The 'Don't Try to Change People' EM Playbook: Supporting Six Subordinate Types in Japanese IT

• Intended readers: EMs and team leads carrying heavy individual-contributor (IC) loads, senior engineers stuck on people development, HR and organization-development staff
• Assumed knowledge: General team-operations vocabulary (1on1s, task delegation, performance reviews)
• Reading time: Full read about 30 minutes / key points about 13 minutes

Overview

The previous article, Engineering Management as a Punishment Game: The Three Separations Japanese IT Companies Need, diagnosed why the EM role becomes a “punishment game” in Japanese IT: roughly 96% (precisely 95.8%) of managers are also acting as individual contributors¹, only 21.4% of employees actively want to become managers², and engagement sits at 6%³. The root cause is forcing one EM to absorb AI management, people management, IC work, professional people-management work, and the legal/psychological/values layer all at once. (“Punishment Game Structure” here refers to a system that nominally rewards a role while structurally penalizing whoever takes it.)

This article is the sequel and focuses on what to actually do in tomorrow’s 1on1.

The conclusion comes up front, in reverse. This is not a playbook for developing your reports. It is a playbook for giving up on “developing” or “changing” them.

The central observation lands at the same place from three independent routes.

First, the labor invested in trying to change a person is usually wasted. Personality, motivation, and values are the individual’s self-driven domain. Big Five research⁴, the Stages of Change model⁵, and Self-Determination Theory (SDT)⁶ all converge: rewriting a person from the outside is a poor return on investment, even on the evidence.

Second, once the verb becomes “change this person,” the available tools narrow to pressure, threat, and coercion, which almost automatically pushes you across labor-law red lines. Implied threats of dismissal, pressuring resignation, mandating after-hours study, attacking personality in performance reviews, intervening in someone’s values—every classic EM mistake reads as a move to “change” something.

Third, every EM verb that stays inside legal lines points the same direction: don’t change, support self-direction—observe, document, communicate, connect, listen, wait, assign, organize. This is not coincidence. Labor law is designed around “protecting individual self-determination from external pressure,” especially in jurisdictions like Japan with strong dismissal restrictions.

Given that, the situations an EM faces fall into three cases.

Case	Situation	What the EM can do	Section
(1)	The report is trying to change, and the direction looks reasonable to the EM	Don’t get in the way; arrange conditions and support (type-by-type Do’s and Don’ts)	§4
(2)	The report is trying to change, but the direction looks wrong to the EM	Sort out what the “wrongness” actually is; climb a ladder from light to heavy verbs	§5
(3)	The report is not trying to change at all	Stop trying to change them. Re-cast the role, escalate, or withdraw	§6

The last key to escaping the punishment game is letting go of the guilt that case (3) feels like “abandoning development.” That self-imposed burden—”I couldn’t develop this person”—is the actual fuel of the punishment game.

A specific trap to watch for: mistaking case (2) for “Type D wearing a Type A mask” (preferences and values dressed up as ‘wrongness’) pulls you into the §1 gravity well of pressure-based intervention. That distinction is the core question of §5.

With that framing, the article runs in this order. §1 shows the structural reason “trying to change people” hits legal red lines. §2 lays out the EM’s boundaries (legal, role, capability). §3 reframes the six-type taxonomy as “a lens to avoid pointing support in the wrong direction,” not “a prescription for intervention.” §4 covers type-by-type Do’s and Don’ts (case (1)). §5 covers how to sort out a “wrong direction” feeling (case (2)). §6 covers reports who are not trying to change (case (3)). §7 covers operational design. §8 lists caveats and limits.

The six-type taxonomy (Insource 2025⁷), the Will/Skill matrix⁸, and Situational Leadership theory⁹ are practitioner frameworks, not psychometrically validated taxonomies. Treat them as scaffolding to make the first move less likely to miss, not as diagnostic instruments.

1. Why “Trying to Change People” Hits Legal Red Lines

This is the spine of the article. The moment an EM steps in to “change” a report, the only tools left are pressure, threat, and coercion—and most of them cross labor-law red lines. This correspondence is what structurally constrains an EM’s verbs.

It helps to notice that “trying to change someone” has two faces.

• (i) Change them through pressure: motivation-style intervention—”work harder,” “show some enthusiasm,” “I won’t rate you well unless you produce.”
• (ii) Correct what’s wrong: direction-style intervention—”your approach is wrong,” “that judgment call is mistaken.”

(i) reads as obviously off-limits. (ii) looks like “legitimate work feedback” at first glance, but once it starts repeating in ways that intrude on a person’s judgment, choices, and values, it falls into the same gravity well as (i). Well-intentioned verbs—”let me correct you,” “let me teach you”—erode the other person’s autonomy and eventually convert into pressure-based intervention. §5 unpacks this in detail. The §1 point is just: both belong to the “trying to change” category, and both carry the same legal and psychological risk profile.

1-1. Common EM mistakes are all “change-the-person” verbs

Under Japanese labor law and case law, there are clear lines around what an EM can do in day-to-day operations¹⁰¹¹¹². Re-stated as “what is this verb trying to change?”, the correspondence becomes obvious.

Common mistake	What it’s trying to change	Legal problem
“Produce results or you’re fired”-style threats	Behavior, by fear	Doctrine of Abuse of the Right to Dismiss (Labor Contract Act Art. 16)10; coercive resignation pressure
Unilateral hints about transfer or demotion	Environment, by threat	Abuse of HR authority; violations of work rules
“Maybe this isn’t for you—why don’t you quit?”	The person’s existence (= exclusion)	Coercive resignation pressure (Shimonoseki Commercial High School case, Supreme Court 1980)11
Mandating study or training outside working hours	Skills, by command	Working-hours regulation (Labor Standards Act Art. 32 / 37)
Treating mental-health issues as “lack of grit / weakness”	State of mind, by willpower-talk	Breach of the duty of care (Labor Contract Act Art. 5)12
Personal attacks or yelling in performance reviews	Personality, by pressure	Power Harassment Prevention Act
Reviewing or penalizing values, ideology, or beliefs	Thought and conscience, by evaluation	Article 19 of the Constitution of Japan (freedom of thought and conscience); Labor Standards Act Art. 3
Implicit obstruction of leave (medical, parental)	The act of taking leave, by pressure	Child Care and Family Care Leave Act / Labor Standards Act
Unilaterally “ruling” on harassment claims	Interpretation of an incident, by fiat	Determinations belong to HR/legal, not the EM

These are areas that require case-by-case judgment, and an EM should not be making the call alone. The correct move at the first whiff of doubt is to hand off to HR, legal, or the occupational physician.

