From IC to EM: What Nobody Tells You About the First 90 Days

Three weeks into managing my first team, I closed a nasty concurrency bug that had been open for a month. I fixed it at 11 p.m., pushed the PR, and felt the old rush of dopamine. The next morning, my most senior engineer asked me a polite question. Why had I taken the most interesting problem on the board away from the person it was assigned to?

That was the moment the job actually started. Not the day the title changed. That morning, I understood something. The thing that made me good enough to get promoted was now the thing most likely to make me a bad manager.

I made this jump myself at Cogoport in 2021. I joined as a Senior Lead Engineer. Six months in, we were scaling the engineering team fast, and my VP of Engineering pushed me to take a manager role. I said yes before I really understood what I was saying yes to. That is how most of these transitions happen. And what I said yes to, it turned out, was 28 engineers. Four tech leads, a couple of seniors, and around twenty people fresh out of the IITs and NITs. Brilliant, and almost entirely green. I had not finished learning the ropes myself, and now I was responsible for twenty people learning theirs. Since then I have promoted and coached other engineers through the same jump. The pattern is always the same. Almost none of it is in the congratulations email. This post is what the first 90 days actually demand, what breaks people, and what I wish someone had told me before I shipped that PR at 11 p.m.

Your unit of output just changed and your instincts didn't

Here is the idea that everything else hangs on. As an IC, your output is the work you produce. As a manager, your output is the work your team produces. Plus the decisions you make that change what they produce. You did not get a bigger version of your old job. You got a different job that happens to share some vocabulary with the old one.

This sounds obvious on paper. It is brutally hard to live. Your instincts were trained over years of being rewarded for solving things yourself. A hard problem shows up, and your hands are already on the keyboard. That reflex was your best asset for a decade. In the manager's chair, it is a trap. Every hour you spend head-down in code is an hour you are not doing the job only you can do now. Unblocking four other people. Catching the design mistake before it ships. Having the conversation that keeps someone from quitting.

The math is stark once you look at it. Say you have five engineers, and you spend a day writing code. You produced one engineer-day of work, and you were probably slower than the person who should have done it. Now spend that same day removing a blocker that was costing each of them two hours. You just produced five engineer-hours. And they now trust you to clear the road, which pays off again and again. The leverage is not close. But leverage is abstract, and a green CI check is concrete. So new managers chase the concrete thing and wonder why the team feels adrift.

So the first discipline of the first 90 days is simple. Catch your own hands reaching for the keyboard. Then ask one question. Is this really the highest-leverage use of me right now, or am I doing it because it is comfortable and feels like progress? In the early days, nine times out of ten, it is the second one.

Write down your imposter syndrome so you can argue with it

Every good new manager feels like a fraud in month one. The ones who say they don't are usually the ones I worry about. You went from being one of the most capable people in the room to being a beginner again. The craft you were great at is gone. The new craft is fuzzy, and you cannot see your own progress. Of course it feels bad. You spent years building mastery, and you just traded it for a job where the feedback loop is measured in quarters. Half the signal is hidden from you, because people manage up.

This fear produces one specific failure: overcompensation. You feel like a fraud, so you try to prove you still have the technical chops. You jump into code reviews you should have delegated and leave nitpicks. You re-architect a design in a meeting to show you still see the whole board. You answer questions nobody asked you. Every one of these is your imposter syndrome trying to grab back the identity you understood. And every one of them undercuts the engineers you are supposed to be growing.

Here is the part nobody tells you. Your value as a manager is now almost invisible to you in real time. That is normal. It is not a sign you are failing. As an IC, you could point at a merged PR. As a manager, your best work looks like a problem that never happened. A person who did not quit. A decision that quietly saved the team a month. You have to trust a scoreboard you cannot see, and that takes far longer than 90 days. What the first 90 days ask of you is smaller. Do not let the discomfort push you back into IC work just to feel competent. Name the feeling. Expect it. Do the uncomfortable job anyway.

Weeks one through three: listen far more than you think you should

I have watched more new managers hurt their credibility by acting too fast than by acting too slow. You show up full of ideas about what should change. The urge to establish yourself by fixing things right away is huge. Resist it. You are working with almost no real data. The things you want to bulldoze exist for reasons you cannot see yet.

