MARK for self-assessment

who this is for

individual contributors and senior PMs who want an honest read on their own judgment shape — not a flattering one.

the problem this solves

most self-assessment frameworks fail in two directions at once: you overstate the things you’re proud of, and you quietly protect the gaps you already know about. the result is a number that feels earned but tells you nothing useful. you walk away confirmed rather than calibrated.

MARK self-assessment is designed to break that pattern — not by adding a harder rubric, but by making the evidence layer impossible to avoid.

how to use MARK for self-assessment, end-to-end

step 1 — take the Skill Scan once, cold

the Skill Scan is your entry point, not your answer. take it once, in one sitting, without going back to change answers. the point is to capture your current self-model before you start interrogating it.

don’t over-interpret the output. the scan is a shaped guess. what it gives you is a starting fingerprint — twelve data points that are claimed, not earned. your job over the next four weeks is to stress-test each one against actual work.

one rule: don’t re-scan for at least 30 days. re-scanning weekly turns the instrument into a mood meter. it measures how you felt about yourself this Tuesday, not how you actually make decisions.

step 2 — walk each competency and name your last call

this is the core exercise. for each of the twelve competencies, ask yourself one question:

what is the last decision I made that clearly tested this competency — and if I scored it honestly against L1–L4, what level was it?

go through each one in order:

• worth — deciding what’s worth building among everything you could ship. your last call: did you source that opportunity yourself, or did it land in your lap from leadership or a customer escalation? sourcing is the test. reactive response is not worth at L3.

• kill — killing your own idea when sunk cost says don’t. your last call: have you killed something in the last six months that you owned and believed in? if the answer is “not really,” you may be holding things longer than the evidence warrants.

• halt — deciding not to build at all. this is the inverse of worth, and often harder. your last call: when did you say “we shouldn’t build anything here, yet” — and hold that position when your team wanted momentum?

• signal — reading conflicting evidence to a decision. your last call: what evidence contradicted your working hypothesis on your last project, and what did you do with it? if you can’t name the contradicting evidence, you may have been filtering rather than reading.

• reframe — spotting and fixing the wrong question before solving it. your last call: when did you push back on the problem definition rather than just solving the problem as handed to you? if the answer is rarely, reframe is likely a developing competency regardless of scan output.

• bet — sizing investment under uncertainty, including reversibility. your last call: did you explicitly reason about reversibility before committing resources? if every decision felt equivalently weighty, you may not be using the one-way-door distinction.

• power — saying no to power, telling the exec the call is wrong. your last call: when did you tell someone above you that their read was wrong — with evidence, not hedging? this one requires you to name a specific instance. vague memory of “pushing back” doesn’t count.

• hold — holding a hard call without folding or bluffing. your last call: when you were pressured to reverse a decision you still believed was right, what happened? did you hold, revise, or fold? a clean hold at L3 means you can name the pressure, name the evidence you re-examined, and name why you kept the call.

• miss — owning a clean miss without deflection. your last call: think of your last significant miss. how did you talk about it in the retro? if your framing started with context (“the data we had at the time…”) before ownership (“I made the wrong call”), you’re looking at developing or competent, not proficient.

• room — reading the room, what people actually mean. your last call: what did someone in your last stakeholder meeting actually want, versus what they said they wanted? if the answer is the same, either you’re incredibly lucky or you stopped listening closely.

• user — reading user truth, who you’re really building for. your last call: your last feature. was it grounded in something you heard directly from users, or was it inferred from usage data or a customer success ticket? direct signal vs. proxied signal is a meaningful level distinction here.

• taste — judging quality you didn’t make, including AI output. your last call: when you reviewed AI-generated work (copy, specs, analysis), what was your actual bar? did you ship the first output or interrogate it? this competency is becoming load-bearing faster than most PMs realize.

step 3 — mark your honest level for each one

write it down. twelve competencies, twelve honest level reads. don’t average them. don’t round up. the fingerprint is meant to be jagged — that’s what makes it useful.

if you find yourself writing L3 across the board, that’s a signal worth pausing on. most experienced PMs have genuine L3 in three or four competencies, credible L2 in several, and real L1s in corners they’ve never had to exercise. a flat profile usually means the exercise is being managed rather than answered.

step 4 — use the 12 chips as a weekly checklist for four weeks

the four-week protocol isn’t arbitrary. one week isn’t long enough to generate new evidence. four weeks gives you enough real decisions to test your self-read against live work.

each week, pick two or three competencies and actively watch for moments where they’re in play. when you’re in a planning meeting and you feel the temptation to add something to the roadmap, that’s halt and worth simultaneously. when you’re in a 1:1 and your skip-level seems to be asking for one thing but clearly wants another, that’s room.

