Skip to main content

Benjamin
Charity

Published: October 18, 2026

Inheriting a Team That Doesn't Trust You Yet

Reading time: 23min

Your first all-hands with an inherited engineering team goes fine. Everyone nods. Nobody asks a hard question. Afterward, three people message the same senior engineer to ask what they should actually do.

That meeting is the whole problem in miniature. The team wasn't being rude, and they weren't being lazy. They've learned that leadership is something that happens to them, that promises from the front of the room don't survive the quarter, and that the safest response to a new plan is to agree with it and wait. A team that's been burned doesn't argue with a new leader. It waits you out.

A rough circle of mismatched office chairs in an otherwise empty, slightly worn meeting room, one chair turned away from the others

I wrote about the systems half of this in Inheriting a Mess: The Technical Turnaround Strategy. That piece covers the code, the deploy pipeline, and the sequence of technical fixes that reduces uncertainty. This is the other half, and it's the half that decides whether the technical fixes hold. You may be walking into a team that has been through a leader or two before you. The system may be fixable in months. The trust can take the rest of the year.

This one is written for two readers at once: the leader about to walk into an existing team, and the CEO, COO, or board member deciding who to hand that team to. They need the same information. The leader needs to know the order of operations. The person hiring them needs to know what they're hiring for, because fixing a system and regaining a team are different jobs, and interviews almost always test for the first.

What a burned team looks like in the first week

Burned teams are easy to misread, because the symptoms look like performance problems. They aren't. They're adaptations, and every one of them was a reasonable response to the environment the team was living in.

  • Performative agreement. Meetings end in consensus and nothing changes. The real decision gets made later, in a direct message thread you're not in, because the last few times someone disagreed out loud it cost them something.
  • Overcommunication as armor. Every message copies three extra people. Every estimate comes with a paragraph of caveats. The team has been blamed before, and they've learned to keep receipts.
  • Low initiative. Nobody proposes anything. People wait to be assigned work, then do exactly the work assigned, because proposing something in this environment meant owning its failure alone.
  • Learned helplessness. "We tried that." You'll hear it a lot, and it's usually true. The team watched the last two improvement efforts die, and they've stopped spending energy on the third.
  • Departures already in motion. The strongest people have been quietly interviewing. Some of them stayed only to see who you'd turn out to be.

That list isn't a diagnosis of the team. It's a record of what happened to them. Every one of those behaviors is what the team learned the last time they did the opposite. Performative agreement means someone got punished for disagreeing. Overcommunication means someone got blamed. Low initiative means the last person who proposed something was blamed when it failed. What you've inherited is a team's memory of its last leader, and your first job is to be visibly different from that memory.

If the org you've inherited is forty people with four managers rather than eight people and a lead, the same record exists, but you collect it differently. Your one-on-one list becomes your managers plus skip-levels, the burn is as likely to sit at the manager layer as below it, and the person everyone defers to may be one of your managers. Everything below still applies. It just takes a second layer of conversations to reach.

Why "here's my vision" is the wrong opening move

The instinct on day one is to set direction. You were hired for a reason, you've got a point of view, and the team seems to be waiting for one.

They are waiting, but not for that. A vision from a stranger is a promise, and this team's recent experience with promises is that they get made in January and nobody mentions them again by March. Your vision lands as the next thing that isn't going to happen. The bigger the plan, the worse it lands. The team knows the constraints you haven't learned yet, and an ambitious plan that ignores them tells the team you haven't done the reading.

Before the team can believe where you're going, they have to see that what you say on Monday is still true on Friday. That sounds small. It isn't. Trust in a burned team is rebuilt one kept commitment at a time, and the first commitments have to be small enough that you can keep them no matter what the quarter throws at you.

There's a failure on the other side of this too. A leader who withholds direction for months, waiting until trust is fully rebuilt, teaches the team something different and just as bad: that this one is a caretaker. The fix is to sequence it, not to go silent. You still say where you think things are going, early, as a hypothesis you expect to revise. What you don't do is ask the team to reorganize their work around it before you've kept a single promise.

What the first-90-days playbook gets right, and where it runs out

If you've been handed a team, someone has probably handed you a copy of Michael Watkins' The First 90 Days as well. It earns its reputation. The core advice holds up: learn before you act, diagnose the situation before you pick a style, and secure early wins so the organization sees evidence rather than intent. Watkins' STARS model sorts transitions into start-up, turnaround, accelerated growth, realignment, and sustaining success, and his observation that each demands a different posture is the most useful thing in the book. In a conversation with HBR, he describes what happens to a leader who charges into a realignment as if it were a turnaround: the organization mounts what he calls an immune system reaction, and anyone who's watched a new executive get rejected by the team in six months knows exactly what he means.