The first three weeks are for listening. I mean real, structured listening, not vibes. Run a proper 1:1 with every person. In that first one, you mostly shut up. I ask a few questions and then get out of the way. What is working that I should not touch? What is broken that I should know about? What do you want to do more of, and less of? And the big one: what have you told past managers that nothing ever came of? That last question is the most valuable one you will ask all quarter. The answer shows you exactly where trust was already broken before you arrived. Fix one of those long-ignored things, and you earn real credibility fast.

While you listen, do not promise. New managers want to please, so they commit to fixes on the spot, before they understand the constraints. Then reality hits, they break the commitment, and now the team has learned that their word is soft. There is a better answer: "That is really useful. Let me understand the whole picture before I commit to anything." That is a stronger reply than a promise you cannot keep. In the first 90 days, trust is built on the gap between what you say and what you do. Keep that gap at zero, even if it means saying less.

There is one thing I do change fast: anything on fire. A person about to quit. A production risk nobody owns. A broken process bleeding the team every week. Emergencies do not wait for your 90-day plan. But the ordinary stuff that just looks suboptimal? Write it down and leave it alone until you have earned the context to change it well.

The delegation you find hardest is the delegation the team needs most

The first 90 days force a painful kind of delegation on you. You have to hand the interesting, high-visibility, technically meaty work to someone else. The exact work that made the job fun. And you have to do it again and again, on purpose, when every instinct says you could do it faster and better yourself.

You probably can do it faster. That is not the point. If you hoard the interesting work, three things happen, and all of them are bad. Your senior engineers stall and start eyeing the door, because growth for them means the kind of stretch problem you keep taking. You become a bottleneck, because everything worth doing runs through you. And you never actually learn to manage, because you are still working as an IC while the management work piles up behind you.

Here is the reframe that helped me. Delegation is not offloading work you do not want. It is your main tool for growing people, and growth is now your product. When I hand a senior engineer a problem I would love to solve, I am not being generous. I am doing my actual job. Watching someone take a slower path to a solution I can already see is uncomfortable. That discomfort is the tax on building a team that does not need me in the room. I pay it on purpose.

There is a real skill to delegating without abandoning or smothering. Abandoning is "here is a hard thing, good luck," and then acting surprised when it goes sideways. Smothering is handing over the task but not the authority. You hover, you correct every decision, and the person learns they are just your hands. The middle path is to delegate the outcome and the decision, and stay available as a resource. Here is the problem and why it matters. Here are the few constraints I actually care about. The how is yours. I am here when you want to think out loud. Then you let them make calls you would have made differently. You only step in when the blast radius is too big to absorb as a learning cost. Getting that balance right is most of the job, and 90 days is barely enough to start.

You are now managing the people who used to be your peers

Most first-time EMs get promoted from within. If that is you, you inherit a landmine nobody defuses for you. Yesterday you were a peer. Today you are the boss. The engineer you used to complain about your old manager with now reports to you. The person who was a stronger IC than you, and knows it, now gets their performance review from you. This is awkward for everyone, and pretending it is not makes it worse.

The move that works is to name it directly and early. One on one with each person: our relationship is changing, here is how I want to handle it, and here is what I am asking of you. There are two ways to get this wrong. You can overcorrect into cold formality to assert authority, which feels like a betrayal to people who liked the old you. Or you can pretend nothing changed and try to stay everyone's buddy, which falls apart the first time you have to make an unpopular call or give hard feedback. Both are the same failure of nerve.

The hardest case is the strong IC who wanted your job, or who is simply better than you at the craft. The instinct is to feel threatened and assert technical dominance. That is exactly wrong. Do the opposite. Say plainly that they are better than you at the thing. Make their expertise visible. Lean on it in public. Reframe your job as making their impact bigger, not competing with it. Great managers are not the best engineer in the room, and they stop pretending to be almost immediately. Being able to say "you know this better than I do, what do you think we should do" is not weakness. It is the foundation of a team that trusts you. The sooner you find it in the 90 days, the better.

Favoritism is the other trap with former peers, and it is mostly invisible to the person doing it. You will naturally be more comfortable with the people you were already friends with. That comfort leaks into who gets the good projects, whose ideas get airtime, who gets the benefit of the doubt. The team notices instantly, long before you do. Nothing poisons a group faster. In the first 90 days, I deliberately over-invest in the people I have the least natural rapport with, to counter a bias I know I have.

