Mentoring Ladder
Adjusting the level of direction given to someone according to their experience with the specific task, from demonstration through to sponsorship.
The common mentoring error is applying one style to everyone — usually the style the mentor prefers rather than the one the person needs.
The progression, matched to experience with the task at hand rather than to seniority in general:
Demonstrate. Do it with them watching and narrate the reasoning. Appropriate for something genuinely new to them.
Guide. They do it, with close support and frequent check-ins.
Delegate with review. They do it and you review the outcome, not the process.
Delegate. They own it, and you are available if asked.
Sponsor. You create opportunities for them and advocate for them where they are not present — which is the level with the largest effect on a career and the one most often omitted, because it costs the mentor something.
Two distinctions worth holding: teaching reasoning beats teaching answers, since answers expire and the ability to reach them does not — which means explaining why a decision was made, and letting people make reversible mistakes; and mentoring is not the same as reviewing, though architects frequently substitute the second for the first.
The multiplier argument: an architect who raises the design capability of twenty engineers has more effect than one who personally designs everything, and the second does not scale in any case.