5 On Being a Good Soldier
This chapter is currently in draft form.
- Why engineering leaders need people who pull with them, not people they have to pull
- The mirror: what you require from your directs, your boss requires from you
- The two most common forms of ego that prevent good soldier discipline
- What good soldier discipline actually looks like in practice
- Where the line is between subordinating ego and abandoning judgment
You have probably managed someone who couldn’t fully align. They pushed back past the point where the decision was made. They agreed in the meeting and then moved slowly, or moved in a different direction, or kept relitigating a call that had already settled. You felt the drag on the organization. You felt yourself spending energy on something that should have been resolved.
You fixed it or you are fixing it, because you understand that you cannot afford people you have to pull. You need people who pull with you. That is a meaningful distinction. People you pull are a cost. People who pull with you are a multiplier. Building a team of the latter is not optional if you want the organization to move.
Now ask yourself: is that what your boss experiences from you?
That is the question this chapter is built around. Not whether you are a good manager. Not whether you have good judgment. Whether you are giving your boss what you require from your own directs. Most engineering leaders, if they are honest, are not.
A word about the term, because it will bristle. “Good soldier” has a bad name in our industry. It sounds like blind obedience, like taking the order and marching into the wall anyway. I use it deliberately, and against that grain. I do not mean the soldier who stops thinking. I mean the discipline of bringing your whole judgment to the fight before the decision and your whole commitment to it after.
5.1 The Cost of a Direct Who Won’t Align
You have experienced a specific version of this. The direct who raises their objection before the decision is made, and then raises it again after, and then again after that. The one who executes the letter of what was decided but not the spirit. The one who has made their disagreement a permanent condition rather than an input.
You have to solve that behavior. Non-alignment at the leadership level will cause your initiatives to fail. The organization cannot move forward if the people responsible for moving it are each pulling in the direction they personally believe is right. At some point you tell that person directly: I need you to align. If they cannot, you find someone who can. That is not a comfortable conversation, but it is a necessary one.
People will have different ideas and approaches. You should entertain those ideas and, where possible, fold them into your plan to make it more effective. However, once the decision is made, regardless of whether their specific objections were incorporated, they need to get behind it and execute.
What you are describing when you have that conversation is the requirement for good soldier discipline. Not blind obedience and not the absence of disagreement. It is something far more specific, and more difficult. Full, open, safe argument before the decision, then disagree and commit. That is what you need from them. It is also what your boss needs from you.
The question is whether you are giving it to them or not.
5.2 The Ego That Knows Better
For most of my career, I was not.
At one company, I was building an accounting system. The requirement was clear: reliable, auditable, able to handle the transactional volume the business needed. It had some other interesting requirements, like an accessible ledger, but Postgres would have handled it just fine. Any competent engineer reaching for the right tool would have been done in a fraction of the time.
I was not interested in Postgres. I was interested in global distribution and transactionality at a scale the problem didn’t require. So I built a database from the ground up, using blockchain-like semantics. It was incredible technology and a tremendous amount of fun, but it made no sense at all given what we actually needed.
My boss became frustrated. I kept building. I knew he was frustrated, but we were just so close to finishing the useful core, just a bit more and productivity would magically open up with all this incredible capability. The architecture was sound. The approach was technically superior. The business just couldn’t see it yet.
What my leader saw was an engineer who had been given a clear problem and chosen to solve a different one. A more interesting one. One that let me demonstrate what I was capable of. I had fused my identity with technical correctness, and once that fusion happened, their feedback stopped being information. It became something to overcome.
The outcome was predictable once I was willing to see it. I stopped getting invited to conversations. I started being ignored, not invited to off-sites and recruiting trips. Eventually, I realized that I was shut out and realized I had poured gas on my capital and lit it on fire. I asked my boss if it was time to go. He confirmed it was.
That was a soft firing. I had earned it.
A version of this pattern appears in The Economy of Political Capital, in the context of a CI/CD initiative where the CEO was clearly signaling that the approach was not landing. The signal was unmistakable. I doubled down, because I knew I was right. The cost was political capital I did not have enough of to spend. The underlying failure was the same: ego fused with technical correctness, experienced by the people around me as an inability to hear.
The psychological mechanism has been studied. When you are personally responsible for a prior decision, you commit more resources to it after negative feedback, not less (Staw, 1976). Staw called this escalating commitment. It does not feel like irrationality from the inside. It feels like conviction.