1-2. The legally safe verbs all line up with “don’t change, support”

Conversely, the actions that sit clearly inside an individual EM’s authority are also well-defined.

• Adjusting scope and assignments: distributing tasks, setting priorities, changing ticket granularity
• Building agreement in 1on1s: clarifying expectations, checking progress, aligning on goals
• Progress management, code review, feedback: evaluating behavior and outcomes (not personality or character)
• Building fact-based evaluation records: documenting observed behavior and its impact, not vague impressions
• Communicating improvement requests: separating fact, impact, and expectation
• Bridging escalations: connecting to transfers, leave, occupational physicians, HR (the EM connects; someone else decides)
• Designing team communication: meeting cadence, reporting cycles, documentation culture

Pulled out, the verbs are: observe, document, communicate, connect, listen, wait, assign, organize. Every one of them lines up with “don’t change, support self-direction.”

1-3. Why this correspondence isn’t an accident

There is a structural reason “trying to change → crosses legal lines” and “supporting → stays legally safe” map cleanly onto each other.

flowchart TB
  CHANGE["Trying to change<br>the person"]
  TOOLS["Tools narrow to<br>pressure, threat, coercion, command"]
  ILLEGAL["Labor-law red lines<br>(dismissal, coerced resignation,<br>harassment, breach of duty of care,<br>intrusion on thought/conscience)"]
  SUPPORT["Supporting<br>self-direction"]
  VERBS["Observe, document, communicate,<br>connect, listen, wait, assign, organize"]
  LEGAL["Legally safe zone"]
  CHANGE --> TOOLS --> ILLEGAL
  SUPPORT --> VERBS --> LEGAL

Three independent routes lead to the same conclusion.

Legal route: Labor law is built around protecting individual self-determination from external pressure. Labor Contract Act Article 16 (Doctrine of Abuse of the Right to Dismiss)¹⁰, Labor Contract Act Article 5 (Duty of Care)¹², Labor Standards Act Article 3 (no discrimination by belief, etc.), Article 19 of the Constitution of Japan (freedom of thought and conscience)—every one of them shields individuals from “external pressure to change them.” The same protective intent shows up in many jurisdictions, even when the specific statutes differ.

Psychological route: SDT⁶’s autonomy need. Gagné & Deci (2005, DOI: 10.1002/job.322) consolidate experimental SDT findings showing that threats, deadlines, directive evaluations, and imposed goals reduce intrinsic motivation⁶. “Work harder” pressure can break the motivation of even the people who were already running on intrinsic drive.

Empirical route: Big Five research⁴ shows adult personality traits are largely stable; external intervention rarely produces large shifts. The Stages of Change model⁵ shows that for someone in the precontemplation stage (no intent to change), external pushing simply doesn’t work. Effort spent on “I will change them” is, on the evidence, a poor investment.

Three independent routes converge on the same conclusion: “don’t try to change people” is correct legally, psychologically, and empirically. That is the foundation under the rest of this playbook.

2. The EM’s Boundaries: Legal, Role, Capability

§1 covered the legal boundary. There is a second boundary an EM must not cross. It is legally safe but operationally unsustainable for one person—the role and capability boundary.

Type of boundary	Character	What happens if you cross it
(a) Legal / institutional boundary	Strict; lines drawn by labor law, case law, work rules	Legal risk (coercive-resignation finding, harassment finding, breach of duty of care); damage to both the report and the organization
(b) Role / capability boundary	Loose; not illegal, but impossible to complete alone	Overload spiral; one person carrying it all is what makes the role a punishment game

2-1. Role and capability boundary—what one person can’t carry

Even where nothing is illegal, there are areas no single EM can fully own. Treating them as “my responsibility” indefinitely is what keeps the punishment-game dynamic running. In general management literature this maps to the distinction between Span of Control and Span of Influence¹³; in the previous article it was the 5-factor matrix (capability, situation, career view, environment, motivation).

Domain	What the EM can carry	What one EM cannot carry alone
Motivation	Articulating expectations, returning competence-feedback, naming achievement	Fundamental rewiring of intrinsic motivation (depends on the person’s life view and values)
Skill development	Providing on-the-job experience, reviewing, recommending	Career re-design, fundamental reorientation of expertise
Mental health	Spotting signals, bridging to occupational physicians / EAP	Diagnosis, treatment, recovery planning (medical domain)
Career design	Short-term role assignment, evaluation records	Long-term career guarantees, executing transfers (executive / HR authority)
Compensation / evaluation	Evaluation records, recommendations, input to HR	Compensation system design, demotion/promotion decisions (executive / HR authority)
Team culture	Team rules, operational design, maintaining psychological safety	Org culture, evaluation system, HR policy (executive / HR domain)

Stepping into the “can’t carry alone” zone is legally safe but operationally fatal—the EM’s overload doesn’t stop. The legal boundary in §1 has clear case-law-drawn lines; this one is fuzzier. The only defense is the EM’s own habit of asking, “Is this actually mine to hold?”

Three diagnostic questions:

1. Authority: Is this within my decision rights? (Am I stepping into HR, executive, or medical territory?)
2. Time: At a 96% playing-manager rate¹, do I actually have the resource to put against this?
3. Expertise: Is this my professional domain? (Mental-health care and legal judgment belong to specialists.)

If even one answer is “no,” that is a signal to hand it off rather than hold it.

