Sponsorship is what you say when they're not in the room
Coaching changes someone's work while you're in the room. Sponsorship changes what happens to their career when you're not, and it only runs on trust you've actually earned rather than trust you've claimed by title.
By JP LeBlanc
core5–150+ engineersAlso a Playbook →

Coaching happens with the person right there. You watch them run a hard meeting, or read their doc, and you tell them what would make it better. They get the benefit on the spot.
Sponsorship happens somewhere they can't hear it. A calibration room, a staffing conversation, a hallway aside about who should own the new thing. What you say there, unprompted, when there's nothing in it for you and no one checking, is basically all of it.
Most managers think they already do this. Most don't, not consistently, because coaching is cheap once you've noticed the thing worth saying, and sponsorship costs you something real every time. Social capital, spent on someone else's behalf, in a room they'll never see.
Credit the person, not the department
Here's the habit that actually moves something, and it's smaller than people expect. Notice what you say about someone's work when they're not there. Then say it on purpose.
Not "the team shipped this." Say who did it, attached to the actual decision. "Priya caught that the migration would've taken the write path down for four hours, and rewrote the rollout before anyone else even saw the risk." One sentence like that, said once in the right room, does more than a year of praise the person never gets to see land anywhere.
Translate it into whatever language the room already trusts, too. A VP doesn't reliably hear "great engineer" and file it anywhere useful. A VP hears "this is the person who kept the checkout redesign from slipping six weeks," because that sentence has a number and a consequence in it, and sentences like that are the ones that get repeated after you've left.
Put people forward for the stuff that's visible from outside their own team. The cross-functional project. The talk. The customer call where a senior engineering voice actually changes the outcome. None of it replaces the work itself. It's the difference between doing excellent work in a room nobody else enters, and doing the same work somewhere the people who'll eventually decide things can see it happen.
And when the story about someone is wrong, correct it, specifics and all, even when it's inconvenient to. "That outage wasn't caused by what Marcus shipped, it exposed a gap that had been sitting there for eight months" is a sentence some people would rather not say, because it complicates a tidy narrative someone else has already settled into. Say it anyway. The story people carry about a person when that person isn't in the room is the actual reputation. Whatever's true underneath it doesn't matter much if nobody's saying it out loud.
What it isn't
Sponsorship isn't a promise. You're not committing to a promotion or a rating or any outcome you don't actually control, and saying that plainly protects both of you.
It also isn't cover from honest evaluation. If the work has a real gap, sponsoring someone means calling out that gap accurately in the room too, not smoothing it into something flattering. The second sponsorship starts meaning "I'll make sure this looks better than it is," it stops being advocacy. It becomes a liability the first time anyone checks the claim against the work.
Calibration is where this gets tested most directly. A room built to normalize ratings across managers is also, structurally, a room that rewards whoever argues hardest for their own people. Walking in there and sponsoring someone well means you've already done the unglamorous part beforehand: the specific example, the real outcome, the thing that happened because this person, specifically, was the one doing it. Advocacy without evidence is just noise cancelling out everyone else's noise.
The work that's easy to miss
Some contributions are loud by default. A shipped feature announces itself.
The most valuable work a senior engineer does is usually the opposite kind. The system redesign that quietly prevented an outage nobody's ever going to hear about, because it never happened. The two hours spent unblocking a junior engineer that shows up on no dashboard anywhere. The proposal that got withdrawn because the senior engineer talked the room out of six wasted weeks before anyone started them. A staff-plus role is defined by reach across teams, and that reach, almost by definition, doesn't leave a record anyone downstream can read without help.
So that's your actual job here. Translate it. Write down what changed and why, in terms a leader two levels removed can follow without the technical context. Credit them in the sentence that explains the decision, not just in the commit that made it. Help them document the outcome themselves, too, since the version they write is usually more accurate than whatever you'd reconstruct from memory eight months later.
Then open the door and step back. Nominate them for the stretch project and let them run it on their own authority, not yours. Sponsorship that needs you in the room to keep functioning isn't sponsorship. It's dependency with better branding.
Trust is the actual constraint
None of this works without trust, and trust isn't something you get to claim because of your title.
People show up to a new manager relationship carrying the last one with them. Someone who was ignored when they raised a risk early. Someone whose manager was blindsided in a review by something that had been building for months, in plain sight. Someone who did the unglamorous work for two years and watched somebody louder take the credit for it. Trust is a price people set based on what they've already watched you do, and a new manager inherits a price that has nothing to do with them and everything to do with the last three people who held the job.
You don't talk your way to a lower one. You earn it slowly, one boring habit at a time. Asking a question and actually remembering the answer three weeks later. Saying "I don't know" instead of inventing something plausible-sounding on the spot. Keeping a commitment, and when you genuinely can't, saying so early rather than letting the date arrive and explain itself. Giving credit away in public and taking responsibility in public when the team falls short, in that order, every single time, not just when it's convenient.
One thing breaks all of this faster than anything else does. Using something someone told you in confidence, later, even gently, even in service of a point you happen to believe is true. Confidentiality kept selectively isn't confidentiality at all, and everyone recalibrates the price the second they notice.
I had a version of this wrong for a while, honestly. I thought one honest, vulnerable conversation, done well, was what built trust with a team. It isn't. It's a data point, and a good one, but it's one data point set against months of small consistent behavior, and treating it as if that one conversation could carry the whole weight is how you end up confused later, wondering why trust didn't follow the good conversation the way it was supposed to.
The confidence that isn't
Watch for the manager who builds their own standing by making the room slightly smaller for whoever else is in it. It reads as confidence. It isn't.
If your certainty depends on someone else feeling less capable, that's not certainty, it's a comparison you're running to feel steady yourself. Real confidence can say "I don't know" out loud, in front of the team, and let the best answer in the room win even when it isn't yours. That's what actually makes it safe for the next person to say the thing they're not sure about, which is the whole supply chain sponsorship runs on underneath everything else. You can't advocate for people's work in a room where nobody feels safe saying out loud what they don't yet understand.
Managing up on someone else's behalf
Sponsorship and managing up run on the same move: make the true thing visible to people who weren't there to see it happen, in language they can actually use.
The difference is where the story centers. Managing up in service of yourself makes you the protagonist of every update. Managing up in service of your team makes outcomes, risks, and needs visible, and keeps you out of the sentence wherever the sentence loses nothing true for it.
The line between coaching and sponsorship comes from a leadership guide a former colleague wrote at CircleCI.
The practice underneath it is simple to state and genuinely hard to keep doing. Say the accurate, generous thing about someone's work in the rooms they can't get into. Do it before you need anything back from them. Do it whether or not anyone's watching you do it. Most of what people end up calling sponsorship is really just this, repeated often enough that it turns into a reputation you never had to build any other way.
Read this if
- What's the difference between coaching and sponsorship
- How do I advocate for someone in calibration
- Staff engineer isn't getting recognized
- How do I earn my team's trust