Promoting Your Best IC to Engineering Manager Without Ruining Two Careers
Last updated: June 12, 2026 | By Kris Drouet, in partnership with KORE1
Promoting your best IC to engineering manager without a real transition plan is how you ruin two careers at once. The move only works when you treat it as a career change, with a written plan, real training, a named coach, and a documented way back.
I ruined the first engineer I ever promoted.
Call him Marcus. The real Marcus is still out there shipping code, better at his job than I ever was at mine, so the name stays changed. Marcus was the strongest TypeScript engineer on a loan-origination platform team I ran in my mid-thirties, the guy who rebuilt our document-intake pipeline by himself while the rest of us were still arguing about the design doc. I handed him the manager title on a Friday afternoon. By Monday he had eleven direct reports, a calendar that went from four meetings a week to twenty-three, and no plan from me beyond a congratulations email that I am still embarrassed to reread. Ten months later he resigned. The senior engineer who trusted him most, the one who had been quietly interviewing since the week Marcus stopped reviewing pull requests, was gone within the quarter.
Two careers. One signature on a promotion form. Mine.
Every director who has asked me about promotions since then has gotten the same line, and I will keep saying it until it stops being true. Promoting your best IC into management without a real transition plan is how you ruin two careers simultaneously. Theirs, and whoever has to clean up after them. Earlier this week I wrote about why engineering velocity stalls, and I named this exact promotion as the hiring mistake that compounds everything else. This piece is the plan I wish someone had handed me before I handed Marcus that title. KORE1’s engineering staffing recruiters asked me to write it down because they keep getting the phone call that comes nine months after a promotion like mine, when the team is down two strong people and nobody can quite explain how.
Management Is a Career Change, Not a Promotion
Moving from individual contributor to engineering manager is a career change, not a rung on the same ladder. The skills that earned the promotion, deep technical judgment and personal output, stop being the job. The new job is leverage through other people, and it has to be learned like any discipline.
Our industry refuses to believe this. We treat the manager title as the default reward for technical excellence, the same way sales orgs hand the team to the rep with the biggest number, then act surprised when the person who loved building things spends every working day in a job that contains no building at all.
Charity Majors wrote the canonical essay on this back in 2017, the engineer/manager pendulum, and the most useful idea in it has barely aged. Going into management is not a step up. Coming back to engineering is not a step down. They are two different crafts, and the strongest technical leaders I know have swung between them more than once. Gergely Orosz has covered the same transition in The Pragmatic Engineer and lands in the same place. New job. Not bigger job.
If you only fix one thing in your org’s vocabulary this year, fix that one. The word “promotion” is doing damage every time someone says it.
Why Your Best Engineer Is Your Riskiest Pick
There is actual research on this, and it is uncomfortable reading for anyone who has ever run a promotion calibration. Economists Alan Benson, Danielle Li, and Kelly Shue tracked sales workers across 214 American firms and published the results in the Quarterly Journal of Economics in 2019. Companies overwhelmingly promoted their best individual performers. A rep who sold twice as much as a peer was about 15 percent more likely to get the manager job. And once promoted, the best sellers made measurably worse managers, with their teams’ performance declining relative to teams that got a manager with stronger people signals and a weaker personal number. The researchers called it the first large-scale empirical confirmation of the Peter Principle. I call it Tuesday.
That study is sales, not software. The mechanism transfers anyway, because the mechanism is not about the work. It is about what promotion committees can see. Personal output is legible. Coaching ability is not. So we promote the legible thing and hope.
Gallup has spent two decades quantifying what that hope costs. Their research puts the manager at the center of roughly 70 percent of the variance in team engagement, and their State of the American Manager work estimates that about one person in ten carries the natural mix of traits the job actually requires. One in ten. Your best engineer might be in that ten percent. Plenty of mine have been. But their commit history tells you nothing about it either way, and the commit history is usually the only evidence in the room when the decision gets made.
Think about what makes a great senior engineer. Depth. Long stretches of focus. The stubbornness to chase a memory leak through a weekend because it offends them personally. Now look at the manager job under the hood. It is interruption as a lifestyle. It is saying the same thing about priorities eleven times without sounding bored. It is sitting with an underperformer on a Wednesday afternoon and caring more about their next quarter than your own backlog. Some people cross that gap and thrive. Nobody crosses it by accident.
What Actually Gets Ruined, and for Whom
Career one belongs to the engineer you promoted.