2-2. Why EMs cross the boundary anyway

Three structural reasons.

First, the “three separations” from the previous article are not institutionalized, so org design defaults to an EM personally owning professional people-management, legal judgment, and values intervention.

Second, the 96% playing-manager rate¹ keeps cognitive load saturated, and “I’ll just do it myself” looks like the shortest path even when it isn’t.

Third, in Japanese management culture, “caring well = intervening deeply” is a common confusion. Stepping into values, personality, and ideology gets read as “passion” and rewarded. Legally this is a minefield; operationally it’s a furnace for overload.

“The EM’s job is not to step in deeper but to hand off well.” That mindset shift is a prerequisite before type-by-type tactics make any sense.

3. What the Six-Type Lens Is Actually For

Once §1 and §2 are in place, the EM’s job collapses to “support,” not “intervention.” So what is type classification for? A lens to avoid pointing support in the wrong direction.

Even when supporting “a report who is trying to change,” what to support varies by person. The same SL-style “Selling” looks completely different for an Energy-Conserving (Pragmatist) report (“explain the reward and the meaning”) versus a Poor-at-Planning report (“co-decompose the task with them”). Type lenses exist only to reduce direction-of-support mismatches.

3-1. Will/Skill complemented by SL

A coarse two-axis cut of a report’s state maps onto Situational Leadership’s four styles⁹.

flowchart TB
  Q["Report's current state"]
  Q --> Q1["Skill: high / low?"]
  Q --> Q2["Will: high / low?"]
  Q1 & Q2 --> M["Four-quadrant matrix"]
  M --> A["High Skill, High Will<br>→ Delegating"]
  M --> B["High Skill, Low Will<br>→ Participating"]
  M --> C["Low Skill, High Will<br>→ Selling"]
  M --> D["Low Skill, Low Will<br>→ Telling"]

Hersey & Blanchard’s 1977 SL theory⁹ cycles Telling → Selling → Participating → Delegating across developmental stages D1–D4. The Will/Skill matrix⁸ follows the same structure: vary the depth of involvement by quadrant.

The key reframing: SL styles are not “intervention intensity.” They are forms of support. Telling is not “teaching to change” but “handing a beginner the information they need.” Delegating is not “leaving them alone” but “not getting in the way of someone who is already running.”

3-2. Six-type self-diagnostic checklist

Practitioner frameworks often split report behavior into six types⁷. Read each type as having a different “direction the report is trying to change” and different “conditions to support.”

Goal-Achievement Oriented (Self-Starter)—energized by results. Pulls hard problems toward themselves. Active in strategy discussions. High Skill × High Will, SL: Delegating.

Risk-Averse (Cautious)—”don’t make mistakes” first. Lots of pre-checking and confirmation. Patient, methodical. Mid-to-high Skill × Mid Will, SL: Participating.

Energy-Conserving (Pragmatist / Cost-Performance Oriented)—avoids waste, calculates so the sparks don’t land on them. Often misread as “no motivation,” but is actually watching cost-performance. High Skill × Low Will, SL: Selling.

Poor at Planning (Eager but Unstructured)—willing, but weak at prioritization and task decomposition. Reporting cadence is off-beat. Low Skill × High Will, SL: Selling / Telling.

Indecisive (Hesitant)—keeps deferring even when the inputs are present. Underneath: aversion to being responsible for failure. Mid Skill × Low Will (low confidence), SL: Participating.

Stubborn / Self-Righteous (Independent but Rigid)—high confidence in their own “rightness.” On individually-bounded tasks, ships well in the short term. Direct confrontation tangles the relationship fast. High Skill × High Will (idiosyncratic), SL: Delegating + careful questioning.

3-3. Types are not fixed labels

The same person shifts type by task or by life phase. A Goal-Achievement Oriented engineer entering an unfamiliar domain shows Risk-Averse behavior the next morning—that kind of phase transition is normal. The right move is less “diagnose the type” and more “check which motivation structure they’re running on right now” in a 1on1.

flowchart TB
  P["Report's current behavior"]
  P --> O1["Which results energize them?"]
  P --> O2["What are they trying to avoid?"]
  P --> O3["What instruction granularity feels right?"]
  O1 & O2 & O3 --> T["Tentative six-type placement"]
  T --> ACT["First support move"]
  ACT --> R["Observe response → update hypothesis"]

Notice that the “Low Skill × Low Will” quadrant is missing from the six types. That’s not an accident. The taxonomy assumes “people who are actually on the team and contributing something.” A persistent Low Skill × Low Will state is outside §4’s scope and belongs in §6 (reports who are not trying to change).

4. Type-by-Type Do’s and Don’ts: Supporting “Reports Who Are Trying to Change”

This section is for case (1) in the overview (the report is trying to change and the direction looks reasonable to the EM). Inside the §1 and §2 boundaries, against the §3 six types, here are the concrete support verbs. The Do’s are oriented toward “don’t block self-direction; arrange conditions.” The Don’ts are oriented toward “support that misreads the type and whiffs.”

Case (2) (direction looks wrong) and case (3) (not trying to change) are §5 and §6. This section assumes the direction is right; only the precision of support is in question.

4-1. Goal-Achievement Oriented (D4 / Self-Starter)

• 🟢 Do: agree on direction and goal, then let go / specifically name what they achieved when you give feedback
• 🔴 Don’t: micromanage, over-decompose progress, hammer “just-checking” pings
• 🚫 Boundary: the report’s intrinsic motivation source (“the feel of achievement”) is not for the EM to supply—stay in the role of striker, not fuel tank

1on1 opener:

“Of next quarter’s themes, which one looks most interesting to you? Once you pick, I’ll spell out the requirements and constraints, and the how is yours.”

Achievement is the fuel, so observing and naming the achievement is non-negotiable. “Glad it shipped” is too thin. Naming what was decisive turns into a catalyst for the next stretch. In Gagné & Deci⁶’s framing, this is “competence-feedback,” not “external reward”—one of the few interventions that reinforces intrinsic motivation without eroding it.

