Lesson 5 ended on a number: at most one or two changes at the end of a quarter. It is the least intuitive instruction in this course, so this lesson is the argument for it.
Larson’s framing turns on one distinction.
His statement: durable improvements depend on building systems that create changes, not performing tactical actions that create the ephemeral appearance of improvement.
The CTO in the scenario removed three gates. That is a tactical action. The system is whatever converts an incident into a new approval gate, and it ran undisturbed for eleven months.
This also explains the number. System changes are slow to show their effects, for good or ill, and each one carries adoption overhead. An organization can only tolerate a small number concurrently. The limit is the organization’s capacity to absorb them, not your ambition or your energy.
The most useful sentence in the chapter removes the obvious objection, which is that the change was correct.
If you skip understanding, you can emulate success in the short term, but you will be back working on the same problems a few months later, with less trust and less feedback from the team.
Read the second half twice. The cost is not a wrong decision. It is a correct decision that treated a symptom, plus a team that has now learned you decide before you listen. That lesson is what shuts off the flow of information you needed to find the real cause, which is why the second attempt is harder than the first.
Larson names three onboarding traps, and they share a structure. Each substitutes something you already have for understanding you have not built.
Rushing to make changes before understanding the problem’s shape. This is the first and the biggest. You substitute momentum.
Judging without context. His example of the sound it makes: “ah, this technology is terrible, what sort of fools made this decision?” You substitute your judgment.
Reminiscing about past employers. The refrain “at my last job, we…” You substitute another company’s answers to another company’s questions.
| Trap | What it substitutes for understanding |
|---|---|
| Rushing to change before understanding the shape | Momentum |
| Judging without context | Your own judgment |
| Reminiscing about past employers | Another company’s answers to another company’s questions |
| Too many lightly-held opinions | Volume |
The fourth path is the one lesson 4 introduced: having too many lightly-held opinions. Its asymmetry is what makes it an executive problem specifically. It is low impact for an early-career engineer, because people take those opinions with a grain of salt. Nobody discounts yours. A lightly-held opinion from an executive arrives as an instruction, and the result is thrash.
That is Grove’s managerial meddling from lesson 3, amplified by a title. Meddling shrinks a subordinate’s view of their job. Opinion volume does it to a whole organization at once.
There is a single sentence that fixes the second trap, and it is worth carrying permanently.
Most bad decisions today were great decisions within a context that no longer exists.
So the useful act is not a verdict, it is a record. Write down what those contexts were, so you and every future hire can understand the evolution. Larson gives this as a technology practice, and it generalises to org structure, process and vendor choices without modification.
The practical argument for it is not politeness. The context behind a decision tells you what this company optimises for, what constraints it was under, and how it makes decisions at all. Those are three of the seven things lesson 5 sent you to learn. A verdict ends the conversation that would have handed them to you.
Larson attaches a related warning about complex problems generally.
Early in a career, most problems are hard because they are new to you, and someone nearby has seen them before. He cites a company that gave new hires the 20-40 rule: always spend at least twenty minutes on a problem before asking for help, and never more than forty before asking. The numbers are not the point. Explicit permission to ask is.
The strategy of working harder holds only while someone else manages the flow of your incoming work. A manager or senior peer used to do that for you. At executive level nobody does, and you meet the unfiltered demands of the business directly. Past that point, working harder is self-defeating: you become too busy to teach, then too busy to learn, and you fall further behind the harder you work.
His prescription for a rapidly changing, complicated situation is to slow down. There is no door in that direction, and the path is gilded with false progress, appearing to work but never arriving. You learn through reflection.
Restraint here is not caution and it is not passivity. It is a claim about where the leverage sits.
Lesson 3 said to rank work by how many people it moves and for how long. A tactical fix moves the people on one project once. Changing the thing that generates the problem moves everyone who will ever hit it, indefinitely, without you in the room. You cannot find that thing in week four, because finding it is exactly what the first ninety days were for.
The next track is where the understanding turns into something the organization can act on, starting with what a strategy is and how to tell one from a list of ambitions.
Source: Will Larson, The Engineering Executive's Primer, Ch. 2 'Your First 90 Days'
Answer to reveal the explanation. Nothing is scored.
1A new CTO personally unblocks four stalled projects in month two. Every one of them ships. Why does Larson call this a trap rather than a success?
Durable improvement depends on building systems that create changes. Four unblocked projects is four unblocked projects. Whatever caused them to stall is untouched, so the same executive gets to unblock the next four, and the organization has learned that stalling is resolved by escalation to him.
2Which of these is NOT one of Larson's three onboarding traps?
The 1:1 timing is an open question Larson explicitly leaves room to experiment with. The three traps are different in kind: each one substitutes something you already have, momentum, judgment, or a previous company's answers, for the understanding you have not built yet.
3Why does skipping the understanding phase fail even when the change itself turns out to be correct?
This is the most useful sentence in the chapter, because it removes the defence of being right. The cost is not a wrong change. It is a correct change that fixed a symptom, plus a team that has learned you decide before you listen, which is what shuts off the feedback you needed to find the actual cause.
4An engineer describes a queue choice you consider clearly wrong. What does Larson's reframe suggest you do?
The reframe is not politeness. The context that produced the decision is live information about how this company makes decisions, what it optimised for, and what constraints it was under, all of which you need. A verdict ends the conversation that would have told you those things, and it teaches everyone what happens when they explain their reasoning to you.
5You are three weeks into an ambiguous, complicated problem and working longer hours on it each week. What does Larson prescribe?
His line is that there is no door in that direction. Working harder is a strategy that only holds while someone else manages the flow of your incoming work, and at executive level nobody does. Past that point, extra hours buy the appearance of progress and cost you the reflection that would actually solve it.