Stretch, sponsorship, exposure

Ask a manager how they grow their engineers and most will describe advice: the design reviews, the hard question answered well in a 1:1. For my first year managing that was my whole picture of development, and I told myself it counted. All of it is useful, and almost none of it advances a career. The development literature has a number for the gap, the 70-20-10 split: seventy percent of how people grow comes from challenging work, twenty percent from other people, and ten percent from formal training. The advice is the ten and the work is the seventy, and I had been spending my energy on the ten for the obvious reason: an answer across a 1:1 table costs nothing and cannot go wrong in front of anyone.

Three instruments live inside that seventy, and managers blur them constantly:

• Stretch is work beyond someone's current ability, the assignment they are not quite ready for.
• Sponsorship is your own credibility spent on their behalf in rooms they are not in.
• Exposure is visibility you engineer for them, putting their work and their name in front of the people who decide things.

Mentorship, the advice mode, helps someone do better at the job they have, while these three reach for the job they don't have yet. Naming the three is the easy part, since all the damage happens in the handing out.

Co-author it, don't hand it down

The default picture has opportunity flowing downhill from you: you decide an engineer is ready for the hard project, so you give it to her; you decide someone belongs in the promotion calibration, so you name him. That direction of flow is the problem.

An assignment handed down gets received as one of two things, and neither is good. If the person reads it as a gift, it looks like favoritism, and the rest of the team starts on the arithmetic: why her and not me, what does she have that I don't. If they read it as more work dropped on a full plate, you get resentment, because from where they sit it is just the thing nobody else wanted. The same stretch, in the same words, can land as either, and you do not get to decide which; the person receiving it does, with far less information about your intentions than you have.

Picture the migration nobody wants to own, handed to a junior engineer with a line in passing about what a great chance it is to own something end to end. The senior who handed it over walks away sure he has done her a favor. What she heard was different: I am now responsible for the job nobody wanted, alone, on top of everything I already had. Weeks later it is behind schedule, and the story that travels is that she couldn't handle the scope. The opportunity was real, and the handover wasted it.

A better speech will not fix this, because the wording was never the broken part. Stop handing the thing down and build it with the person instead: ask where they want to be stretched and what kind of visibility matters to them, then shape the assignment around the answer. Co-authored, the same migration carries consent and ownership, because the engineer had a hand in choosing it. Where that conversation happens is its own instrument.

Consent is the line between developing someone and setting them up, and every mistake in the rest of this essay is a version of skipping it.

Ration it honestly, the capital is finite

You cannot do this for everyone equally, and pretending you can is a fiction your team sees through. You have a fixed amount of credibility to spend and a limited number of visible projects to hand out, and sponsorship is the tightest of the three constraints: the credibility you spend advocating for one person is credibility you are not spending on another. Sponsor all six reports in the same review and the word stops meaning anything.

So you ration, and the honest move is to ration in the open rather than behind a pretense that everyone gets the same. The people who did not get the visible project should know there was a project, that it went to someone, and why. The pretense is tempting because it defers a cost: telling someone they were passed over is one awkward conversation this week, while hidden rationing builds the exact favoritism story you were trying to avoid, among people who can see the outcome and were never told the rules.

This is the growth version of a harder question, where your best people go. That one asks who you put on the work that matters most to the business; this one asks who you put on the work that would grow them most. The two pull apart more than you would like, because the person who would grow the most from the hard project is frequently not the safest pair of hands for it, and choosing growth means choosing some risk to the outcome, on purpose.

Stretch: match it to readiness, take something off the plate

A stretch is supposed to be uncomfortable, work slightly beyond reach, where the person is operating without full competence yet. But the same hard assignment can grow someone or grind them down, and the support around it decides which.

Hand that same hard, visible project to someone who is genuinely not ready, attach no coaching, take nothing off their existing load, and leave no way back if it goes wrong. They are now carrying a high-stakes project in front of an audience, with none of the support that would let them grow into it, and when it fails, which it often does, the failure reads as theirs, a personal flaw rather than a staffing decision. That pattern has a name, the glass cliff, and it stays common for an unlovely reason: the person carries the blame, and the manager who built the setup rarely does.

Three things keep a stretch on the developmental side of that line:

• Coaching, so they are reaching with a hand nearby.
• Removed load, because something has to come off the plate to make room.
• Reversibility, a way back that does not end a career if the thing goes wrong.

Get those right and the same difficult assignment grows someone instead of breaking them.

The hardest judgment inside a stretch is telling struggle from drowning. Struggle is the point, because you are meant to let people work at the edge of their ability, and rescuing them too early is how you waste the stretch. Drowning is the stretch failing, and from the outside it looks almost identical to struggle. Reading that difference, and not pulling the plug on someone who is just in over their head for a while, is most of the skill.

Sponsorship: risk you carry, not praise you give

Praise is cheap, being public and costing you nothing. Sponsorship runs the other way: it is often private, and it is a bet whose downside you personally carry.

The clearest version is the room the person is not in, a calibration meeting or a promotion committee, where their next opportunity is being decided and they have no voice in it. You put your name on them, saying in effect: I vouch for this person, route the opportunity here. From that moment their stumble is partly your miss, and if they don't deliver, you were wrong in front of the people whose opinion of your judgment matters.

It would be nice to engineer that cost away, and it would ruin the instrument. Your word carries weight in that room only because everyone present knows you pay when it is wrong; strip out the risk and you are back to praise, which the room has already learned to discount. This is also why you should only sponsor what you have co-authored: advocate for someone's readiness without having built the work alongside them and you are vouching for a readiness you assumed rather than one you watched. The craft of making that case upward, or sideways to peers, belongs elsewhere.

Exposure: engineer it for the ones who won't chase it

There is a popular framing of career advancement, sometimes called the PIE model, that weights performance at ten percent, image at thirty, and exposure at sixty. Its claim is that being seen by decision-makers matters more than the quality of the work, and taken as personal advice it becomes a license to self-promote, which is why the people who follow it most eagerly are rarely the ones doing the best work.

Read it as a leader instead and the same numbers describe a gap you are well placed to close for someone else. The strong, quiet engineer who does excellent work and never mentions it goes unseen for exactly the reason PIE names, and they will not chase the exposure themselves, which makes engineering it your job.

On the ground this looks like a plateau that makes no sense on the merits: someone whose work is consistently among the best on the team keeps getting passed over for the visible role and the named credit, because they do the work and then go quiet, and nobody with influence has seen any of it. You correct it by routing visibility on purpose: hand them the design review to present, put their name on the thing they built, and send them into the room instead of carrying their work in for them.

Exposure goes wrong in two directions. Visibility with nothing behind it gets seen through, and it costs the person credibility rather than earning them any. Exposure handed to someone not ready for the scrutiny is the glass cliff again in a different costume. The exposure worth engineering is good work that already exists, made visible, for someone who can stand the scrutiny.

What it comes down to

Stretch, sponsorship, and exposure fail the same way: each does harm in its unilateral version, where the stretch becomes a dumping, the sponsorship becomes a bet you had no business making, and the exposure becomes a glass cliff. Each of them works when the person had a hand in choosing it, which brings all three back to one conversation small enough to feel skippable: where do you want to be stretched, and what kind of visibility matters to you.