4-2. Risk-Averse (D2–D3 / Cautious)

• 🟢 Do: standardize the three-step “appreciate → check progress → next small goal” / make the safety net visible
• 🔴 Don’t: rush them, dismiss the risk they raised, label it “overthinking”
• 🚫 Boundary: caution is closer to a personality trait than a habit; you can’t forcibly remold it—deploy it well instead

1on1 opener:

“Thanks for being thorough on this. Let’s set the next checkpoint here. What might trip you up before then?”

Risk-aversion is often a role-level strength, not a weakness—pre-release quality gates, production operations, contract closure. Rather than retrofit the person, find roles where the trait is the asset. Better return on investment.

4-3. Energy-Conserving (Pragmatist) (High Skill × Low Will / Rational)

• 🟢 Do: pair “why now” with “what they get out of it” in every ask / welcome rational pushback
• 🔴 Don’t: emotional appeals (“show some motivation”), abstract goals (“contribute more”), pushed-down busywork
• 🚫 Boundary: redesigning compensation systems and rewriting values are out of EM authority (executive / HR)—escalate when you hit the “system wall”

1on1 opener:

“This task has a slightly elevated priority. Here’s how it shows up in your evaluation. If there’s something we can stop in exchange, name it.”

Energy-Conservers aren’t unmotivated; they’re watching cost-performance. Build a “return that justifies the cost” and the high-Skill side flows straight into utilization. Conversely, in orgs where reward and evaluation are decoupled from output, this type quietly leaves first.

4-4. Poor at Planning (Will High × Skill Low / Structure-Support)

• 🟢 Do: scaffold with structure (templates, Jira/Linear ticket granularity, reporting cadence) / ask “what’s blocking you?”
• 🔴 Don’t: write it off as “lack of effort,” ignore them, dump a large work item without scaffolding
• 🚫 Boundary: skill acquisition takes time and reps—off-hours bootcamps are a labor-law risk (Labor Standards Act)

1on1 opener:

“I’d line up the priorities like this first—anything off about that? Also, if you’re stuck on something, tell me earlier rather than later.”

Will is plenty; a few rounds of structured pairing produce the success experiences that tip them into self-direction. Skip the structure and they look like “all motivation, no output,” which damages both their evaluation and their self-image. In the AI era, ToDo decomposition and prioritization are also areas that pair well with AI assistants.

4-5. Indecisive (Low Will / Push-from-Behind)

• 🟢 Do: narrow to 2–3 options and let them choose / give them decision reps in areas where mistakes are recoverable
• 🔴 Don’t: leave them alone, hand them total freedom, frame decisions as a “responsibility” they should fear
• 🚫 Boundary: confidence rebuilds slowly / a company-wide “tolerate failure” culture is not in one EM’s gift—stay focused on small safe zones inside the team

1on1 opener:

“Plan A or Plan B—which would you prioritize? If we miss, we can roll back to here.”

Indecision is usually rooted in “I don’t want to be on the hook for the failure.” That is exactly why psychological safety matters—Edmondson¹⁴ shows that on teams where people believe interpersonal risk is safe, frequency of decisions, speech, and experiments rises. What an individual EM can do is keep explicitly stating, “the structure absorbs the cost of failure.”

4-6. Stubborn / Self-Righteous (High Skill × High Will, Idiosyncratic / Independent)

• 🟢 Do: open the lens with “what would this look like from the other side?” questions / build roles where their depth pays off
• 🔴 Don’t: head-on confrontation, emotional argument, public negation
• 🚫 Boundary: forcibly changing values or beliefs is legally out of bounds (Article 19 of the Constitution of Japan; freedom of thought and conscience). If they don’t fit collaboration requirements, work with HR on placement, not on “fixing” them

1on1 opener:

“Your call holds up technically. I just want to hear how this looks from the operations team’s side.”

Stubborn types ship in the short term, but costs accumulate where collaboration is required. The realistic move is “deploy them well,” not “fix them.” If you start to feel “the direction itself is wrong,” step into the §5 sort, and if the report still won’t move and is jamming the team, escalate per §6.

5. When the Direction Itself Looks Wrong: Sorting Out Case (2)

§4 was for case (1): trying to change, direction looks fine. This section is case (2): trying to change, but the direction looks wrong to the EM.

This is the most trap-rich territory in the article. The well-intentioned verb “let me correct this” is the easiest entry point into the §1 gravity well of trying to change a person. “Let me correct you” / “let me teach you” sounds educational, but repeated against a person’s judgment, choices, and values, it converts into pressure, command, and coercion.

The sort runs in three steps: (1) classify what kind of “wrongness” this is; (2) detect “Type D wearing a Type A mask”; (3) climb a ladder from light to heavy verbs.

5-1. Four kinds of “wrong”

The first question to ask the moment “they’re going the wrong way” pops into your head is: wrong how? The classification entirely determines what verbs are available.

Kind	What it is	Objectivity	What the EM can do
A. Factual error in the work	Design is technically broken; estimates are weakly grounded; will fail in production; data interpretation is wrong	High (reproducible, third-party verifiable)	Communicate based on facts. Inside authority; inside support
B. Process / team impact	Skipping reviews, broken status updates, blocking other members, violating Definition of Done	Medium (observable)	Address the rule and structure, not the person. Update process, agree on contracts, tighten team guidelines
C. Their career choice	“If they keep going this way, they won’t be rated well / their growth will plateau / their market value drops”—as the EM sees it	Low (EM subjectivity is in the loop)	Acknowledge it’s only your prediction; offer information; return the decision to them
D. Mismatch with the EM’s values or preferences	“I wish they were more growth-oriented” / “I want more energy” / “I wouldn’t do it that way”	Near zero (this is on the EM’s side)	Don’t intervene. This is the dead center of the §1 gravity well

5-2. Spotting “Type D wearing a Type A mask”

The most common practical trap: expressing a Type-D motive in Type-A or Type-C language.