Learn to run 1:1s like they are the job, because they are

New managers treat 1:1s as status meetings and run them into the ground. You do not need a meeting to learn that a ticket is in code review. That is what the board is for. The 1:1 is the highest-bandwidth tool you have for the real work of management. Spending it on status is like using a scalpel to butter toast.

A few rules I locked in early and have never regretted. It is the report's meeting, not yours. They set most of the agenda, and you resist the urge to fill the silence with your updates. You never cancel it. Canceling on someone again and again tells them, louder than any words, that they are not a priority. A protected, recurring slot is itself a sign of respect. And you use it for the things that fit nowhere else. Growth. Friction. The half-formed worry someone will not raise in a group. The feedback that needs a closed door. If all your 1:1s are tactical, you are not managing. You are running a smaller standup.

Here is the uncomfortable truth about the first 90 days. Most of what matters in a 1:1 only shows up after the person trusts you. Trust takes months to build and one broken confidence to destroy. So your early 1:1s may feel shallow and unproductive. That is fine. You are laying track. The consistency, the real listening, the small things you follow through on, all of it compounds. Eventually you get the kind of relationship where someone tells you they are burning out before they hand you a resignation letter. That relationship is the whole game. You cannot rush it. You can only start it well.

Your relationship with your own manager is now a tool, not a formality

As an IC, your skip-level was a distant figure. As a new EM, your relationship with your own manager becomes one of your most important tools. Most first-timers underuse it, out of a misplaced wish to look like they have everything handled.

Bring your manager the messy stuff early. The hard people decision you are unsure about. The priority conflict you cannot resolve at your level. The political dynamic you do not understand. This is not weakness. It is exactly what the relationship is for, and a good manager wants the hard problems while they are still small. The new EMs who hide their struggles to look competent are the ones who blow up in month five, when a problem they buried in month two has grown out of control. I would much rather coach someone through a mess they surfaced early than clean up one they hid.

Use the relationship to calibrate, too. Your own sense of how you are doing is unreliable in the first 90 days. The scoreboard is invisible and the feedback loop is slow. So ask directly and specifically. Where am I overweighting IC work? Where is my read on this person off? What am I not seeing? When you cannot see your own output clearly, an experienced outside view is not a luxury. It is instrumentation, and you would be foolish to run without it.

Your manager gives you one more thing: air cover. And you have to learn to ask for it. Your team will run into priorities set two levels above you, deadlines you never agreed to, reorg rumors you cannot confirm. Part of your job is to absorb that pressure so it does not hit your engineers as raw anxiety. And you cannot absorb what you do not understand. So push upward for the context. Why this deadline? What is the real constraint? What can actually move? Then translate it down into something your team can act on without panicking. A manager who just forwards executive pressure downhill is a router, not a shield. The team feels the difference within a week.

Ninety days in, the honest scorecard

At the 90-day mark, the numbers that tell you whether you are becoming a manager are not the ones you are used to. Do not grade yourself on how many problems you personally solved. That number should be going down, not up. Grade yourself on other things. Does your team ship more predictably than before? Can your people tell you what they are working toward and why it matters? Is the thing you were most worried about on day one now owned by someone other than you? Has anyone trusted you with something genuinely hard and personal? Those are the real signs of a functioning team, and none of them show up in a commit history.

One instinct got me through the transition, and it is the one I would hand to anyone about to make it. Whenever you feel the pull to jump in and do the work yourself, treat that pull as a signal, not a plan. It is almost always pointing at the real work. The delegation you are avoiding. The conversation you do not want to have. The trust you have not built. The decision you are dodging by staying busy. In the first 90 days, the comfortable thing and the right thing are rarely the same. Learning to tell them apart, and then choosing the right one, is the whole job.

I stopped shipping 11 p.m. PRs. It was the hardest professional habit I ever broke. Breaking it was the moment I actually became a manager. The dopamine is gone. What replaced it is slower and quieter, and in the end much bigger. You watch people you invested in do things they could not have done without you. None of it has your name on the commit. Nobody tells you about that trade in advance. It is the whole trade.