This is the most common form of ego in engineering leadership. It does not feel like ego from the inside. It feels like standards. It feels like rigor. It feels like caring about doing things well. The people around you experience it as someone who is not quite reachable. And the research suggests that the people most prone to it are disproportionately likely to rise: leaders who project the strongest conviction in early contexts are selected for that confidence, but the same traits that drive leadership emergence predict declining effectiveness at the top (Grijalva et al., 2015). The selection process favors the pattern before its costs become visible.
5.3 The Ego That Wants to Be Good
There is a second form. Harder to see, because it looks like virtue.
Later in my career I was managing a team that included a director. A genuinely nice person. Everyone liked him. He was working hard, visibly trying. He also was not effective in the role. My boss moved him into my organization with a clear directive: work him out.
I understood the directive. I agreed with the assessment. I did not act on it.
I kept looking for a way to help him grow into the role. I kept finding reasons to give him more time. What I told myself was that I was being fair. What I was actually doing was protecting my image of myself as the leader who finds another way. The leader who doesn’t discard people. The one whose team can trust will fight for them. If you are a remotely empathetic person, those conversations just suck. To this day I hate firing people no matter how well deserved it is. So I dragged my feet. I was also spending political capital I could not afford to lose, for no good reason. This person had already proven that they couldn’t perform. The decision had been made, just not by me.
That is ego too. It just wears compassion as a disguise.
The director stayed. The organization continued not getting what it needed from that role. That unresolved situation was one of several things that contributed to my not succeeding in that role.
The person you cannot bring yourself to let go is costing the organization more than you are admitting. And the reason you are not admitting it is frequently not about them. It is about what executing that decision would mean about you. The technical ego wants to be right. The relational ego wants to be good. Both are the self in the way of the mission.
5.4 But My Boss Is Wrong
Here is the objection you have been making since the first page. Your boss is different. They don’t have the domain knowledge. They don’t understand the technical reality the way you do. The call they just made is wrong, you can see exactly why, and aligning behind it means walking the organization into a wall you could have pointed out.
Maybe. It might be precisely that. It might also be that you are missing context you don’t know exists: pressure from the board, a constraint on the balance sheet, a bet that hasn’t been made public, a read on the market you don’t have. You usually cannot tell which of these is true. You almost never have the information to be as certain as you feel.
And here is the part that matters: it doesn’t change what you owe. Whether your boss is wrong, or whether you simply can’t see why they’re right, the obligation is identical. Nothing in the organization moves without alignment. A leadership team each privately convinced the direction is wrong, each quietly hedging, produces an organization that goes nowhere slowly and expensively, congratulating itself on its prudence the whole way.
It is usually better to move in the wrong direction than in no direction at all, provided the direction is correctable. A correctable wrong direction announces itself, you feel the resistance, you see the results, you correct, and the act of moving is what produces the information that lets you correct. No direction announces nothing. It only compounds. The team that commits to a flawed plan and executes it together will usually beat the team still litigating the perfect one, because the first team is learning and the second is only arguing. The directions worth fearing are the ones that don’t announce themselves: the slow, compounding kind you recognize only in hindsight. They are rarer than the ego pretends, but they are the ones to watch.
That is the rule for nearly everything. The exception is narrow, and it is real: you do not move in a direction that is illegal, unethical, or irreversibly destructive to the company. Those are not wrong directions to be corrected later. They are different in kind, and there, refusing to move may be the right thing, but it is not the free thing. Refusal has a cost, sometimes the role itself, and you go into it knowing you may pay. The Principle of Right Action deals with that cost directly. What you do not get to do is widen that line to cover the decisions you merely dislike: the temptation to file the call you’d have made differently under “this one crosses the line” is the most respectable excuse for non-alignment you will ever give yourself. Where the line actually sits is worth being precise about, and we will get there.
5.5 What a Good Soldier Actually Does
Years after the director situation, I was in a similar position. I had just hired a very senior engineer. He was in his first month. The data was not fully in yet. A reasonable person could have argued it was too early to be certain. I was one of those reasonable people.
He had two conversations with my boss that went badly. Not dramatically. He simply did not let my boss get a word in either time. My boss came to me with a clear read: this person is not a fit. He needs to go.