• “Your design is wrong” (= it’s actually just different from my style / possibly D)
• “Your career choice is wrong” (= I’m scoring it on my values / D dressed up as C)
• “That approach is bad” (= it just diverges from existing convention; technically defensible / possibly D)

The instant this happens, you are inside the §1 “trying to change → pressure-based → legal-line crossing” gravity well. The well-intentioned verbs—”let me correct you,” “let me teach you”—are exactly the easiest way to slide in without noticing.

Four self-questions to sort it:

1. Can the “wrongness” be described in facts a third party can reproduce? → Yes, then A.
2. Does it violate the team’s published rules or agreements? → Yes, then B.
3. Is it a prediction about their future career, where I might well be wrong? → Yes, then C (information-only).
4. Is this about “I wouldn’t do it that way”—preference and values? → Yes, then D (don’t intervene).

If 1 and 2 don’t immediately answer “yes,” there is a high chance this lives in C or D territory. You need the discipline to admit, “I do feel something is off, but that doesn’t make it ‘wrong.’”

5-3. The light-to-heavy verb ladder

Even when you’ve classified it as A or B, don’t jump to the “correct it” verb. Climb from light verbs first: observe → ask → share fact → explain impact → present options.

flowchart TB
  L1["1. Observe<br>'In this implementation, when X runs, I think it goes like this—does that match?'"]
  L2["2. Ask<br>'How are you imagining the behavior in the Y case?'"]
  L3["3. Share fact<br>'In production we've seen Z happen with a similar pattern before'"]
  L4["4. Explain impact<br>'If we keep going this way, post-release costs are likely to look like this'"]
  L5["5. Present options<br>'Plan A or Plan B—I want to hear your call'"]
  STOP["Proceed in the direction the report chose<br>→ run a learning cycle from the result"]
  L1 --> L2 --> L3 --> L4 --> L5 --> STOP

Each step invades autonomy less the lighter it is. Steps 1–2 alone are often enough for the report to notice and self-correct. Conversely, starting at step 4 or 5 collapses straight into pressure-based intervention.

The critical rule: even after you climb to step 5, if the report decides to go in their original direction, respect it. The EM’s job is no longer “stop them” but “run a learning cycle together when results are in.” Removing the failure removes the only mechanism by which their self-direction grows.

5-4. Exception: cases where you do have to stop them

That said, there are situations where the EM has legitimate grounds for stronger intervention mid-ladder. In any of the following, start light, but the EM has the authority (or the obligation to immediately involve someone above) to issue a stop.

Exception case	Response
Irreversible operational risk	Destroying production data, security incident, contract breach, regulatory violation
Impact on interpersonal safety	Harassment, behavior that destroys psychological safety, designs likely to cause mental-health harm
Team-wide blockage	One person’s call has been blocking other members’ work for an extended period
Material deviation from commitments	Significant overrun against agreed effort, budget, or deadline

These go beyond “respecting the report’s call.” Stopping, sending back, and escalating are EM responsibility. The crucial framing: even here, the verb is “showing the operational red line,” not “changing this person.” Communicate it as “you can’t cross this line,” not “I want to change you,” and as an organizational rule, not a personal one.

5-5. If they won’t move past 5-3, hand off to §6

If you’ve climbed to step 5 in 5-3, none of the 5-4 exceptions apply, and the report still isn’t moving—you’ve likely crossed into case (3): not trying to change. Apply the §6 escalation criteria (same issue not improved over three consecutive cycles / six-month stagnation / mental-health signals / legal points) and shift to role re-cast, escalation, or withdrawal.

The on-the-ground flow is “case (2) (wrong direction) → case (3) (not trying to change)” in that order.

6. When You Decide They’re Not Trying to Change

§4 (case (1) Do’s) and §5 (case (2) sort) both presuppose “a report who is trying to change.” Specifically: if you climbed the 5-3 ladder to step 5, no 5-4 exception applies, and the report still doesn’t move, you are likely now in case (3)—the report is in precontemplation⁵, “I don’t see a problem / I don’t intend to change.” This section covers that.

The headline: the right move is to stop trying to change them. §1 already showed why: stepping in here narrows your tools to pressure, threat, and coercion, and legal risk wakes up. Stages of Change⁵ research adds: direct intervention against people in precontemplation does not work.

6-1. Three options

Option	What it is
Re-cast the role	The current task or role may not match the report’s self-direction. Re-evaluate Person-Job Fit. Use the previous article’s 5-factor matrix (capability / situation / career view / environment / motivation) to identify which factor is dominating
Escalation	If any of “same issue not improved over three consecutive cycles,” “six-month stagnation,” “mental-health signals,” or “legal points” appears, hand off to HR, occupational physician, legal
Withdrawal (cost containment)	If the above two don’t move it, stop additional EM personal investment. This is not abandonment of development; this is the correct withdrawal—returning the call to HR and executive

6-2. Handling the “I couldn’t develop them” guilt

The crucial part of §6 is letting go of the guilt that withdrawal feels like personal failure.

As the previous article showed, the fuel of the punishment-game structure is the EM’s “I couldn’t develop this person” self-blame. Japanese management culture confuses “caring well = intervening deeply,” so withdrawal carries strong psychological resistance.

What to remember here is the §1 structure: keep trying to change them and you eventually cross legal red lines. Withdrawal isn’t giving up; it is a defensive action protecting both the EM and the report from legal and psychological risk, and it correctly returns the issue to executive and HR for organizational handling.

