The first quarter has one deliverable, and it is an accurate model of the company. Everything else in this lesson is a technique for building that model faster than it builds itself.
Larson’s list is organised as questions, because the questions are the work.
How does the business work? Where does the money come from, where does it go, how much is in the bank. What needs to be true in the next year for the business to make a step change in value. What the knowledge gaps are between engineering and the people running the business. If the domain is new to you, banking or real estate or insurance, dig into the domain itself.
What defines the culture? Not the values on the wall. What the true values are, what key decisions the company made recently, how decisions really get made, who is valued and why, whose role has grown and whose has stagnated.
How do you build healthy relationships with peers and stakeholders? What each peer needs from engineering and how you can help them succeed. How they define success. And the timing question that makes this urgent: how do you build the relationship before you hit your first source of conflict.
Is the team executing effectively on the right work? How an idea turns into finished work. How work gets assigned. Who steps in for emergencies.
Is technical quality high? Lesson 4 covered the practices for this one, so use them rather than inventing new ones: ask where the tools steal time, learn how technical decisions get made, ship a trivial change, attend incident reviews. Add one question specific to this priority, which is which projects the team considers impossible due to technical limitations.
Is it a high-morale, inclusive team? Who is succeeding, who is not, and why. What the active inclusion efforts are, who leads them, and whether that work is valued. What is energising the team and what is stealing energy.
Is the pace sustainable for the long haul? What you personally need to stay engaged. And which lines you are temporarily crossing that you will explicitly walk back once you have ramped up.
Six of those are about the company. The seventh is about you, and it is on the same list on purpose: an executive who burns out in month eight has failed at the job, not at self-care.
For each priority, identify a couple of measurable goals. Larson notes that the act of identifying specific goals is itself part of the learning process, because you cannot name a measure for something you do not yet understand. Execution can use the delivery metrics from lesson 11. Team health can use skip-levels and sentiment surveys. Pace is the awkward one: pick the conditions you personally need and measure how often you are meeting them. One executive he worked with needed a single thirty-minute working block each day to stay engaged. Yours will be something else.
Learning is not passive, and most of it happens through a small number of deliberate moves.
Ask your manager to write down their explicit expectations. Usually the CEO. The question is “what will success look like for my role?” A broad answer like “go figure out how to be useful” is genuinely fine. An unstated specific expectation is what ends people.
Check whether something is really wrong. Companies are surprisingly resilient and endure wounds that look unbearable. But you may have joined one with a dire problem, and Larson notes the jarring case where people are so used to it they have stopped seeing it.
Go on a listening tour. Under thirty people, meet everyone individually across the ninety days. Larger, mix team and individual meetings. Then get out of your own organization, using lunches and other informal settings, to meet people you will not work with directly.
Set up recurring 1:1s and skip-levels. Budget skip-levels by time per week, not by frequency, so the commitment survives growth. Include peers and stakeholders, not only your own team. And get to know people as people rather than as functional gears. Some executives suggest waiting until after the first ninety days to set recurring meetings, so you learn broadly before you learn deeply.
Share what you are observing. A weekly email to your organization works well. The purpose is showing both that you are listening and that you are not jumping to judgment, which are two separate messages and both need sending.
Attend routine forums. Observe how they work, including forums outside your organization. Know that a new executive’s presence generally breaks the meetings you are expected to start leading. Ask people to keep running them as before for the first several iterations anyway, and read the result knowing your presence changed it.
Shadow support tickets. Larson’s phrase is that customers are always a reality reservoir, because companies often have a distorted sense of how things are going. Shadow the customer success team, or get a login and dig yourself.
Shadow customer meetings, partner meetings, or user testing. The priority is learning how the company interfaces with the outside people essential to its success.
Find the business analytics and learn to query them. Data is a subtle liar, so do not assume your queries are right. But being able to pull your own numbers for an initial look is worth the setup.
The further into your career you get, the fewer people at your company have ever held your role. Lesson 2 said the CTO is not the top of a ladder; one consequence is there is nobody above you on it to ask.
Find a small leadership community: a private Slack, a weekly breakfast, anything with peers in similar roles. Larson’s argument for it is that decisions in senior roles are slow to show results, which makes borrowed experience unusually valuable. An executive coach is the compressed version of the same thing, with the added property that the company pays them to tell you things you would rather not hear.
Two adjustments to everything above.
Scale it to the company. Ninety days is a default, not a law. A 20-person company should not take three months to understand. A 2,000-person one likely will, and finishing early there means you stopped asking, not that you finished.
And if the learning uncovers something deeply unwell, stop going down the list and fix it. The goal is downgrading it from dire to distressed, not resolving it completely. Larson attaches one warning to that permission, and it is the sharper half: do not spend all your time on the fire. Fixing emergencies is legible, urgent and satisfying in a way that strategic work never is, and a quarter spent entirely inside one has quietly traded the long term for the short.
Not a transformation. A model of the company you can act on, two measurable goals per priority, a set of relationships that predate your first conflict, and at most one or two changes underway.
That last number is the subject of the next lesson, which is about why the instinct to arrive with answers is the most expensive instinct a new executive has.
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 joins a 2,200-person company and finishes her learning plan in five weeks, ahead of schedule. What is the most likely reading?
Larson is explicit that the complexity of a 20-person company should not take three months, while a 2,000-person company likely will. Finishing early at that scale means the questions stopped, not that the answers arrived. There is no version of a 2,200-person business, culture, peer set, execution model, technical estate and inclusion picture that resolves in five weeks.
2Which of Larson's seven learning priorities is about you rather than the company?
He asks what you need to stay engaged and energised, and which lines you will temporarily cross and explicitly walk back once you have ramped up. Putting this on the same list as business model and technical quality is the point: an executive who burns out in month eight has failed at the job, not at self-care.
3Why does Larson suggest budgeting skip-levels by hours per week rather than by frequency?
Committing to see everyone quarterly is a promise whose cost rises with every hire, so it either breaks or eats the calendar. Committing three hours a week is a promise that survives growth; what changes is the coverage, which is the honest thing to let change.
4A new executive starts attending the weekly staff meeting and asks everyone to run it exactly as before. What does Larson say actually happens?
The request is still worth making, and it buys you several iterations closer to the real thing than taking it over would. But the observation is not free. Knowing you are looking at a changed meeting is the difference between reading the room correctly and mistaking your own effect for the baseline.
5In week three you find something genuinely dire. What does Larson permit, and what does he still forbid?
His framing is downgrading the situation from dire to distressed, not resolving it. The trap he names is the joy of fixing emergencies, which is legible, urgent and rewarding in a way strategic work never is. Executive leadership requires succeeding in the short and long term, and a quarter spent entirely in the fire has silently traded one for the other.