On being a first-time mentor: the five mistakes I made so you don't have to

The jump from "I know this stuff" to "I can teach this stuff" is bigger than most first-time mentors expect. A year into mentoring on VisNavigans, here are the five mistakes I made so you don't have to.

1. Assuming shared context

The curse of expertise: when you know something deeply, you forget what it was like not to know it. You skip the three concepts your mentee needs, land on the fourth, and wonder why they look confused.

The fix is boring and effective: ask "what's your background?" at the start of every session. Then, more critically, ask "is this making sense?" every 5–10 minutes. Not as a rhetorical check-in — as a real yes/no question that you're prepared to hear "no" to.

I lost three mentees to this in my first month before I caught on. They all said the sessions were "great", then ghosted. The feedback form said "too fast, felt dumb". Which is my fault, not theirs.

2. Giving answers instead of asking questions

Your mentee asks "should I use React or Vue for this project?" and you know the answer, so you say it. They nod, go build in React, and six months later you're having the same conversation about their next project.

The mentor version of this is: "what's driving the choice?" Let them work through the reasoning out loud. You're training them to make decisions like you, not creating a dependency where every choice requires a consultation.

Socratic method. Ancient. Still undefeated.

3. Being too kind to the work

Your mentee sends you code. It's bad. You want to be encouraging. You write "looks good! small nits below..." and then bury the actual issues in gentle language.

They don't fix the issues because they can't tell which feedback is important. They think their code is 90% right; it's 40% right. The gap between their perception and reality widens until some embarrassing failure forces reality in — and they wonder why you didn't tell them sooner.

The fix: separate kindness (as a default) from honesty (as an obligation). You can be kind about the person while being brutal about the code. "This logic has a race condition and won't work under load. Separately — good move on extracting that helper." Two sentences, two moods.

Most mentees want honest feedback. They're paying for it specifically because their friends and colleagues won't give it. Don't be another "nice" voice.

4. Talking more than listening

The trap is pattern-matching. Mentee says "I'm stuck on X". You've seen X before. You launch into a 10-minute monologue about X and all its variations. But their X isn't the X you've seen — they have a specific constraint, or they're further along than you assumed, or they have context from their team you don't know.

The one-sentence fix: every time the mentee starts describing a problem, let them finish. Resist the urge to interrupt with the solution. Ask one clarifying question before offering any advice. Your first hunch is wrong 40% of the time; the clarifying question is usually where the real issue surfaces.

5. Not having a structure

My first sessions were aimless. Mentee showed up, we chatted, they left having learned something, but there was no arc. The value per minute was low and I could tell they were politely satisfied rather than excited.

Now I run sessions on a rough structure:

1. 2 min — quick check-in on any homework from last session.
2. 3 min — what do they want to focus on today? (Whatever they wrote in the brief.)
3. 35–45 min — the work. Could be code review, discussion, working through a specific problem.
4. 5 min — synthesis. What are the 2–3 takeaways? What are they doing this week?
5. 5 min — open questions, random topics, whatever they want.

The structure feels rigid at first, but mentees love it. They leave knowing exactly what to do next. Sessions get denser. Ratings go up.

Bonus: the meta-mistake

Not asking for feedback from your mentees. A mentor who doesn't improve at mentoring isn't a mentor — they're an expert giving advice. After every session, ask: "what worked, what didn't, what should I do differently next time?" Most won't say anything useful the first few times. Keep asking. Eventually they'll give you the gold.

I now have a running doc of notes from mentee feedback over the past year. Half of what I do well came out of it.

You'll make new mistakes

That's the deal. Every mentor I've talked to has their own "five mistakes" list, mostly different from mine. The point isn't to avoid all of them — it's to notice when you're making one and course-correct fast, before it costs a mentee their trust.