Reword “I couldn’t develop them” to “this exceeded what my authority, time, and expertise can address.” This is the operational implementation of the previous article’s “three separations” (especially separation 2: IC work vs. professional people-management).

flowchart TB
  S1["1on1: confirm motivation,<br>weak spots, preferred delegation"]
  S2["Tentative type placement +<br>support direction"]
  S3["First support move (§4)"]
  S4["Observe response, log effect"]
  S5{"Trend toward improvement?"}
  E1["Hold hypothesis,<br>continue support"]
  E2["Update hypothesis,<br>try different type lens"]
  E3{"Meets §6 criteria?"}
  ESC["Re-cast role<br>or escalate<br>or withdraw"]
  S1 --> S2 --> S3 --> S4 --> S5
  S5 -- "Yes" --> E1 --> S1
  S5 -- "No" --> E2 --> E3
  E3 -- "Yes" --> ESC
  E3 -- "No" --> S2

6-3. What the EM should be asking from their boss, HR, and executives: update the evaluation criteria

Letting go of the guilt individually is not enough. As long as the EM’s manager, HR, and executives score correct case-(3) calls as “abandonment” or “doesn’t care about people,” this playbook works only at the individual level. The EM protects themselves and their report from legal and psychological risk, but takes the hit on their own evaluation. That is exactly the punishment-game structure from the previous article. Without an evaluation update, EMs eventually stop adopting the playbook.

What is needed is organizational updating of evaluation criteria. Replace the legacy intervention-centric criteria with withdrawal- and support-centric criteria.

Legacy criteria (interventionist)	New criteria (withdrawal- and support-oriented)
“Never gave up”	“Withdrew or escalated at the right time”
“Grew everyone”	“Didn’t block those trying to change; placed the rest by role-fit”
“Cared deeply about people”	“Didn’t cross legal or psychological boundaries”
“Lifted reports through sheer drive”	“Didn’t intrude on self-direction; returned decisions to the report”
“Kept attrition low” (at any cost)	“Avoided mental-health incidents, leaves, and legal trouble”
“Development outcomes” (number of people who grew)	“Role-matching outcomes” (number of placements that produced results)

Withdrawal judgment is itself the EM’s professional skill, the same way “knowing when more surgery would harm the patient” is part of the surgeon’s professional skill. An organization that scores this as “giving up” is structurally forcing the EM to cross legal lines.

Three concrete asks of the organization:

1. Add “appropriate withdrawal / escalation calls” as an explicit evaluation item: How many times, at what point, did this EM hand off to HR, occupational physician, legal—and did that handoff avoid legal or psychological risk? Make it part of performance review.
2. Add “incidence of legal trouble, mental-health incidents, and leaves” to org KPIs: Tracking only “low attrition” induces unhealthy retention pressure and pressure-based intervention. Track “the rate of legally and psychologically healthy exits and transfers” instead.
3. Score “outcomes from role-matching” in parallel with “development outcomes”: Outcomes from placing someone in a fit role should be rated equally—or higher—than outcomes from “trying to change someone who didn’t change.”

This is not something an individual EM can implement alone. It is a system-design matter for HR and executives. If any reader of this article is in those seats, this is the place to start the evaluation review. Individual EMs can still actively negotiate with their own manager: in your 1on1, make “how withdrawal calls are evaluated” an explicit topic. Stay silent and you’ll keep being measured on the legacy axis.

Without an evaluation update, EMs who execute this playbook personally take the hit for doing the right thing—a reproduction of the punishment-game structure the organization should be breaking, not perpetuating.

7. Operational Design to Run the Cycle

To keep type-by-type support from being a one-off 1on1 stunt, build it into operational form.

7-1. 1on1 template

Fill these in over the first three sessions:

1. Motivation check: “Most energizing recent piece of work?” “Heaviest piece of recent work?”
2. Weak-spot check: “What’s eating too much of your time right now?” “What do you want to stop or reduce?”
3. Delegation-style check: “Is the current granularity of instructions right? More detailed? Just the big picture?”

The third item is the SL-theory-critical one: letting the report describe their own preferred style sharpens the EM’s style choice. Re-asking every 3–6 months reduces the chance of missing a phase transition.

7-2. Tooling

• Task management: in Jira / Linear / Asana, vary ticket granularity by type (small, decomposed child tickets for Poor-at-Planning; big parent tickets for Goal-Achievement)
• Observation notes: in Notion or similar, log type hypothesis, observed responses, and support effects as EM-private notes. Do not share with the report or the team (avoids labeling and ossification)
• AI assistants: 1on1 summarization, mining motivation patterns from past comments—use AI to reduce the EM’s cognitive load side, not to “diagnose” reports

7-3. HR escalation criteria (linked to the previous article)

To keep the EM from bottling everything up, keep mechanical triggers ready to fire.

Signal	Escalate to
Same issue not improved over three consecutive cycles	Share with HR; bring in third-party perspective
Six-month stagnation; type hypotheses keep missing	Discuss transfer (re-evaluate Person-Job Fit)
Mental-health signals (sleep, lateness, increased away-from-desk time, expression changes)	Loop in occupational physician / EAP
Legal points (harassment, contract terms, leave)	HR / legal

This is the operational implementation of the previous article’s separations 2 (IC work vs. professional people-management) and 3 (legal / psychology / values).

7-4. Bias check on yourself

Quarterly, audit your own response patterns.

• Do I lean on Delegating? (Am I leaving Poor-at-Planning reports unsupported?)
• Do I lean on Telling? (Am I eroding the autonomy of Goal-Achievement reports?)
• Which type do I personally find hardest to handle, and have my moves toward them gone monotone?
• Am I noticing when I shift into “trying to change them”? (Pressure, command, coercion sneaking in?)

The EM is also one of the six types, with bias toward leniency on similar types and severity on dissimilar types. Including yourself in the observation surface is what keeps this operation working over the long term.

8. Caveats, Reservations, Limits

This playbook is a strong tool, and overconfidence causes accidents. Stating limits explicitly:

Risk of label ossification. Behavior is dynamic; the same person varies by project, health, life phase. The instant “she’s an Energy-Conserver” calcifies as a label, the EM’s observation stops—the most common failure mode. Treat type as “current hypothesis” and keep updating with observation and dialogue.

The six-type taxonomy itself is a practitioner framework, not a peer-reviewed taxonomy. It originates in a practitioner article from Insource⁷; it has not been validated through psychometric construct-validity testing or factor analysis. The Will/Skill matrix⁸ and SL theory⁹ are widely used but academically contested—SL theory in particular has faced critique on empirical grounds since the 1980s¹⁵. Treat these as tools that sit between “useful for practitioners” and “scientifically validated taxonomy.”

