection the other directors have, and a founder confident in his own performance should be the one insisting on it.
Advisory boards are a different instrument — label them that way
A great deal of small-company confusion comes from an advisory board setup that was never written down. Advisers have no fiduciary duty, no vote, no authority to bind anyone, and no protection from the company's indemnity or insurance unless someone deliberately extended it. Directors have all four. People drift between the categories in everyone's mind, and the drift is discovered at the worst possible moment.
Give advisers a short letter that says what they are, what they are not, what they are paid, how long it lasts and what they may not repeat. Do not call the group a board if you can call it a council. Use advisers for expertise you cannot afford to seat, and directors for authority you cannot delegate. And before you ask anyone to take the harder of the two roles, read your own invitation from their side — I keep a standing list of questions I ask before saying yes to a board seat, and the organisations that answer them well are invariably the ones already doing the work described above.
Build the file you would want in a deposition
Everything here fits in an afternoon and a folder: a purpose paragraph, a one-page schedule of decision rights, a conflicts policy with a register, minutes worth reading, a fixed calendar, and letters that say who is a director and who is not. None of it makes a small organisation slower. All of it makes the organisation legible — to a lender, a funder, a successor, a buyer, or the version of yourselves that has stopped agreeing.
The founders who resist this usually say the same thing: we all trust each other. I believe them. That is exactly the condition under which the documents are easy to write, because nothing is at stake yet and nobody has to concede anything. The same conversation held after the disagreement costs a year. Whether the board is four people around a table or a volunteer committee for an alumni community you are trying to make stick, we get the governance we install before we need it — never the governance we improvise once we do.