Work On the Org, Not In the Org
I mentor a lot of CTOs and CPOs. It is the advice I give more often than any other piece of advice, and it is the advice that most often goes unheeded. Work on the org, not in the org.
Your job — if you are a senior engineering or product leader — is not to pull the train. Your job is to build an engine that can pull the train. That is the organising principle, and almost every other piece of CTO or CPO advice — about setting roadmaps , running OKRs , sizing squads, or handling the board — falls out of it.
A lot of leaders get this wrong, and I have a view on why. Pulling the train is visible. It’s tangible. You can see the output. People thank you. The emails answer themselves. The problem gets solved. There is an immediate dopamine hit to being the person who rolled up their sleeves and fixed the database or unblocked the sales deal or wrote the architecture doc. Building the engine is the opposite. It is slow, indirect, low-feedback work. You spend six months hiring a VP of Engineering, and on most days you cannot point to what you did. But over years, it is the only thing that actually matters.
Almost every CTO I have watched get ejected from a senior role had the same root cause. They failed to build a world-class team of direct reports, because they were too busy doing the work themselves. The train kept moving, right up until the moment it didn’t, and then the board realised that the whole machine had been dependent on one person pulling harder and harder, and that person was now exhausted, and nothing had been built. The visible symptom is often a feature factory — lots of output, no one thinking about what the output adds up to — but the root cause sits higher up the org chart.
A’s hire A’s, B’s hire C’s
The reason building the engine is high-leverage is a cascade effect in hiring that is almost mechanical in its predictability. A’s hire A’s. B’s hire C’s. Some version of this rule has been attributed to dozens of people over the years. It doesn’t matter where it came from. It is true, and it is true to an uncomfortable degree.
An A-grade senior hire is someone who is genuinely better at their craft than the leader who hired them is at that craft. They are secure enough in their judgement to want people around them who will challenge them, because that’s how they continue to get better. They hire A-grades too.
A B-grade senior hire is someone who is quietly threatened by people who are better than they are. They will not consciously admit this — nobody admits it — but the pattern shows up in their hires. They hire C-grades. People who are competent enough to get the work done, but not good enough to expose the B-grade’s limits. Over three hiring cycles that B-grade senior quietly builds a department of C-grades, and the organisation has silently degraded underneath the leader’s feet.
This is why one mediocre senior hire is a much bigger problem than one mediocre junior hire. A junior mistake costs you a year of output. A senior mistake costs you a decade of hiring. If you take one thing from this article and ignore the rest, take this: spend a disproportionate share of your calendar on senior hiring, and do not compromise the bar. Leave the seat empty for another three months rather than fill it with a B. You will pay for the compromise for years.
The 1:1 cadence
The team is constantly evolving so how do you ensure that you’re effectively herding the cats. My answer — and this is the specific, concrete, copy-it-if-you-like part of my playbook — is a cadence of 1:1s that scales across the organisation. Note that these run alongside regular SteerCo’s and ExCo sessions. Yes it’s a lot of meetings, yes it’s your job to work on the org not in the org.
When I was running teams of 600 to 800 people, the cadence was:
- Weekly 1:1s with my direct reports. Non-negotiable. Every week, every time. If the meeting got bumped it got rescheduled, not skipped. This is the smallest number of people in the organisation who need your highest-bandwidth attention and the highest-bandwidth feedback channel both ways. Skip this cadence and you are flying blind.
- Monthly 1:1s with their direct reports. My N minus 2. The leaders of the leaders. This is the layer where execution actually lives and where the culture actually gets set, and if you only hear about this layer through your directs you are hearing a filtered version of reality. Monthly is enough to keep the signal flowing.
- Quarterly 1:1s with selected people in the next layer down. My N minus 3. You cannot do this exhaustively — in an 800-person organisation there are hundreds of people at that layer. But you can do it selectively: the high-performers to make sure they feel seen, the people who are rumoured to be wobbling, the people whose skip-level context you specifically need, and people’s whose deliverables disproportionately impact the outcomes you’re driving. A quarterly cadence gives you four of these per year per person, which is enough to keep the organisation feeling reachable from the top without eating the calendar.
The pattern is a geometric one. Weekly at N-1, monthly at N-2, quarterly at selected N-3. Each layer gets one-quarter the attention of the layer above. That’s the right calibration — it’s enough to stay in touch without flattening the organisation, and it’s little enough that the intermediate layers don’t feel routed around. Respecting that hierarchy is also load-bearing for optimal squad size — if you’re flattening the org by skipping layers, you’ll tend to also flatten the team structure, which produces squads that are too big to function.
The diary trick
Here is the operational trick that actually makes this work. Put every 1:1 in the diary as a recurring event, and shift the individual occurrences around as the date approaches.
This sounds trivial. It is load-bearing. The problem with 1:1s that happen “when there’s time” is that there is never time. The diary fills up with other things, and the 1:1 gets dropped, and then it’s been three weeks since you spoke to a direct report, and then it’s been six, and then you have genuinely lost touch with a person whose weekly attention you were depending on.