Structural limits of Japanese firms. Inside membership-based employment (where employees are hired into “the company” rather than “a job,” generally with strong dismissal restrictions), seniority-driven systems, and Japan’s labor-law context, investing equal effort in every type is unrealistic. As the previous article argued, the end goal at the org level is “making HR escalation and specialist tracks real,” and this playbook is interim operation until that is in place.

Lack of quantitative effect data. As of 2026, no public study specific to Japanese IT firms quantifies how six-type operation affects attrition or productivity. Overseas SL/coaching-intervention research exists, but applying it directly to the Japanese organizational context requires local validation. Run “did this work or not” as a measurement on your own team.

Legal caveats. When considering transfers, PIPs, or demotions, confirm with HR against current case law, work rules, and the Labor Contract Act. The “type hypothesis” in the playbook must not be used as the basis for performance records or legal determinations.

To avoid over-generalization and label ossification, read this framework as “scaffolding so the first support move is less likely to miss.”

Summary

For an EM to absorb overload under the punishment-game structure, the starting move is letting go of the illusion of “developing / changing” people. The points of this article:

1. Once you step in to “change a person,” your tools narrow to pressure, threat, and coercion, and you almost automatically cross labor-law red lines. Three independent routes—legal, psychological, empirical—converge on the same conclusion. Both “change them through pressure” and “correct what’s wrong” fall into the same gravity well.
2. The EM’s boundary is two-layered: legal boundary (common mistakes) and role/capability boundary (what one person can’t carry). With both in place, the EM’s verbs converge to “observe, document, communicate, connect, listen, wait, assign, organize.”
3. Situations break into three cases: (1) trying to change, direction OK; (2) trying to change, direction looks wrong; (3) not trying to change. §4, §5, §6 respectively.
4. The six-type taxonomy is not a “prescription for intervention” but a lens for not pointing support in the wrong direction. When the type hypothesis misses, update it.
5. The guilt over case (3) is the punishment game’s fuel. Reword “I couldn’t develop them” to “this exceeded what my authority, time, and expertise can address.”
6. An individual-EM playbook alone cannot break the punishment-game structure. Managers, HR, and executives have responsibility to update evaluation criteria—score withdrawal, escalation, and role-matching as “appropriate professional judgment,” not “abandonment of development.” Without this, EMs who execute the playbook take a personal hit for doing the right thing.

Three-case map:

Case	Situation	Move	Section
(1)	Trying to change, direction OK	Tighten the precision of support via type-by-type Do’s and Don’ts	§4
(2)	Direction looks wrong	Sort A–D, climb the light-to-heavy verb ladder, spot “Type D in Type A clothing”	§5
(3)	Not trying to change	Re-cast role, escalate, or withdraw. Drop the guilt	§6

Case (1) summary (by type):

Type	Main Do	Main Don’t	Boundary watch
Goal-Achievement Oriented	Name what was decisive in the achievement; communicate direction and let go	Micromanage	Can’t supply intrinsic motivation
Risk-Averse	Appreciate + make the safety net visible	Rush them, dismiss their concerns	Can’t forcibly remold caution
Energy-Conserving	Pair “why now / what they get” in every ask	Emotional appeals (“show motivation”)	Compensation system / values intervention is out of authority
Poor at Planning	Co-decompose tasks; structural scaffolding	Write them off as “not trying”	Off-hours bootcamps are a labor-law risk
Indecisive	Narrow to 2–3 options; small wins	Leave them alone; full delegation too early	Org-wide “tolerate failure” culture is not in one EM’s gift
Stubborn	“How does this look from the other side?”	Head-on confrontation, public negation	Forcibly changing values intrudes on freedom of thought / conscience

Case (2) summary (the four “wrongs”):

Kind	What it is	Move
A. Factual error	Broken design, production risk, misread data	Communicate based on facts (in authority)
B. Process / team impact	Skipped reviews, blocking, DoD violations	Address the rule and structure, not the person
C. Career choice	Predictions about evaluation, growth, market value	Information-only; return the decision to them
D. EM’s values / preferences	“More energy” / “I wouldn’t do that”	Don’t intervene (dead center of the §1 gravity well)

One action to try today (EMs): In your next 1on1, pick one report and tentatively place them in one of the six types using this article’s checklist. Then ask just one question: “Is the granularity of my instructions right for you?” The answer tells you whether you’re picking the right SL style. Then ask: “Is anything in your current direction giving you pause?”—that’s the entry point into §5.

One action to try today (managers, HR, executives): In your next review with each EM, alongside their “wins,” ask for one “correct withdrawal / escalation” and score it as a positive. Add “withdrew correctly” to the evaluation axis, alongside “didn’t give up.” That single change starts to shift EMs’ choices.

And one level deeper: “not trying to change people” is the maximum support you can give to people who are trying to change. The EM’s verb is not “develop.” It is “observe, document, communicate, connect, listen, wait, assign, organize”—only actions that fit on that list sustain team operations under the punishment-game structure. And the organization has a responsibility to evaluate that correctly. Only when both are in place does an exit from the punishment-game structure become visible.

Related Articles

For related themes, see:

• Engineering Management as a Punishment Game: The Three Separations Japanese IT Companies Need — the previous article. Three separations and structural amplifiers
• Conditions for the “Decisive Leader” Who Changes the Status Quo — leadership as the upper concept for handling Indecisive types
• A Transformation Guide for Companies That Can’t Leverage Deep-Diving Talent — avoiding misdiagnosis with the 5-factor matrix
• Why “Boiling Water” Training of the Ice Age Generation Doesn’t Reach Gen Z — SDT and inter-generational mismatch
• How “Instruction-Waiting” Types Survive the AI Era — redefining Poor-at-Planning and instruction-waiting types

References

References, numbered to match in-text citations.

