4dx-d2-lead-measures

name: 4dx-d2-lead-measures description: | Coaches identifying weekly lead measures (predictive + influenceable) for a WIG + audit of a lead list. Use when about to track lag metrics or a generic dashboard: 'find lead measures', 'are our KPIs real leads?'. Stale leads → momentum-rescue. source_book: "The 4 Disciplines of Execution (2nd ed., 2021) — McChesney/Covey/Huling/Thele/Walker" source_chapter: "Chapter 3 — Discipline 2: Act on the Lead Measures; Chapter 7 — Translating Organizational Focus Into Executable Targets; Chapter 12 — Applying Discipline 1 (member operating frame); Chapter 13 — Applying Discipline 2" source_language: en tags: [d2, lead-measure, multi-scope, two-axis-test, predictive-influenceable, goodhart, solo, team-leader, team-member, 4dx] related_skills: - 4dx-d1-wig-formulation - 4dx-d1-wig-formulation - 4dx-d3-scoreboard - 4dx-d4-cadence - 4dx-sustain-momentum-rescue - 4dx-meta-team-leader-onboarding - using-four-dx-coach

4dx-d2-lead-measures — Lead-measure work across all roles (multi-scope)

Mission

Coach the user through Discipline 2 — Act on the Lead Measures — in whichever of three roles they hold against a Wildly Important Goal: picking their own personal leads (solo); facilitating their team's selection without dictating (team-leader); or mapping their personal sphere of influence on leads someone else already chose (team-member). The same two-axis test (predictive AND influenceable) and the same Goodhart self-check apply across all three; what differs is who picks, who facilitates, and who works with what's already given.

The book frames D2 as "the most-misunderstood discipline." Across all three modes, the central confusion is the same: lead measures are not generic "leading indicators" (predictive only), not habits (untied to a specific WIG), not sub-WIGs (lag on a smaller clock), not tasks (one-shot deliverables), and not OKR Key Results (mix of lag and lead). This skill's job is to surface the two-axis test loudly and protect the user from each of those collapses.

When this skill activates

Multilingual trigger phrasings (covering all 3 modes)

Solo (personal lead-measure discovery, voice = personal coach):

• EN: "I have the goal but I don't know what to do every day", "What's a leading indicator for [my goal]?", "I'm working hard but nothing's moving — what am I missing?", "What should I track to know I'm on the right path?"
• JP:「目標は決まったが、毎日何をすればいい？」「先行指標って何を選べばいい？」 「頑張ってるのに数字が動かない、何が間違ってる？」
• zh-TW:「我有目標了，但不知道每天該做什麼才會有結果」「該追蹤什麼指標才知道 我走在對的路上？」「我以為我在做對的事，但結果沒變化」

Team-leader (facilitate team's selection, voice = consultant-to-leader):

• EN: "How do I run a lead-measure session with my team without just telling them the answer?", "My team brainstormed 12 candidates — how do I narrow to 2-3?", "Should I prepare draft lead measures before the team meeting?", "My team can't agree on what we can actually influence."
• JP:「team で lead measure を一緒に決めたい。どう facilitate すればいい？」 「メンバーが提案した先行指標、どう veto する？」「6 個候補が出たが、2-3 に 絞れない」
• zh-TW:「帶 team 一起找 lead measure，但我不想直接給答案，怎麼引導？」 「成員提的指標感覺不對，但我不想 dictate，怎麼 veto？」「team 在吵某個指標 到底我們控不控得到」

Team-member (influence-map on inherited leads, voice = personal coach to member):

• EN: "Our team's lead measures are set — how do I influence them?", "I'm on a 4DX team, leads are picked, what do I actually do?", "Which lead measure is mine to move?", "I support all of them but feel stretched thin."
• JP:「team の lead measure は決まったが、自分はどう関われる？」「自分の立場 でどの lead を動かせるか分からない」「全部の lead を見てるが手応えがない」
• zh-TW:「團隊的 lead measure 已經訂了，我怎麼影響？」「lead 已經選好但我不 知道哪些跟我的位置有關」「我支援全部 lead 但完全被攤薄」

Audit (existing artifact + optional stakeholder critique, voice = consultant-to-stakeholder):

• EN: "Audit our lead measures", "Got feedback our leads aren't quite right — help me see what's wrong", "Stakeholder wants to revisit our lead measures — diagnose", "Manager flagged we should sharpen the leads", "Boss says these aren't real lead measures", "Are our 12 KPIs the right leads?", "Diagnose this list against 4DX", "Which of these pass the two-axis test?"
• JP:「うちの lead measure を診断して」「これ本当に lead measure？」 「上司にフィードバックされた、lead どこ直すべき？」「マネージャーから 方向性を直したいと言われた、lead 見て」「ボスに『これは lead じゃない』 と言われた、どこが違う？」「KPI 12 個あるが、どれが本当の lead？」
• zh-TW:「主管覺得我們的 lead 要調整，幫看哪裡」「review 後 stakeholder 想 討論 lead 方向」「老闆說我們的 lead 不對，幫看哪裡」「我們列了 12 個指標， 哪些是真 lead？」「幫我用 two-axis test 過一遍這份清單」「診斷這份 lead 清單」

