There’s a reason surveyors don’t take one bearing and call it a position. One point on a map tells you where something might be. Two or three, taken from different angles, tell you where it actually is. Triangulation isn’t extra rigour bolted onto navigation. It’s the only way navigation actually works.
Most boards never do this. They take one bearing, the founder’s, and treat it as the truth of the business.
That’s not a criticism of founders. Every founder curates what they bring to a board meeting, not out of dishonesty but because that’s what happens when you’re the one being assessed every quarter. You lead with what’s working. You frame what isn’t working in a way that shows you’re already on top of it. None of that is lying. It’s what any competent person does when they’re presenting their own performance to the people who oversee it.
The problem isn’t the founder doing that. The problem is a board that only ever hears it.
What the board is actually governing
If the only information reaching a board comes filtered through the founder, the board isn’t governing the business. It’s governing the founder’s account of the business. Those two things drift apart quietly, long before anyone notices, because the drift never shows up in a board pack. It shows up in a support ticket thread the founder didn’t think worth mentioning, in a supplier relationship going stale, in a senior hire who’s been trying to say something for two board cycles and has stopped bothering.
I’ve sat on boards where that drift ran for two quarters before it became unavoidable. Nothing was hidden exactly. It just never came up, because nobody had asked anyone who wasn’t the founder.
Triangulation is what closes that gap. Not once, as a project. Continuously, as a discipline.
What it actually looks like
In practice it’s unglamorous. A NED who calls a senior hire directly, without the founder on the line, just to ask how things are going. Someone who reads a handful of actual customer support exchanges rather than the NPS summary someone built a slide around. A conversation with a supplier or investor where the tone tells you more than the words.
And the data itself needs the same treatment. A retention percentage on a slide is one bearing. The cohort curve underneath it is another. A NED who accepts the summary number with the same trust they’d give a founder’s verbal account hasn’t triangulated anything, they’ve just swapped one single source for another that happens to look more objective. Ask for the underlying data, not the digest of it. The pipeline export, not the forecast slide. The cohort retention, not the blended number. The aged debtor report, not “cash is fine.”
None of this replaces the founder’s account. It sits alongside it. Where the two agree, the board has real confidence. Where they don’t, the board has found the thing worth a proper conversation, before it becomes the thing that sinks a quarter.
Do it in the open, or it’s just backchannelling
There’s a version of everything above that’s genuinely useful, and a version that’s corrosive, and the difference isn’t the behaviour. It’s whether the founder knows it’s happening.
A NED who quietly calls the ops lead without ever mentioning that’s how they work is running a backchannel. A NED who says, plainly, at the start of the relationship, “this is how I work, I’ll be talking to your team and your customers directly, that’s not a mark against you, it’s how I do the job properly,” has turned the exact same behaviour into partnership. Same phone call. Completely different relationship.
This also tells you something about the founder. One who makes the introductions, encourages their team to speak freely, and treats a NED’s questions as useful rather than intrusive is showing you real security. One who quietly manages access, decides who the NED gets to speak to and when, is showing you something too. Neither is subtle once you know to look for it.
It works better as a board, not as an individual
One NED triangulating alone is weaker than NEDs comparing notes with each other before the meeting even happens. The most useful signal in governance often isn’t what one person picked up. It’s that two board members, working separately, picked up the same thing without being prompted to look for it. That kind of convergence carries more weight than anything either of them could raise alone, and it only shows up if board members are talking to each other between meetings, not just performing agreement in the room.
Why most NEDs don’t actually do this
Here’s the uncomfortable part, and it’s aimed at board members more than founders.
Triangulation takes real time. Not board-meeting time, the ninety minutes everyone can point to on a calendar, but the time in between: a call with a customer, a coffee with someone on the leadership team, a message to another NED to compare what each of you is separately noticing. None of that shows up on an agenda. None of it is visible to anyone but the person doing it.
Most NEDs aren’t given that time. And most aren’t paid in a way that makes spending it feel proportionate to what’s being asked of them. So the role quietly shrinks to fit what’s actually being compensated: read the pack, show up, ask a sharp question or two, leave. That’s not a NED doing the job badly. That’s a structure making the real job almost impossible to do at all.
What the time actually looks like
Most NED agreements list the visible half of the job: four to six board meetings a year, half a day’s prep for each. That’s real, but it’s probably a third of what proper engagement actually takes. The rest is the unlisted part this piece has been describing.
A NED genuinely doing this well at a fast-scaling company is closer to fifteen to twenty-five days a year, more if they’re chairing. That’s not a number to be alarmed by. It’s a number worth writing down, because right now almost nobody does.
If you’re a founder setting up a NED role, or someone placing NEDs into portfolio companies, this is the practical move: put an expected engagement day count into the agreement itself, not just meeting cadence and fee. Not because anyone’s going to log hours against it, but because naming the number is what turns “I’ll be available if you need me” into an actual expectation both sides can point back to. Vague commitments get quietly renegotiated down to whatever’s easiest. Numbers don’t.
The other failure, and it’s a worse one
Time and money explain a lot of underperforming NEDs. They don’t explain all of them.
Some NEDs have the time, and choose not to spend it this way, because they never took the seat to do this kind of work in the first place. They took it for the title. For the network. For a line on a bio that says non-executive director somewhere impressive-sounding. These are, in my experience, usually the people who talk most in the board meeting itself and know least about the business in between meetings, because the meeting is the only part of the role they ever actually signed up for.
If you’re hiring or placing a NED, there’s a better question than “what boards have you sat on.” Ask them to tell you about a time they found out something a founder hadn’t told them, and what they did with it. Anyone who’s actually done this job has an answer ready. Anyone who hasn’t will reach for something about “adding strategic value” instead, and that gap is the whole diagnosis.
What good actually looks like
A board doing this well doesn’t feel adversarial. It feels like a founder who trusts their board more, not less, because they know something real is happening in the background, not because they’re being managed but because someone is genuinely trying to see the whole picture rather than the version presented to them once a quarter.
For founders: the discomfort of a NED who stays close to the business is the price of having someone who’ll actually catch what you can’t. That’s not a cost of good governance. It’s most of what good governance is.
For board members: if the only place you’re engaging with the business is the boardroom itself, you’re not triangulating anything. You’re taking one bearing and calling it a position. It might be right. But you won’t know that, and neither will anyone else, until it isn’t.