Other references

• Previous article: Engineering Management as a Punishment Game — the structural diagnosis (three separations, four amplifiers, 5-factor matrix) on which this sequel is built

1. Over 90% of department heads are also playing managers — HR Pro / Sanno Institute of Management Research (survey conducted 2019, published 2020; sample: department heads at listed companies with 100+ employees). Playing-manager rate is precisely 95.8% (referred to in this article as “about 96%”); weighted average of IC work is 39.9%. [Reliability: Medium-High] ↩︎ ↩︎² ↩︎³

2. The “punishment game” of management: can the “strong middle” come back? — Yuji Kobayashi, Persol Research and Consulting (Dec 15, 2025). [Reliability: High] ↩︎

3. State of the Global Workplace 2024 — Gallup (2024 edition). Japan engagement rate 6%, East Asia average 18%, global average 23%. The same URL is updated annually, so the 2024 data can also be confirmed via PR Newswire. [Reliability: High (official international survey)] ↩︎

4. Patterns of mean-level change in personality traits across the life course: a meta-analysis of longitudinal studies — Roberts, B. W., Walton, K. E., & Viechtbauer, W., Psychological Bulletin, 132(1), pp. 1–25 (2006, DOI: 10.1037/0033-2909.132.1.1). Meta-analysis of 92 longitudinal studies showing that Big Five traits do change across adulthood, but at year-scale, slowly. Related: Roberts, B. W., & Mroczek, D. (2008) Current Directions in Psychological Science, 17(1), pp. 31–35 (DOI: 10.1111/j.1467-8721.2008.00543.x). The article’s claim that “external intervention rarely produces large short-term shifts” rests on the small magnitude of change and the year-scale time horizon in this meta-analysis. [Reliability: High (peer-reviewed meta-analysis)] ↩︎ ↩︎²

5. The Transtheoretical Model of Health Behavior Change — Prochaska, J. O., & Velicer, W. F., American Journal of Health Promotion, 12(1), pp. 38–48 (1997, DOI: 10.4278/0890-1171-12.1.38). Stages of Change model: external intervention does not work for individuals in precontemplation. The need to design intervention according to the person’s stage of readiness. [Reliability: High (peer-reviewed; widely adopted model)] ↩︎ ↩︎² ↩︎³ ↩︎⁴

6. Self-Determination Theory and the Facilitation of Intrinsic Motivation, Social Development, and Well-Being — Ryan & Deci, American Psychologist, Vol. 55(1), pp. 68–78 (2000). Related: Self-determination theory and work motivation — Gagné & Deci, Journal of Organizational Behavior, 26(4), pp. 331–362 (2005, DOI: 10.1002/job.322). [Reliability: High (peer-reviewed; heavily cited)] ↩︎ ↩︎² ↩︎³ ↩︎⁴

7. Type-based guidance for subordinates: vary how you say it to match their behavior and thinking patterns — Insource (2025). Practitioner article classifying subordinates into six types and organizing characteristics, instruction styles, and coaching tips. Direct basis for this article’s six-type lens. [Reliability: Medium (practitioner article, not peer-reviewed)] ↩︎ ↩︎² ↩︎³

8. Framework: the “Will/Skill matrix” for subordinate development — ecoslyme.com. Four quadrants of skill × motivation and development direction per quadrant. A widely used practitioner framework derived in part from Blanchard and others. [Reliability: Medium (practitioner explainer; original is classical management theory)] ↩︎ ↩︎² ↩︎³

9. What Is SL Theory: Switch Style by Type — Nihon no Jinjibu. Explainer of Situational Leadership, proposed by Hersey & Blanchard in 1977. Original: Hersey, P., & Blanchard, K. H. (1977). Management of Organizational Behavior. Four styles (Telling/Selling/Participating/Delegating) and four developmental stages (D1–D4). [Reliability: Medium-High (widely adopted management theory; classical original)] ↩︎ ↩︎² ↩︎³ ↩︎⁴

10. Labor Contract Act (Act No. 128 of 2007), Article 16 — e-Gov Law Search. “A dismissal shall, where it lacks objectively reasonable grounds and is not considered to be appropriate in general societal terms, be treated as an abuse of right and be invalid.” (Doctrine of Abuse of the Right to Dismiss.) [Reliability: High (statutory text)] ↩︎ ↩︎² ↩︎³

11. Shimonoseki Commercial High School case (Supreme Court, July 10, 1980) — Courts in Japan. Leading case on resignation pressure, holding that persistent resignation pressure is unlawful (tort). Defines the boundary between unlawful coerced resignation and lawful resignation requests. Related: Ministry of Health, Labour and Welfare “Working-Style Reform Handbook”. [Reliability: High (official material; case law)] ↩︎ ↩︎²

12. Labor Contract Act, Article 5 (Duty of Care) — “An employer shall, in conjunction with a labor contract, give such consideration as is necessary to enable workers to work while securing the safety of their lives and bodies.” Related case: Dentsu case (Supreme Court, March 24, 2000) and others. [Reliability: High (statutory text and case law)] ↩︎ ↩︎² ↩︎³

13. Span of Control / Span of Influence — Harvard Business Review (April 2012) and others. The Span of Control / Span of Influence distinction was widely popularized by Stephen Covey’s 7 Habits of Highly Effective People (1989). [Reliability: Medium-High (practitioner / management theory)] ↩︎

14. Psychological Safety and Learning Behavior in Work Teams — Edmondson, A., Administrative Science Quarterly, 44(2), pp. 350–383 (1999, DOI: 10.2307/2666999). Classic study on the impact of psychological safety on learning behavior and interpersonal risk-taking. [Reliability: High (peer-reviewed; heavily cited)] ↩︎

15. The Situational Leadership Theory: A Critical View — Graeff, C. L., Academy of Management Review, 8(2), pp. 285–291 (1983, DOI: 10.5465/amr.1983.4284738). Related: Evolution of situational leadership theory: A critical review — Graeff, C. L., Leadership Quarterly, 8(2), pp. 153–170 (1997). Classical critique of SL theory’s empirical basis (1983) and a follow-up reassessment (1997). [Reliability: Medium-High (academic critique)] ↩︎