I had reservations. I had hired him. It was month one. I said so.
Then I made it happen.
It happened sooner than it would have if we had let it play out naturally. But it happened, and it was the right thing. My boss was willing to move more quickly than I was. I raised my concern. The decision was made. I executed it fully.
That is the part worth sitting with. I was not certain my boss was right. I am still not. I committed anyway, not because I deferred, but because I had made my case, the call had been made, and my uncertainty was not a license to slow it down. Committing when you are sure is easy. Committing when you genuinely might be right and are overruled is the discipline.
That is what a good soldier does. You bring everything you have to the problem before the decision. You raise your objections. You make your case clearly and completely. You give the decision-maker every piece of information that might change the outcome. And then, when the decision is made, you execute it. Not slowly. Not with reservations seeping into your body language or whispered to your team. Fully, to the best of your ability.
When you say to your direct “take that hill,” you need to know what is going to happen. You need them moving, not standing at the base of the hill explaining why they think it is the wrong hill. You need that because the organization cannot afford anything else. Your boss needs the same from you, for the same reason.
Good soldier discipline is not about suppressing disagreement. Suppressed disagreement is a problem in and of itself. It means decision-makers are getting incomplete information and organizations are accumulating silent resistance underneath surface compliance. What you want from your directs, and what your boss wants from you, is not quiet compliance. It is full engagement before the decision and full commitment after.
Those are different things. The first is a gift to the decision-maker. The second is a gift to the organization.
This works best when the decision-maker does their part. A leader who wants commitment after the decision owes real argument before it: an explicit call, the reasoning behind it where that helps, and no punishment for surfacing inconvenient information. That is what good leadership owes the people under it. But owing it and giving it are two different things. Your boss may not give you the hearing they owe you, and that is just life. It is not a release. The discipline you owe upward does not dissolve because the person above you failed at the discipline they owe downward, not unless they push you across the line of law, ethics, or irreversible harm. You can name what you are owed. You cannot make your own commitment conditional on getting it.
The practiced version of disagree-and-commit is not effortful. In the novice form, the discipline is still carrying the disagreement, managing it so it doesn’t bleed into your body language and your team’s ears, spending energy on containment. The mature form is not starting the tally. Once the decision is made, you stop tracking whether you were right. You still track outcomes; you simply stop tracking vindication. Not because you are suppressing the question, but because running that process is a continuous drain on the energy you need for execution. The scoreboard is the cost. That is a learned practice. It takes time.
5.6 Where the Line Is
The line is not whether you think the decision is right.
Once your objection has been heard and the decision has been made, your continued opinion is not the point. “I disagree with the strategic direction” is not grounds for withholding execution. “I think this will fail” is not grounds for dragging your feet. “I don’t respect the decision-maker” is not grounds for non-alignment. All of those are the ego finding better framing.
The line is ethics and law. You do not do things that are ethically wrong. Not gray. Wrong. You do not break the law. You do not compromise your integrity to comply with a directive. That line is real and it is not negotiable.
But it is narrow. The Principle of Right Action chapter covers the ethics question in depth. What belongs here is the simpler observation: that line should be invoked rarely. If you find yourself invoking it frequently, the more likely explanation is that your ego has colonized your ethics. You are using the language of principle to avoid the discipline of alignment.
A useful test: if the same directive came from someone you trusted and respected completely, would the ethical objection still stand? If the answer changes depending on who is giving the directive, it was not an ethical objection. It was a preference with a more respectable label.
There is a cleaner frame for handling the “but I might be right” objection. The person above you does not have more of everything. They have more breadth. You have more depth. Most decisions turn on the breadth axis: what the business actually needs solved, which tradeoffs are acceptable at the organizational level, which problems matter now. You are probably right about the thing you can see closely. You are probably wrong about the axis the decision depends on.
The blockchain story is the cleanest illustration, and the harsher version of the point above. The architecture was sound. The approach was technically superior. But “technically superior” was not a property the engineering held on its own; it only means anything relative to the problem, and I was solving one no one had. It is tempting to say I was right on the engineering and merely wrong about which problem to solve, as if those were two separate scorecards. They weren’t. Sound architecture for a problem the business didn’t have is not a right answer on the wrong axis. It’s a well-built answer to a question nobody asked, which is no answer at all. And the person who could see that was the person I was not listening to.