The damage starts quietly. Their technical sharpness begins decaying the day the calendar fills, and they know it, which is why so many new managers keep a private branch going at midnight, long after anyone has stopped asking them for code. Marcus did. Meanwhile nobody has defined what good looks like in the new role, so they grade themselves on the old metric, personal output, and the grade keeps coming back worse. The Chartered Management Institute and YouGov surveyed 2,524 UK workers with management experience and found that 82 percent had walked into the role with no formal management training at all. Accidental managers, the report calls them. The same study found a third of workers had quit a job over the culture those accidents create.
Career two belongs to whoever inherits the mess.
A few years ago I joined an org where a promoted IC had been failing slowly for fourteen months. Nobody had told him. Six engineers, two open reqs that had not seen an interview in ninety days, and retro notes that read like ransom letters. He did not wait for rapport. He asked to go back to writing code about ninety seconds into our first one-on-one, while the coffee between us was still too hot to drink. We moved him to a staff engineer seat that month. He is the best technical lead in that company today, which tells you the failure was never his. The director who promoted him without a plan spent the next year rebuilding a team that did not need to break, and the manager we eventually placed into that seat started fourteen months behind on trust she did nothing to lose.
Count the careers in that story. I get to four.
The Transition Plan I Wish More Directors Wrote
None of this argues against promoting ICs into management. Some of the best engineering leaders I have worked with came up exactly that way, including people I promoted after Marcus taught me what not to do. The argument is against doing it with a title change and a prayer. A real transition plan fits on one page. Here is the shape mine take.
| Phase | When | What Happens | The Checkpoint |
|---|---|---|---|
| Audition | One to two quarters before any title change | Acting tech lead on a single initiative. They run planning, run an incident review, and sit in on two interview loops. Still shipping code. | They can tell you whether they want the work, not just the title. |
| Day-one reset | First two weeks | Written role definition. Calendar rebuilt around one-on-ones. Off the PagerDuty rotation and off the critical path. Coach assigned by name. | Every engineer on the team can say who owns what now. |
| First 90 days | Months one through three | Real management training, not a lunch webinar. Weekly sessions with the coach. First feedback conversations happen with backup standing by. | One hard conversation completed without you in the room. |
| The verdict window | Months six through twelve | Full ownership of hiring, performance, and delivery. A scheduled pendulum conversation at month nine, on the calendar from day one, not improvised mid-crisis. | Both of you can answer the question: would they take the job again? |
The table is the skeleton. Five commitments from you, the director, put muscle on it.
- Write down what changes. Success metrics, decision rights, what they stop doing. If the role definition lives in your head, you have not defined a role. You have made a wish.
- Training before the title. The 82 percent number above is not a UK quirk. It is the industry default, and you fix it for a few thousand dollars before day one or you pay for it in attrition later.
- A named coach who has done the job, and not you. You are their boss. Bosses are the last people new managers practice honesty with.
- A return path in writing, with pay parity. This is the one almost everyone skips. Glassdoor’s 2026 averages put engineering managers around $226K against roughly $193K for staff engineers, which makes management look like the money move. Matched level for level inside a company with a real dual track, that gap mostly evaporates, and every comp review I have sat in confirms it. If the only way your ladder gives a great engineer a raise is handing them direct reports, the ladder is the bug. KORE1’s engineering manager salary guide breaks the bands down, and their salary benchmark assistant will tell you what the seat pays in your market before the conversation starts.
- A backfill plan for the work they carried. Promote your best engineer with no backfill and you have engineered the worst trade in staffing. The team loses its strongest builder and gains a manager who is still secretly doing the old job at night.
One page. Maybe a day of your time to write. I have watched the absence of that page cost a company two staff-level engineers and a year of roadmap, so the return on the day is hard to beat.
The Conversation to Have Before You Offer the Title
None of the plan survives a candidate who never wanted the job. The trouble is that asking “do you want to be a manager?” tells you almost nothing, because nearly everyone says yes to it, and the yes arrives long before they have pictured what the days actually contain. The title flatters. The raise tempts, even when the raise is mostly mythology.
Ask what they think you do all day. The answer is diagnostic. If they describe your job as “unblocking the team and setting direction,” they have been paying attention. If they describe it as “going to meetings,” they are about to be very unhappy, because they are right and they hate it already.
