Failure Was the Safe Ground. A Field Note.
- Tom Kooy
- May 15
- 6 min read
Updated: May 21
A field note on the question that breaks organisations.
About this field note.
Many of the pieces I have published have been atmospheric essays.
The Hinge, set in a restaurant. The Weight, set in a gym. The First Room, which covers the neuroscience underneath all of it. Essays about what happens to people in the eleven seconds before words are spoken. The read at the door. The Welcome, they do not yet know they are receiving at Besk, in West Leederville...
Those pieces observe.
This one applies.
A field note is a different form. It carries the same lens, what the room communicates before anyone speaks, what the nervous system registers before the mind catches up, but it moves from observation toward something you can do with the idea on Monday or Tuesday.
It will have sections. It will name specific moves. If that feels less like an essay and more like a memo, that is deliberate.
The thinking underneath it is the same. The register is different.
Enjoy....
Here is what I know after two decades in a plethora of leadership roles.
The talent is rarely the problem. Neither is the product. Market timing...occasionally. But far less often than the post-mortems suggest.
What kills companies, and I have watched this from the inside at seed, at Series B, at the point where the cap table is full of serious people and the stakes are real, is something nobody in the pitch deck ever measures.
It kills major corporates too. Sporting clubs. Committees of any kind (if you can be bothered having committees). The building changes. The rot is the same.
What dies first is not the strategy. It's the willingness to say the true thing. The board presentation where the bad number gets reframed instead of named. The committee meeting where three people share the same concern in the car park and none of them said it at the table. None of those moments are lies. But they set a precedent. People stop bringing the true thing because they have learned what happens when they do. And the organisation starts running on its own mythology instead of its actual situation. The higher the stakes, the more mythology it needs...and before anyone speaks in any room, the nervous system runs a scan.
Is it safe here? Do I belong?
There are two rooms in every organisation. The one you've built. And the one every person carries with them before they walk through the door. This field note is about what happens when neither one is safe.
Amy Edmondson at Harvard has spent twenty-five years studying this. Her most counterintuitive finding: the best-performing teams didn't report fewer errors. They reported more. Not because they were worse. Because it was safe to say what was actually happening. Safety creates space for honest conversation, which enables learning, and learning is the only thing that actually compounds...
Google ran the same study on a hundred and eighty teams. The number one predictor of team effectiveness wasn't talent density or the compensation structure, it wasn't even the brilliance of the founding vision.
It was whether people felt safe enough to speak...
The CTO who won't tell the CEO that the architecture decision made eighteen months ago is now a debt ceiling the product can't grow past. The early hire who watched three people raise concerns and get quietly managed out, and now sits in every planning meeting running the calculation: is the cost of honesty higher than the cost of silence? The co-founder who has started going quiet in investor meetings. Not because she's lost conviction. Because she's lost the room.
The physiological scan is automatic and pre-conscious. What Jonathan Haidt maps is the layer above it: the moment identity fuses with belief and honest feedback stops landing as information. Which means the founder's conviction, the thing that got capital, got talent, got press, is not a belief held lightly in the mind. It's fused with identity. Fused at the level of who they are. Which means when someone challenges the thesis, it doesn't land as information. It lands as attack.
Cathie Wood built ARK Invest on something extraordinary. A 150% return in 2020. A thesis nobody else was running at scale. Then came the 80% drawdown through 2021 and 2022. And the question that surfaced wasn't really about the portfolio. It was about the room. In a conviction-based culture built around one person's singular worldview, that's the structural question the architecture creates: how safe does it make it to say we might be wrong? That's not a criticism of Wood. It's the same question every founder faces the first time the market stops agreeing with them.
The room either holds that conversation or it doesn't, but most don't.
A few years ago, I invested in a data intelligence company at angel stage. The company iterated three times to find product market fit. The problems were endless in the way that real building is always endless: not dramatic, just relentless, one thing after another that needed to be solved or called or rebuilt from the ground up.
What I watched in that company was something I hadn't seen named before. The psychological safety didn't live in the wins. It lived in the failures. The three of us found, over time, that failure was where the safe ground actually fitted. Where the honest conversation could happen without the weight of ego or capital or reputation collapsing on top of it.
That's what enabled three iterations instead of one. It wasn't talent or market timing, but it was the willingness to sit in a failed version of the thing and say plainly: this isn't it and then go again.
We didn't always see eye-to-eye, but what that friction enabled was something that grew larger than any of us individually could have built. The data intelligence play found its market and the thing succeeded.
Not every room which holds the hard conversation stays intact. But the conversation still had to happen and someone had to make it safe enough to have it.
Here's what that means practically. The question leaders need to ask isn't: are we performing? It's, is it safe to stop performing and say what's actually true?
Those are not the same room.
Building the second one is the work.
Implement Now
1. Audit the room before you fix the process.
If performance is flat, attrition is rising, or your best people have gone quiet, resist the instinct to add more structure, more scripts or more incentives. The break is probably upstream. Ask instead: when did someone last say something actually difficult in this room? and what happened to them when they did?
2. Make failure the safe ground, not the shameful one.
For startups and scale-ups especially: build the explicit expectation that calling a failed version of the thing is the job, not a threat to it. The companies that iterate to product market fit aren't smarter. They are safer. Run a monthly ritual where the question isn't what is working, but what are we not saying yet?
3. Watch the leader's identity.
The skeleton key to most leader failure isn't strategic. It's personal. Fuse the idea to the person and the room starts emptying. Slowly and politely. The way people leave a party that stopped being fun an hour ago but nobody wants to say it...
When a thesis is integral to how someone sees themselves, not just what they believe but who they are, disagreement stops being information and starts being threat. The market saying "no" doesn't land as data. It lands as an attack on the self. The response is defensiveness, then dismissal, then the slow building of an 'us-and-them' inside the company where loyalty to the vision and loyalty to the founder become the same thing.
This is where boards go silent, where co-founders start self-editing, and where the best early hires, the ones honest enough to see what's actually happening, start calculating whether the cost of saying it is worth it...
The practical move is very unglamorous. Find one person whose entire role is to say the thing the room won't. A board member, an investor, a peer who has nothing to gain from your certainty. Build that relationship in trust before you need it in a crisis. And when they speak, notice your first instinct. If it is to defend rather than to hear, the identity has fused. That's not a character flaw, it is a warning signal and it's worth more than any post-mortem you will ever run.
4. The quiet ones are data.
In any transformation, the people who stop speaking in meetings are not disengaged. They have answered the question. They don't feel safe. Get them in a room individually, not in a survey, not in a pulse check. Ask them directly what it would take for this to be a place where they would say the hard thing out loud.
5. Stop measuring performance. Start measuring permission.
The metric most organisations never track is the simplest one: do people here feel permission to be honest? You can't survey your way to that answer. You find it in the texture of your meetings, your post-mortems, your hallway conversations. If every post-mortem ends with lessons learned and nobody visibly changed their mind, the room isn't safe. It's performing safety.
I wrote an essay on this in early-May called, The First Room. It covers the neuroscience, the early failure patterns in our lives and how they carry on to wider organisational issues, and what it takes to actually build an environment where the question gets answered. If you want the full picture, read it. A warning though, it reads like you are drinking from a fire hydrant...
--TK
Comments