The Stakes/Familiarity Matrix: A Better Way to Decide What to Delegate
A simple 2x2 framework for deciding how closely to supervise a delegated task, based on two things that actually predict how it goes — not gut feel.
Most delegation advice boils down to "match the task to the person," which is true and not very actionable. The harder, more useful question is how closely to stay involved once something's handed off — and that decision is more predictable than it feels once it's broken into the two things that actually drive it.
The two questions that matter
How bad is it if this goes wrong? Not "how would I feel" — the actual, concrete consequence. A typo in an internal draft and a miscalculated client invoice are both "mistakes," but they sit at very different points on this axis.
How familiar is this person with this specific kind of work? Not their seniority in general — their actual experience with this particular task. A senior employee doing something genuinely new to them is, for this specific task, a novice, and treating them as an expert because of their title elsewhere is exactly how avoidable mistakes happen.
Plotted against each other, these two questions produce four distinct zones, each calling for a different, specific level of involvement.
The four zones, in practice
Pair closely (high stakes, low familiarity). This is where real, hands-on support belongs — working through the first attempt together, not just explaining it and checking the result afterward. This is also the zone most often under-supported, because a manager assumes a generally capable person doesn't need this level of help, missing that capability elsewhere doesn't transfer to a genuinely unfamiliar task.
Check in lightly (high stakes, high familiarity). The person can clearly handle this — a brief check-in at a natural milestone is enough, not because the stakes don't matter, but because their track record on this specific work has already earned a lighter touch.
Let them practice (low stakes, low familiarity). The ideal place to build familiarity, precisely because a mistake here is genuinely cheap. This zone is also where over-supervision does the most quiet damage — hovering over low-stakes work someone's still learning removes the safe practice opportunity the low stakes were supposed to provide.
Hand it off fully (low stakes, high familiarity). No meaningful supervision needed. Checking in here anyway is a pure trust signal in the wrong direction — it tells a capable, experienced person their competence on familiar, low-risk work is still in question.
Why this beats going on instinct
Instinct tends to calibrate supervision to how anxious a task makes the manager personally, which correlates only loosely with the task's actual stakes and the person's actual familiarity. An anxious manager over-supervises the "hand it off fully" quadrant because the task still feels uncomfortable to release, while a confident manager under-supervises "pair closely" work because their general trust in the person overrides the fact that this specific task is genuinely new to them. The framework forces both questions to be asked explicitly, rather than letting a single gut feeling stand in for both.
Applying it to a real decision
Before delegating something, place it honestly on both axes: what's the real cost if this goes wrong, and has this specific person done this specific kind of work before. The quadrant that results should set the starting level of involvement — not a fixed rule for that person going forward, since the same person moves between quadrants constantly as they take on new kinds of work throughout their role.
This connects directly to building genuine delegation habits over time — the matrix answers "how much involvement right now," while the broader skill is recalibrating that answer honestly as familiarity builds task by task, rather than defaulting to the same supervision level regardless of what's actually being handed off.
The short version
Two questions — how bad is it if this goes wrong, and how familiar is this person with this specific task — sort any delegation decision into one of four zones, each calling for a different, specific level of involvement. Using both dimensions explicitly, rather than a single gut-feel judgment, corrects the two most common delegation mistakes: over-supervising safe, familiar work, and under-supervising genuinely risky, unfamiliar work.
Frequently asked questions
What is the Stakes/Familiarity Matrix?
A framework for deciding how closely to supervise a delegated task, based on two questions: how bad is it if this goes wrong, and how familiar is this person with this specific kind of work. The answer to both questions together, not either alone, should set the supervision level.
Why isn't seniority a good basis for deciding how closely to supervise someone?
Because familiarity is task-specific, not a fixed trait of the person. A senior employee doing something genuinely new to them needs real support on that specific task, regardless of how independent they are on their usual work.
What's the most common mistake managers make using intuition instead of a framework?
Calibrating supervision to how anxious the task makes the manager feel, rather than to the task's actual stakes and the person's actual familiarity with it — which tends to over-supervise low-stakes work by an anxious manager and under-supervise genuinely risky work by a confident one.
Does this framework apply the same way to every team?
The two dimensions apply universally, but where a specific task falls on each axis is specific to your team and situation — the framework is a way of asking the right two questions, not a fixed answer key.