Where the framework runs out for an inherited engineering team is that you're usually in two of those situations at once. The system is a turnaround: it's in visible trouble, everyone agrees, and it needs decisive action. The people are a realignment: they've stopped believing, the damage is underneath the surface, and nobody above you lists it as the problem you were hired to solve. Watkins pairs the turnaround with what he calls a heroic leadership style and the realignment with a quieter, more diplomatic, almost educational one, and he's right about both. The trap is that the heroic style is the one the team has already seen fail. The last leader looked decisive and confident right before the promises broke, so the posture the system needs is the one the team trusts least.

So you run both, and you keep them separate. Be decisive about the system: kill the top recurring alert, make the deploy boring, fence off what's stable. Be patient about the people, which means slower judgments, fewer reorganizations, and no conclusions about who's good until the environment has changed enough to tell you. A leader who lets one posture leak into the other either listens their way through a production fire or reorganizes a team they've known for three weeks.

Start with the record of promises

The companion article tells you how to assess the system, and that assessment starts in week one. A second assessment starts the same week, and it's the one to open with, because it has nothing to do with the stack and it explains everything on the list above.

What you're assessing is the record of promises. What did the team get told in the last two years, and what happened to each of those things? That record tells you which promise you absolutely can't be the next person to break. Three conversations get you most of it.

  1. Your boss. Ask what they were told about this team when they decided to hire you, and what they think is wrong with it. Listen for the story they tell about why it broke. It will usually be about the team or about the last leader, and it will rarely be about the decisions made above the team.
  2. Each person on the team. The three questions from the systems piece (what's working, what's painful, what would you change) still apply. Add one: "What was the last thing leadership promised you, and what happened to it?" Then stop talking. The pause after that question is where you learn what the team is waiting to see.
  3. The person everyone's answers point to. By the third conversation, one name will keep coming up as the person who really decides things. Talk to them after the other two conversations, but still within your first two weeks, and ask them what they've been holding together and which parts of it they'd like to keep.

The finding isn't in any single conversation. It's in the gaps between the three stories. Wherever your boss's version and the team's version disagree, that's where trust broke. If your boss says the team is slow and the team says they spent two quarters on a project that was cancelled without explanation, you've just found the first promise you need to keep: that cancelled work gets explained.

The first 30 days: fix the thing nobody believes will change

The systems piece makes the case for boring early wins and cites the research on the power of small wins, which found that visible progress on meaningful work is what keeps people engaged. The same mechanism works on trust, with a tighter definition of the win. A good first win is something the team has been asking for, repeatedly, and been told no on. It's usually small, and it's usually been "not a priority" for a year. The content matters less than the history, because the message you're sending is that a request from this team now has a different outcome.

Fix it fast and fix it visibly. The systems piece says not to make the person who's been asking for it lead it, and here's why: if you do, you've handed them the same work with a new boss attached. That holds when the team was burned by being blamed for what they owned. If they were burned the other way, by having every proposal overridden from above, then doing it for them repeats that wound instead, and the right move is to let them own it with you visibly clearing the path. Either way, say out loud that it happened because they asked. And don't make a habit of delivering every early fix yourself, because a team that watches the leader close every ticket learns to wait for the leader, which is the low-initiative pattern you're trying to undo.

The difference between a quick win and a foundational change is what each one asks the team to believe. A quick win asks them to believe you can deliver, and you can prove that in a week. A foundational change (a new on-call model, a real planning process, a different team structure) asks them to believe you'll still be here, still consistent, in six months. They can't believe that yet, and asking them to is how a good foundational change gets performative agreement and quiet sabotage.

You can also overdo this. A leader who only does quick wins, month after month, becomes a very senior ticket-closer. The team likes them and nothing structural changes. The quick win is a down payment, and the foundational change is what you're buying with it. Roughly, by the end of the first month you should have one visible win, one explicit "no" that you've kept to (something the team was told was coming that you've confirmed is not), and one foundational change announced with a date, not yet started. The no you keep does as much for you as the win you deliver. A burned team would rather hear a no they can count on than a yes that keeps getting pushed back.

Psychological safety without the workshop

Psychological safety gets turned into a poster, which is a shame, because the underlying behavior is concrete and you can watch for it. People say the thing in the meeting instead of in a DM afterward. That's all the definition you need. After your next meeting, watch where the real conversation happens.

