I mentor people. Some of it is the formal kind, where a company pairs two names and puts a recurring meeting in the calendar, and some of it is the informal kind that happens because somebody kept asking good questions and I kept answering them. In my experience the formal ones are far more likely to quietly die.
They die in a recognisable way. The first few months are genuinely useful. Then the meeting starts getting moved. Then it gets moved again. Nobody ever ends it; it just stops, and both people feel slightly guilty and slightly relieved. I have been on both sides of that.
The usual diagnosis is that people got busy. I think the real problem is structural: the standard model of mentoring contains its own expiry date.
Three reasons the one-way version stalls
The content runs out. The genuinely transferable material — how this organisation actually makes decisions, who needs to be spoken to before a proposal goes anywhere, how to scope a piece of work so it can ship, what a review committee is really asking when it asks a question — is a finite body of knowledge. It is extremely valuable and you can hand most of it over in a year. After that, if the only defined purpose of the relationship is transfer, you are both improvising, and improvised meetings lose to real deadlines.
The mentee stops being honest. This is the one people underestimate. If your mentor is senior in your own organisation and might be consulted, formally or not, when your promotion comes up, then "I am not sure I want the management track" is not a free thing to say. Neither is "I think the project I have been given is a dead end." What you get instead is a version of the mentee edited for safety. It is worth naming this directly: a mentor with influence over your career and a mentor with none are two different relationships, and the second is often the more useful one for the hard questions. Most people need both.
The mentor gets nothing they can feel. Goodwill is real, but it is slow-burning fuel, and it competes with everything else on a senior person's calendar. If a mentor never experiences the relationship as interesting — only as a duty performed well — attendance decays on a predictable curve.
The mentor side is not charity
There is reasonable evidence that mentors gain materially, not just morally. The most-cited example is Sun Microsystems, whose long-running programme was studied internally over more than a decade: mentors were promoted markedly more often than employees who did not take part, and both mentors and mentees were retained at higher rates.
Read that carefully, though. It is one company, and the selection problem is obvious: the people who volunteer to mentor are often already the people on a promotion track, and the study cannot fully separate the two. What it does is undercut the framing of mentoring as a favour the senior person performs at their own expense. Whatever the causal story, mentors are not the ones paying.
The harder question is why the benefit exists at all, and my own answer is fairly mundane.
What I actually get out of it
Three things, and none of them are noble.
The first is that mentoring forces me to articulate things I do by instinct. Explaining why I would reject a particular experimental design, or why a model that looks good on the validation set makes me nervous, means converting a reflex into an argument. A fair share of the time I discover, halfway through the sentence, that my reasoning is thinner than I assumed. That is uncomfortable and extremely useful.
The second is information. The further you get from the work, the more of what reaches you has been filtered — by your own managers, by people who want to look competent, by the natural human instinct not to bring bad news upward. Someone junior enough to still be inside the work, who trusts you enough to be blunt, will tell you what the review process actually feels like, where the tooling is quietly wasting everyone's afternoon, and which of your stated priorities nobody believes. I have changed decisions because of those conversations.
The third is the naive question. "Why do we do it that way?" asked by somebody with no investment in the existing answer is one of the few reliable ways to notice a convention you have stopped seeing.
Notice what none of these require: the junior person does not have to formally teach me anything. This is where a lot of reverse-mentoring programmes go wrong. Staging a twenty-minute slot in which a graduate explains social media to an executive is not reciprocity, it is a performance of reciprocity, and everyone in the room can tell.
Designing for it without the theatre
A handful of things make the difference, in my experience, and they are all small.
- Ask what they know that you do not, and only ask if you mean it. The question is worthless if you have already decided the answer is nothing.
- Let them set the agenda. If the mentee brings no agenda, that is information — either they are not getting value, or they are not comfortable naming what they actually want to discuss. Ask which.
- Say specifically when they have changed your mind. Not generic thanks. "I dropped that requirement after what you said about the review queue" is worth more than a hundred appreciative noises, because it is checkable.
- Say out loud at the start that ending it is allowed. Most mentoring relationships should end. Giving both people permission in advance is the difference between a clean finish and six months of rescheduling.
- Keep one mentor outside your reporting line. For the questions that are genuinely risky to ask, this matters more than seniority or subject match.
Where reciprocity is the wrong frame
I want to be careful not to oversell this, because the mutual-learning framing fails in at least three situations.
Someone in real difficulty does not need a balanced relationship. A person in their first job dealing with something frightening — a hostile manager, a project collapsing under them, a visa problem — needs help, and telling them the relationship must be two-way is a way of adding homework to a crisis.
Hierarchy is not the same everywhere. In workplaces with a strong seniority culture, and that includes plenty of Indian and East Asian organisations as well as parts of academia anywhere, instructing a junior person to teach their senior does not read as an invitation. It reads as an instruction, and it will be complied with rather than meant. Reciprocity in those settings has to be earned quietly over time, not announced in a kickoff meeting.
And sponsorship is asymmetric by design. Spending your own credibility to put someone's name into a conversation they are not in is not mutual, is not supposed to be, and is probably the single highest-value thing a senior person can do. Do not let a tidy theory about balance talk you out of it.
If you are the one being mentored
The most common mistake I see is treating the meeting as a place to receive advice. It is a much better place to test thinking. Arrive with a decision you are genuinely stuck on and an opinion about it, rather than a vague request for guidance, and the hour becomes worth something to both of you. A mentor who is being asked to break a tie is engaged. A mentor who is being asked to fill an hour is drafting emails in their head.
Second: your mentor is not one person. The person who can tell you how promotions really work here, the person who understands the technical domain you are moving into, and the person you can admit doubt to are rarely the same person, and expecting one relationship to carry all three is how you end up disappointed in a perfectly good one.
Third: you are allowed to disagree, and it is usually the most useful thing you can do. When a mentee pushes back on something I have said, one of two things happens. Either I explain the reasoning properly, which is the actual teaching, or I notice I cannot, which is more useful still.
The tell
The mentoring relationships of mine that lasted did not last because they were well designed. They lasted because at some point they stopped being describable as mentoring. The meeting stopped having a defined direction, the agenda came from whoever had the more interesting problem that month, and eventually one of us asked the other for help on something real.
That is the goal, and you cannot mandate it into existence with a template. What you can do is stop building relationships whose only stated purpose is one person emptying their knowledge into another, because those finish, and the ending is rarely dignified.
Tags
Taresh Sharan