Recurring events solve this. The meeting is in the calendar. It has a slot. If the slot is inconvenient that week, you move it to a slot that isn’t. The point is not that the 1:1 happens at the same time every week — it often doesn’t. The point is that the 1:1 happens, because it exists as a scheduled commitment that has to be actively rescheduled rather than passively dropped. It is always easier to move a meeting than to remember to schedule one.
I have used this trick for thirty years. It is the single highest-leverage calendar discipline I know. It is also the one almost every senior leader I mentor does not do, and the first change I ask them to make.
Hire for energy — and remove the negativity
One more pattern worth naming explicitly. When you are building the engine, you are making hiring decisions about the character of the team, not just its competence. Two rules have served me well.
Hire for energy. I will take a slightly less technically credentialled person who is genuinely excited about the work every single time over a more credentialled person who is phoning it in. Energy compounds. It changes the weather in the team. It shows up in the thousand small decisions that nobody will ever tell you about. Credentials matter, but energy is the multiplier.
Remove negativity quickly. People who consistently resist change, who reflexively oppose new ideas, or who bring a pall of cynicism into every meeting are — in my experience — almost never rehabilitated. They will not become positive contributors because you have gone on a journey with them. They will just infect the people around them. This is hard advice because these people are often very loyal, often very knowledgeable, often long-tenured. The kindest thing you can do — to them and to everyone else — is to move them on. The damage they do to the team’s energy is larger than any institutional knowledge they carry out the door with them. I have never regretted removing a negative hire quickly. I have regretted many times waiting too long.
This is also where the Deming principle bites — in knowledge work, where I think it’s closer to 70% system, 30% individual , the individual still matters. A relentlessly negative individual in a senior role becomes part of the system, and their effect compounds. Deal with it.
The ejection pattern
Let me close by naming the failure mode this article is designed to prevent. Here is the pattern by which I have seen CTOs, CPOs, and senior engineering leaders get ejected.
The leader joins a business that genuinely needs their technical and product judgement. They dive in. They solve the visible problems. The board is delighted for the first twelve months. Things are getting fixed. The train is moving, and — the board believes — the leader is making the train move.
Year two, the problems get harder. The leader doubles down on doing, because doing is what got them praised. The direct reports they inherited are adequate but not brilliant, and the leader is too busy pulling the train to invest in upgrading them. Year three, the leader is tired. The cracks show. Deliveries slip. The leader works harder, not smarter.
Somewhere in year three or four, the board has a crisis of confidence. They look at the senior team under the CTO and realise that it is mediocre. They look at the CTO’s calendar and realise it is full of tactical work. They look at the company’s resilience to the CTO being hit by a bus, and they realise there is none. The leader has not built an engine. The leader is the engine. And an engine that cannot be replaced is a business-continuity problem, not an asset.
The ejection, when it comes, is rarely about the recent results. It is about the board’s retrospective realisation that the machine under the leader has not been built, and the leader is now in the position of being the only person who knows how everything works. The business is one exit interview away from catastrophe. The board does the exit interview first.
Every time I see this pattern play out — and I have seen it many times, in diligence, in board work, in mentoring — the counterfactual is the same. A leader who had spent 30–40% of their calendar on hiring, on 1:1s with their directs and skip-levels, on deliberately building the engine rather than pulling the train, would have ended year three with a team that could operate without them. The engine would have been built. The board would have been looking at a company, not a dependency.
The 70/20/10 discipline
For the CTOs and CPOs I mentor, I will often recommend a deliberate calendar audit against a simple target:
- 70% working on the org. Hiring, 1:1s, team structure, performance management, strategy, governance, the shape of the organisation. This is the work only you can do, because you are the only person with the authority to set organisational shape. It is also the work that the capacity of your organisation depends on over years — a well-built team is a capacity multiplier, and a poorly-built one is a capacity drain.
- 20% working in the org. Technical or product judgement calls that genuinely require your expertise. Architecture reviews, major product decisions, escalations that need your judgement, the occasional firebreak sprint where you roll up your sleeves alongside the team. This is also where you reach into Type A decisions the team is about to get wrong — see sometimes people have to grab hold of the electric fence . This 20% should be finite, deliberate, and bounded.
- 10% working on yourself. Reading, thinking, peer networks, skill development, the things that keep you from going stale. Most senior leaders drop this first. It is the thing they most regret dropping.
If your calendar audit shows 30% on the org and 65% in the org, you are a leader on the way to being ejected. You don’t know it yet, because the train is still moving. But the board will know it in year three.
The organising idea
Your job is not to pull the train. Your job is to build an engine that can pull the train. Everything downstream of that — the hiring discipline, the 1:1 cadence, the calendar audit, the relentless energy-positive-culture building — is just a way of turning that organising idea into action.
It is the single most important thing I tell the leaders I mentor, and it is the single thing most of them eventually wish they had started sooner. The train will keep moving in the short term either way. The engine only gets built on purpose.