Non-activation signals (DO NOT fire when…)

• No defined WIG yet — leads need a lag to be predictive of. Route to 4dx-d1-wig-formulation (solo) or the relevant team-WIG skill first.
• Member doesn't know team WIG / leads — comprehension upstream; route to 4dx-d1-wig-formulation first.
• Scoreboard / visualization — 4dx-d3-scoreboard is the artifact that DISPLAYS chosen leads.
• Weekly review / commitment / WIG Session — 4dx-d4-cadence is the cadence that REVIEWS lead performance.
• Lead measures already exist but lag is flat ≥4 weeks — that's a diagnosis-and-redesign job for 4dx-sustain-momentum-rescue, not first-time selection.
• Stroke-of-pen WIG — goal achieved by single decision/purchase (sign contract, hire, reorg). No daily behavior to discover.
• Out-of-4DX — generic "leading indicator" finance talk, OKR Key Results, sprint goals, KPI dashboards without 4DX framing → hand off via using-four-dx-coach.
• Multi-team / leader-of-leaders cascade — out of plugin scope; read book Part 2 directly.

Scope detection

When this skill activates:

1. Determine role with respect to the lead measures: solo (no team or personal-WIG context) / team-leader-facilitator / team-member.
2. Load the matching protocol file from protocols/.
3. Follow that protocol's E section step-by-step.

If the role is ambiguous after reading the user's query, ask ONE Socratic disambiguation question:

EN: "Quick clarification — what's your role with respect to the lead measures: picking them for yourself (solo, your own goal), facilitating your team picking theirs (you lead 3-9 reports), or working with leads your team already has (you're a member receiving what was chosen)? I'll route to the right protocol."

JP:「先に確認 —— lead measure に対するあなたの役割は: 自分の goal 用に 自分で選ぶ（solo）、team を率いて選定を facilitate（leader、3-9 名 の direct reports）、それとも 既に team にあるものに対して関わる（member）？ 適切な protocol に振り分けます。」

zh-TW:「先確認一下 —— 你跟 lead measure 的關係：自己挑給自己用（solo）、 帶 team 一起挑（leader，3-9 名 direct report）、還是 接收 team 已經 訂好的（member）？我幫你導到對的 protocol。」

If the signal in the original query is strong (e.g. "my own goal", "facilitate my team", "leads my manager set"), skip the question and load the protocol directly.

Protocol routing table

After loading the protocol, follow its E section step-by-step. Each protocol carries its own R / I / A1 / A2 / E / B sections; this orchestrator does not run any D2 content directly.

Mode detection (coach vs audit)

• No artifact provided → coach-mode. Detect role (solo / leader / member) per the table above, ask one Socratic disambiguation question if ambiguous.
• Artifact provided (named lead list / KPI dashboard / OKR-KR sheet) → audit-mode (protocols/audit-mode.md), regardless of role. The audit's terminal recommendation routes back to coach-mode if the artifact is irreparable.
• Artifact provided AND user explicitly asks for Socratic rebuild → coach-mode (user override). Audit-mode is the default for artifacts unless the user opts out.

Edge-case routing

• WIG not yet defined → fire 4dx-d1-wig-formulation first; D2 has no anchor.
• Member-but-can't-name-team-WIG-or-leads → fire 4dx-d1-wig-formulation first; influence-mapping over unknown leads produces theatre.
• Member acting as facilitator (leader delegated) → fire protocols/team-facilitate.md with reduced veto authority noted; the actual veto-on-strategic-fit belongs to the leader.
• Leader-of-leaders / multi-team cascade → out of plugin scope; route via using-four-dx-coach.
• Leads already chosen but stale (lead green, lag flat ≥4 weeks) → fire 4dx-sustain-momentum-rescue, not this skill.
• Cross-layer audit (artifact spans WIG + leads + scoreboard + cadence) → fire 4dx-audit (5-layer consultant audit), not audit-mode.md (which is L2-only).

Two types of lead measures (book p 224)

Both coach-mode and audit-mode protocols depend on this typology. Most users default to one type without realizing the other exists; the choice shapes scoreboard design and accountability.

Both types are equally valid applications of D2. The question is not which is better — it's which fits this team's WIG and culture.

Worked examples (book pp 227-228)

Younger Brothers Construction — WIG: Reduce average monthly accidents from 12 to 7 by December 31

• Small outcome: "Achieve average safety-compliance score of 97% each week."
• Leveraged behavior: "Ensure that 95% of associates wear safety boots every day."