Ask when they last enjoyed helping someone ship something they got zero credit for. Not tolerated. Enjoyed. The engineers who light up at that question are already doing the job informally, maintaining the onboarding doc nobody asked them to write, untangling the GitHub review queue because mess bothers them. The ones who go quiet were hoping management meant more authority over the architecture. It does not. It means less.
Then run the cheapest experiment in leadership. Hand them your one-on-ones for two weeks while you travel. A fortnight of real management work teaches both of you more than a quarter of career-development conversations.
When the Right Manager Is Not on Your Team
Sometimes the audition fails. Sometimes nobody on the team wants the job, which is a healthy outcome and not a crisis. Sometimes you simply cannot afford to lose your best builder’s output this year, and the org is too fragile to absorb a first-time manager learning the job in production while the roadmap is already slipping.
Then you hire the manager and keep the engineer. A first-line leader who has already made their mistakes on someone else’s team walks in with the one thing your promoted IC cannot have, which is pattern recognition from doing it before. The vetting is its own discipline. Interviewing a manager is nothing like interviewing a coder, and I wrote up some of what to listen for in KORE1’s engineering manager interview questions guide. Their direct hire staffing team does this search constantly, with recruiters averaging 15-plus years in the business, and 92 percent of the people KORE1 places are still in the seat a year later. In a decision where the failure mode costs you two careers, that retention number is the one I would anchor on.
Promote or hire, the math is the same. The expensive option is the unplanned one.
Where to Start Monday Morning
Do not start with the org chart. Start with the one-page plan. Write what changes, name the coach, book the month-nine pendulum conversation, and put the return path in writing before you say the word “promotion” out loud to anyone. If the plan feels like overkill, you have never watched a Marcus resign.
Staring at one of these decisions right now? Connect with me on LinkedIn and tell me what you are weighing. I read those messages. And if the honest answer is that the right manager is not on your team yet, talk to KORE1’s hiring team before you talk yourself into the easy promotion.
The Questions I Get After a Promotion Goes Sideways
How long before you know whether a new engineering manager is going to make it?
Two quarters for the first honest signal. Twelve to eighteen months for the real verdict, and rushing that timeline causes the exact failure you are trying to detect.
The 90-day signals are narrow but readable. Is the team’s work still legible to you? Has the new manager had one genuinely hard conversation? Are their old teammates still bringing them problems, or have they started routing around them? By month six the trend is usually set, and across 25 years of watching these transitions I can count on one hand the cases where a manager who was drowning at month nine learned to swim by month fifteen. The scheduled pendulum conversation exists for exactly that reason.
They hate it and want their old job back. Now what?
That is the system working. Move them back to a senior IC seat at full pay, say thank you in public, and treat the whole episode as a successful experiment instead of a failure anyone needs to apologize for.
The pendulum only functions if the return trip carries no shame and no pay cut. Punish the first person who swings back and you will never get an honest answer from a struggling manager again.
Do we take the code away from them on day one?
Take them off the critical path, not out of the codebase. A new manager who owns sprint-blocking work is a bottleneck with a fancy title, but one who never reads a pull request goes blind inside a year.
The line I draw is simple. Nothing the sprint depends on. Plenty that keeps their hands warm, code review, prototypes, the internal tool nobody prioritizes. The leaders who stay close enough to the code to smell a bluff are the ones whose teams never bother bluffing. That is a longer argument, and it is the next piece I am writing in this series.
Does the manager title have to come with a raise?
Parity, not premium. Pay the level, not the org-chart shape, and keep the senior IC band overlapping the manager band so nobody takes on direct reports just to get the market adjustment they deserved for the work they were already doing.
If your bands force a raise to justify the move, every comp cycle quietly manufactures more accidental managers. Fix the bands once and the promotion conversation gets honest forever after.
Can someone really manage the engineers who were their peers last sprint?
Usually, yes, and it is awkward for about a quarter no matter what anyone does. The transitions that fail are the ones where nobody names the awkwardness out loud.
The new manager needs to have an explicit reset conversation with each former peer in the first two weeks. Friendships get renegotiated, not pretended away. The hardest case is the peer who also wanted the job, and that one cannot be delegated. You, the director, owe that engineer a direct and early conversation about their own path, because resentment that gets no acknowledgment goes shopping for a new employer instead of a new argument. Skip it and you will be backfilling two seats by spring, which is the phone call KORE1’s recruiters tell me they get more often than any other in IT staffing.