You don't rebuild it with a workshop. You rebuild it in three moments, and all three will show up in your first few weeks whether you're ready for them or not.

The first is the first time someone disagrees with you in front of others. What you do in the next thirty seconds will teach them more than anything you say about culture. Take the disagreement seriously, ask them to say more, and visibly change your mind when they're right. When they're wrong, hold your position out loud and give the reason, because a leader who folds on every public disagreement is teaching a different lesson, and a team burned by indecision has already learned it. If you're defensive once, in front of the team, you've confirmed the old rule.

The second is the first incident. Something will break in your first few weeks, and the team will watch where you look. If you look at the person, you're the last leader again. If you look at the system that let one person's mistake reach production, you've told them something about what incidents are for. I've written a whole series on blameless post-mortems, and the psychological safety piece in it is the one to read before your first incident review. The trust version is simpler: the first incident on your watch is a test of what you do with blame, and the team already knows the answer they expect.

The third is your own first mistake. You'll make one, probably a judgment call you made too early on incomplete information. Say so out loud: what you decided, when, and what you're changing. A leader who talks about their own mistakes makes it safe for everyone else to talk about theirs. Do it for the calls that mattered, not for every misstep, because a running commentary on your own uncertainty reads very differently to a team that was burned by a leader who could never commit.

Two people you will definitely meet

The companion article covers the hero, the burned-out engineer, the skeptic, and the predecessor's loyalists. Two more people show up specifically when the problem is trust, and both of them need a conversation in the first two weeks, not the first two months.

The one who's been waiting to leave. Someone on the team stayed through the chaos for one of two reasons: they wanted to see who the next leader would be, or leaving in the middle of the mess felt disloyal to their teammates. Either way, they've been half out the door for a while, and your arrival was the trigger they were waiting for. Ask them directly. "Are you here to see what changes, or have you already decided?" It's a blunt question and they'll respect it. If they've decided, give them a clean exit and a good reference, and don't try to counter with vision, because vision is the thing they've stopped believing. If they happen to also be the hero engineer holding the system up, the clean exit still stands, and you've just moved knowledge transfer to the top of the technical plan. If they're waiting to see, give them something specific and dated to watch for, and then make sure it happens. This is the person your first kept promise is for.

The one who's been running things. When a team goes without real leadership for long enough, someone fills the gap. Usually it's the most senior engineer, and usually they didn't ask for it. By the time you arrive, they're making the calls, holding the context, and deciding what ships. The team trusts them, which means the team's trust in you will route through them for a while whether you like it or not.

There are two versions of this person. One is relieved you're here and wants to hand it back. The other has built something and isn't sure you deserve it. You can't tell which from the outside, and you won't know until you ask, so ask: what have you been holding, and what do you want to keep holding? Then decide with them, not about them. Say publicly what they carried, because the team already knows and your silence on it would be the first thing they noticed.

You can get this wrong two ways. Leave the shadow leader running the team while you "get up to speed" and you never actually take the job; the team will keep going around you, and by month three the arrangement is permanent. Remove or sideline them to establish authority and you've just removed the only person the team trusted, which the team will read as the new leader clearing out anyone with context. Take the job back on purpose, and in the same week say publicly what they carried.

Wrong people, or good people in a bad environment?

You'll be asked this, usually by your boss, usually early. "Is it the team?" Sometimes it is. Most of the time the honest answer in week one is that you can't tell yet, because a burned team is telling you about the environment it's been in, not about who's on it.

The way to find out is to change the environment in one concrete way and watch who moves. Keep a promise, kill the alert that pages someone every night for nothing, fix the thing they'd given up on, and then look at who re-engages. The people who were adapting to a bad environment tend to come back within a few weeks, and usually with a backlog of ideas they'd stopped voicing. The people who don't come back after the environment has demonstrably changed need a different conversation, and that conversation is fair now in a way it wasn't on day one.

Put a deadline on that test before you start, and write it down: two full sprints or release cycles, whatever your rhythm is. The one exception is active harm. Someone bullying a teammate, leaking, or quietly undermining the handback from the shadow leader doesn't get two cycles. Both answers to "is it the team?" have a cost if you commit to them too early. Assume everyone is salvageable and you'll keep someone who's wrong for the role, and the team will watch you tolerate it, which costs you trust in a different way. Assume nobody is, clear the roster, and you've lost the context and sent the message that you're the last leader with a new face. Change the environment first, then judge the people, and don't let your boss rush you past the first step.