Grocery chain — WIG: Increase average weekly sales from $1 million to $1.5 million by December 31

• Small outcome: "Limit out-of-stocks on top items to 20 or fewer per week."
• Leveraged behavior: "Complete 2 additional shelf reviews each day, filling all holes on top items."

Shared standards

All four protocols (3 coach-mode + 1 audit-mode) share the same canonical D2 standards — coach-mode applies them via Socratic discovery, audit-mode applies them as PASS/FAIL synthesis on a provided artifact. Load on demand:

• standards/predictive-and-influenceable.md — the book's two-axis test: predictive of the lag AND influenceable by the user/team without permission. Both axes must be 4-5 out of 5; failing either disqualifies completely.
• standards/goodhart-self-check.md — once stakes are attached, the measure gets gamed. Three documented industry collapses (Wells Fargo / Phoenix VA / Atlanta APS) and the three personal mitigations (frequency-quality pair, 4-week causal-chain re-check, written forecast).
• standards/lead-vs-lag-distinction.md — what a lead measure is NOT: not a lag-on-a-shorter-clock, not a sub-WIG / Battle, not a one-shot task, not a generic "leading indicator," not an OKR KR.

Cross-skill relations

• Upstream prerequisite — 4dx-d1-wig-formulation (solo / leader) defines the WIG; 4dx-d1-wig-formulation confirms a member can name team WIG + leads verbatim. D2 has nothing to operate on without a clearly-stated WIG.
• Downstream composers — 4dx-d3-scoreboard displays the chosen leads; 4dx-d4-cadence runs the weekly review of lead performance vs targets and converts a focus lead into a specific weekly commitment.
• Diagnostic neighbour — 4dx-sustain-momentum-rescue runs after leads have been chosen and proven stale (green lead, flat lag for 4+ weeks). The boundary is sharp: this skill picks/maps leads; rescue redesigns existing ones.
• Plugin-router fallback — using-four-dx-coach handles cold-start triage and out-of-4DX queries (KPI / OKR KR / sprint goal); not a substitute for this skill, but the right hand-off when the question turns out not to be 4DX D2.

Boundary (cross-mode common)

The mode-specific boundary lives in each protocol's B section. The cross-mode common boundary:

• Two-axis test is mandatory across all 3 modes — predictive AND influenceable; failing either axis disqualifies. Most-common collapse: the user / team picks something predictive but not influenceable ("market direction", "rainfall") or influenceable but not predictive ("drink 2L water", "more meetings"). Reject both.
• 2-3 leads cap, hard — one frequency-based + one quality-based is the canonical pair. More than 3 recreates the whirlwind under a 4DX label. Solo: cap at 3; team: cap at 3, prefer 2; member focus: cap at 1-2 of the team's leads.
• Goodhart self-check applies in all 3 modes — once stakes are attached, leads get gamed. Industry: Wells Fargo / Phoenix VA / Atlanta APS. Personal scale: speed-reading without comprehension, screen-staring counted as writing, voicemails counted as cold calls. Defend with frequency-quality pairing and 4-week causal-chain re-check.
• Coach is Socratic; never authors leads — solo agent doesn't invent the user's leads; team-facilitator-coach doesn't author the team's leads; member-coach doesn't invent the member's influence map. The ownership mechanism that makes lead-keeping work depends on the user / team / member doing the picking.
• Honest "no influence" beats faked focus (member mode) — when a member's per-lead influence audit comes back genuinely empty, the right move is escalation to the leader, not picking a focus to look engaged. The protocol surfaces that signal explicitly.

Audit metadata

• Skill type: multi-file orchestrator with dual-mode posture (3 coach-mode protocols: solo / team-leader / team-member; 1 audit-mode protocol: artifact diagnosis)
• Verification status: V1 ✓ for solo + team-leader modes (both leader-POV in source book Ch 3 / Ch 7 / Ch 13); V1 ⚠️ partial for member-influence mode (book authors leader side of selection; member- POV inferred from Ch 3's two-axis test read from individual seat, Ch 12's daily question, Ch 13's frontline implementation cases)
• Industry grounding: Goodhart 1975 / Strathern 1997 (mechanism); CFPB 2016 / VA OIG 2014 / GBI 2011 (gaming cases); Edmondson 1999/2012 (psychological safety, facilitator mode); Lencioni 2002 (trust → conflict → commitment, facilitator mode); Pfeffer & Sutton 2006 (evidence-based skepticism, facilitator mode); Pfeffer 2010 (power axes, member mode); Cialdini 1984/2007 (six-principle leverage, member mode); Argyris HBR 1986 (skilled incompetence / Model-II inquiry, member mode)
• Created: 2026-04-30
• Output language: SKILL.md body + protocols + standards in English; description + scope-detection prompts multilingual EN / JP / zh-TW