you’re not looking for performance. you’re logging evidence. at the end of week four, revisit your level reads from step 3. some will hold. some will drop. occasionally one will rise — but be strict about what earns a level. moving from L2 to L3 on signal requires that you can point to a decision where conflicting evidence changed your position, not just that you “considered multiple perspectives.”

step 5 — use the Brief as your evidence layer for Worth and Reframe

the Brief is the one surface where your reasoning is fully externalized and scorable. if worth or reframe are the competencies you’re uncertain about, write a Brief response to a relevant prompt. don’t treat it as a test you’re gaming — treat it as a structured way to get your actual thinking out of your head and onto the page where you can look at it.

your written response is your self-assessment for these two competencies in particular. the clarity of your problem framing, the quality of your opportunity reasoning, the sharpness of your restatement of the question — all of it is visible in what you write. the rubric reads what’s there, not what you meant.

what to do in week 1

• take the Skill Scan once, cold, in a single sitting. note the output but don’t treat it as final.
• write down your honest level reads for all twelve competencies using the “last call” exercise from step 2.
• pick two competencies where your scan output and your honest read disagree — those are your first two watch-points.
• write a Brief response to one prompt that tests worth or reframe. read the result. compare the scored levels against your self-reads.

what to expect by week 4

by week four, the fingerprint you started with will have moved — usually downward on two or three competencies, occasionally upward on one. that’s not failure. that’s the exercise working.

what changes is precision. at week one, “I’m probably L3 on Hold” is a feeling. at week four, you can say: “I have two examples of holding a position under executive pressure, and on one of them I folded — which is L2, not L3. the other one held cleanly.” that’s a self-assessment worth acting on.

at a planning meeting, you’ll start using MARK language naturally — not as jargon, but as actual handle words for decisions you’re watching in real time. “this is a halt call” or “I’m reading this as a reframe problem” becomes a way of naming what you’re doing, which is the first step to doing it at a higher level.

the five common pitfalls

1. treating the scan output as the result. the Skill Scan is input, not output. it gives you a starting shape, not a certified level. PMs who stop at the scan and start acting on it are using a self-reported compass as a GPS. the evidence layer is everything — the scan alone is nothing.

2. averaging to a single number. your MARK is not a GPA. a flat 2.8 across all twelve tells you less than knowing you’re L4 on hold and L1 on reframe. the shape is the data. the average erases the shape.

3. sharing self-assessed levels externally. this one matters: self-assessed levels are for your own use. they are not your MARK. the MARK is panel-upheld. if you tell a recruiter you’re L3 on signal based on your own scan, you are making a claim the standard does not uphold. if it comes up, be explicit: “I self-assessed at L3 on signal; I haven’t run a panel Brief yet.”

4. re-scanning to feel better after a hard week. the re-scan reflex usually follows a decision that didn’t go well. you take the scan again hoping the output lands higher. it won’t tell you anything useful — the instrument is too slow to respond to single-event mood. the 30-day minimum isn’t arbitrary. honor it.

5. protecting the gaps you already know about. confirmation bias on strengths is the obvious failure mode. the subtler one is narrative protection — you know your miss is weak because you’ve deflected in the last three retros, but you score yourself L2 instead of L1 because “at least I acknowledged it eventually.” the protocol isn’t designed to shame you. it’s designed to stop you from managing your own story at the cost of clarity.

a worked example

a PM at a mid-stage B2B company had been in the role for three years. she’d taken the Skill Scan twice — both times landing in the proficient band across Map, which covers worth, kill, and halt. her self-read matched: she thought of herself as someone with strong product sense, good at knowing what to build.

she ran the four-week protocol as part of a MARK self-assessment cycle. the exercise that changed everything was step 2 on worth: what is the last decision I made that clearly tested this competency, and how would I score it honestly under L1–L4?

she went back through her last three quarters of roadmap decisions. what she found was consistent: every item on the roadmap had a clear external origin. one came from a major customer threatening churn. three came from a competitor’s feature release that made leadership nervous. two came from a sales team request. one came from an engineering push to retire technical debt. she had prioritized all of them well — real L3 prioritization work, good signal reading, solid framing. but she had discovered none of them.

the worth anchor at L3 requires sourcing opportunities, not just ranking them. she was scoring as a strong L2 — excellent at evaluating what was handed to her, genuinely capable at the prioritization layer, but not yet running her own opportunity discovery process. every “decision” was reactive. that’s not a failure of competence — it’s a failure of posture.

the fix wasn’t more scanning. it wasn’t even a skills gap in the traditional sense. it was building a real discovery process into her work week — customer conversations without an agenda, time blocked for market scanning that wasn’t triggered by a competitor move, deliberate attempts to surface problems before they surfaced themselves.

three months later, the first item she generated proactively made it onto the roadmap. that’s the shift from L2 to L3 on worth. it has nothing to do with intelligence or effort. it’s a structural change in how she operated.

citation

PL Standard v3.1 · using MARK for self-assessment