When listening first is the wrong call

Everything above is about order, and you can get the order wrong by going too fast or too slow. Too fast looks like announcing a vision before you've kept a promise, reorganizing before you understand why the current structure exists, or replacing people before you've changed the environment that shaped them. Each of these feels decisive to you. To the team it looks like the last leader all over again. Too slow looks like listening until the team concludes nothing will change, which is a different kind of broken promise, while the shadow leader consolidates and your boss starts to wonder whether they hired a caretaker. "Listen first" only works for so long. Past that point, the listening itself becomes the thing the team holds against you.

The systems piece calls the first two weeks a listening tour, and tells you to sit through a deploy and an on-call shift without intervening. That advice assumes the system is stable enough to be observed and the team is waiting to see what you'll do with what you learn. Three situations break that assumption, and the usual advice for new leaders skips them:

  • An active incident. If production is on fire when you arrive, lead the incident. Don't observe it. A team that's been burned learns more about you from one well-run incident than from a month of one-on-ones, and standing back to "learn the system" during a fire looks like the same absence they've been living with.
  • A departure already in motion. If someone has a resignation half written, the listening tour doesn't reach them in time. That conversation happens in your first week, whatever else you'd planned.
  • A team burned by indecision rather than churn. Some teams weren't burned by a leader who promised too much. They were burned by a leader who never decided anything, and every request died in a backlog of maybes. For that team, another month of you listening is the same thing they just lived through. They need you to make one small decision in the first week and then stick to it, and the listening can run alongside it.

None of these is the premature decision the systems piece warns about. The warning there is about big calls made on incomplete information. A small decision, made fast and kept, is the opposite: it's the first piece of evidence the team gets that decisions here now have consequences.

Fix the system first, or the team?

You will almost always inherit both at once, a system that's fragile and a team that's stopped believing, and you'll be asked which to fix first. Neither. You don't finish one and then start the other. Each one makes the other easier, and that overlap is the thing to use.

The first system fix you choose should be the one the team has been asking for. A technical fix the team requested is a kept promise, so it does double duty: it reduces uncertainty in the system and it proves something about you at the same time. In the systems piece I described a team that scheduled every release around specific people being online, and the fix was to tighten the release path until a deploy was something anyone could do on a Tuesday afternoon. That was a technical change, but the reason it mattered to the team was that it was the first thing anyone had fixed for them in a long time.

Here's what a technical fix that didn't stick looks like, so you'll recognize it. The new on-call rotation exists in the document and not on the calendar. The planning process gets filled in after the work is already done, to make the board look right. The code review rule is followed to the letter and routed around in a DM. Every one of those is a sound fix that the team agreed to and never believed in, and that's the shape every foundational change takes when it lands before the trust does.

Let the trust that first fix buys you pay for the foundational changes the technical turnaround needs. Fixing the system gives the team a reason to trust you. The trust is what keeps the fixes from being undone. Do either one alone and you'll be back in a year, explaining to the board why the rewrite didn't take.

If you're the one doing the hiring

If you're the CEO, COO, or board member handing someone an existing engineering org, the most useful thing you can do before the search starts is to be honest with yourself about which job this is. A team that's shipping badly on a sound foundation needs a different leader than a team that's shipping fine on a broken one, and both are different again from a team that's stopped believing anything you say. If the job description describes the first and the exit interviews describe the third, you're hiring for the third.

Before you ask a candidate anything, tell them which job it is. Describe the team as it actually is: what it was promised, what fell through, who's been holding it together, who's half out the door. A candidate who only knows the job description will answer the job description, and you'll be judging them on a situation you never described. Then ask how they'd spend the first thirty days in that team, and listen for what they'd do first, second, and third. A candidate who opens with vision will re-burn a team that was burned by broken promises, unless the team was burned by a leader who never decided anything, in which case the candidate who can't name a direction is the bigger risk. A candidate who opens with "I'd listen" and can't say when the listening ends will lose the team to the shadow leader. The answer you want names a first kept promise, a first explicit no, and a date for the first foundational change, and it's honest that the system and the people are two different turnarounds that have to be run at the same time with two different postures.

Then give them the one thing a burned team's new leader needs most from above: don't make them break a promise in their first ninety days. Go back to the first of the three conversations. The story your new leader hears from you about why the team broke will rarely include the decisions made above the team, and the promises that broke it were, more often than the org chart admits, made there. Their credibility with the team is the only tool they have, and it only has to be spent once to be gone.


Build, Scale, Succeed

Join others receiving expert advice on
engineering and product development.

Newsletter Subscription

No data sharing. Unsubscribe at any time.