This is why post-decision resistance is almost always a bad bet. You may be right on the dimension you can see. The decision likely turned on the dimension you can’t. And if the pattern persists, if being right and overruled is not an incident but a steady state, that is information about fit, not a call to deeper soldiering. That question belongs in a different kind of conversation, run when you are off the field, not during execution.
5.7 Curating the Right People
The discipline you are demanding of your directs is the same discipline you owe your boss. You cannot hold them to a standard you are not practicing yourself, not credibly and not for long. The team watches whether what you require of others is what you live.
What you are looking for in your directs is a specific combination: full argument before the decision and full commitment after. Both. The people who argue hardest before a call are often the hardest to redirect once it is made. The people who execute cleanly are not always the ones who will catch your mistakes beforehand. When you find someone who can do both, invest in keeping them.
When you do not have it, fix it. Name what you need. Give the person a real chance to close the gap. If they cannot, make the change, before it becomes acute, not after.
I let it become acute once. I had a manager over an operations team who had deep experience and strong convictions about how the team should work. For a long time those convictions served us well. As the organization matured, we needed that team to shift from being a gate to being a facilitator. He believed the gate model was the only thing maintaining stability and was not open to reconsidering it. Other problems pressed harder, so I let it go.
He eventually left before it became a crisis. But by the time he left, he had shaped the team in his image. The beliefs were not just his anymore. They were the team’s. The new manager I hired faced significant, invasive work to reorient people who had been trained by years of the previous approach.
Leaders transmit culture through what they consistently pay attention to, what they reward and ignore, how they respond when things go wrong, not primarily through what they say, but through what they do repeatedly (Schein and Schein, 2017). A manager who persistently models non-alignment is not just holding a position. They are teaching it. Non-alignment rarely announces itself; it operates as public agreement paired with private subversion: slow movement, selective compliance, friction in the informal channels, and from the outside it can look like ordinary implementation drag until the direction simply never arrives (Ybema and Horvers, 2017).
When you eventually address it, you are not just replacing a person. You are rebuilding a team. Address it sooner.
Your CTO will disagree with you. That is expected and healthy. The question is whether they can execute decisions they disagreed with, fully, without reservation bleeding into how they run their organization.
The signs that alignment is breaking down are specific: initiatives agreed to in the room that move slowly or sideways in implementation; objections that re-emerge after decisions were made; mixed signals reaching the engineering team about what the actual direction is; cultural drift in the organization that doesn’t match the company’s stated priorities.
These have business consequences. Slow execution compounds. A team receiving mixed signals from its leader stops moving with confidence and starts waiting to see which direction wins. You lose quarters to this, not days. Other leaders on your team observe how you handle it and draw conclusions about whether organizational commitments mean anything.
This is an ego problem, not a judgment problem. More information will not fix it. If the CTO can see the pattern and wants to close the gap, coaching may help. If they cannot see it, you are looking at a leadership fit issue: they may be technically excellent and genuinely unable to operate within a hierarchy. Both things can be true simultaneously. The longer you leave it, the more expensive the resolution becomes.
5.8 The Mirror
The arc of this chapter is simple. You know what it costs when a direct cannot align. You will solve it when you have to, because you cannot afford otherwise.
The question is whether you are giving your boss what you require from your own people. In most cases, if you are honest, the answer is that you have not always been. Neither have I.
The work of being a good soldier is not glamorous. It does not feel like the high-leverage leadership work described in the rest of this book. It feels like subordinating yourself. That is what it is. The organization moves forward when the people in it put the mission above themselves, not once, not when it is easy, but as a discipline. The good soldier delivers that discipline to the person above them as reliably as they expect it from the people below.
There is a return on it, though it rarely arrives as thanks. The people who argue hard and then commit cleanly are the ones you stop managing and start trusting. You pull them further into the room, not out of it, because you have learned that their disagreement is information and their commitment is reliable. The directs I trusted most were never the most agreeable; they were the ones who would fight me to a standstill before a decision and then execute it as if it had been their own. That trust is the exact inverse of the exile I earned at that accounting company. It is what the discipline buys, on both sides of the relationship.
That is the standard. Most of us spend years learning it by failing to meet it. The question the chapter opened with is the one it closes with: not whether the people below you give you alignment, but whether you are giving it to the person above.