Resistance Diagnosis
When to use
Diagnose observed resistance — separate symptoms from root causes; offer response options across sponsor, comms, training, and structural levers without mandating a single fix.
What this skill does not do
- Does not log governance decisions — route continue/escalate calls to
/pmo:decision-log when installed.
- Does not label people as blockers — describes behaviors and incentives, not character.
- Does not recommend punitive action by default — escalation is an option, not the default response.
Preconditions
| Input |
If missing |
| Observed resistance (quotes, behaviors, metrics) |
Ask for specific examples |
| Stakeholder segment affected |
Ask who; link to map if available |
| Practice profile resistance taxonomy |
Use loss/competence/habit/politics defaults |
Provisional mode
With anecdote only: hypothesize root causes; flag [review] — need corroboration before major escalation.
Trust spine
- Confidence bands (
hypothesis-driven-analysis):
- High: Multiple independent observations; root cause tied to named loss or incentive.
- Medium: Single source or manager report; hypotheses labeled.
- Low: Vague "they don't like it" — ask for behaviors before prescribing.
- Failure modes:
- Analytical rigor: Distinguish active sabotage from passive workarounds from capacity overload.
- Strategic advice vs. support: Present response options; leadership owns escalation and trade-offs.
Workflow
Read practice profile — resistance taxonomy, escalation thresholds, sponsor model.
Document observed resistance — behaviors (skipping training, shadow IT, public criticism, quiet non-use), not labels ("toxic," "blocker").
Classify root cause (may be multiple):
- Loss — status, autonomy, relationships, job security.
- Competence — fear of failing at new process; insufficient skill.
- Habit — old way is faster under pressure; no muscle memory yet.
- Politics — incentive misalignment; another leader's priority conflicts.
- Capacity — change stacked on BAU with no time budget — often misread as attitude.
Test hypotheses — what evidence would confirm or falsify each cause; what question to ask the segment or manager.
Response options (pick subset relevant — do not recommend all):
- Sponsor — visible removal of conflicting priority, personal ask to holdout influencer.
- Comms — reframe WIIFM, address loss explicitly, manager talking points.
- Training — job aid, coaching, safe practice environment.
- Structural — metric change, role clarity, decision-rights fix (hand off to operating-model if needed).
- Timing — narrow scope, pilot, defer — with trade-off stated.
- Escalation — per profile threshold; what decision is needed from steering.
Flag if resistance is rational — when the new process is genuinely worse for the segment, say so [review] — comms cannot fix a bad design.
Output format
SEGMENT: [name]
OBSERVED BEHAVIORS: [specific, sourced]
ROOT CAUSE ANALYSIS:
Loss: [evidence or "not primary"]
Competence: [...]
Habit: [...]
Politics: [...]
Capacity: [...]
PRIMARY HYPOTHESIS: [cause] — [mechanism]
FALSIFICATION TEST: [question or evidence to gather]
RESPONSE OPTIONS:
Sponsor: [option or "N/A"]
Comms: [...]
Training: [...]
Structural: [...]
Timing: [...]
Escalation: [decision needed, or "below threshold"]
RATIONAL RESISTANCE FLAG: [yes — design issue | no]
Worked example
Input: "Sales won't use the CRM."
Expected output: Split competence (slow on mobile) vs habit (spreadsheet faster); primary hypothesis = habit under quota pressure; sponsor option = remove parallel reporting ask; not "mandatory training again" as sole fix.
Quality checks before delivering
Propose profile update
When a stable convention surfaces during this run (thresholds, naming, tone, output format, or recurring corrections), propose a profile update: show the exact diff against ~/.claude/plugins/config/claude-for-strategy/change-management/CLAUDE.md (org-wide facts go to org-profile.md), ask for confirmation, and write only on yes. Only /change-management:practice-setup auto-applies a full profile write.
Outputs
Follows plugin CLAUDE.md § Outputs. Next: sponsor-roadmap, communications-plan, or /pmo:decision-log for escalation.
1---2name: resistance-diagnosis3description: This skill should be used when the user describes "resistance to the change," "people pushing back," "adoption stalling," or "blockers on the people side" and needs root-cause split and response options before escalation or comms rework.4---5# Resistance Diagnosis67## When to use89Diagnose observed resistance — separate symptoms from root causes; offer response options across sponsor, comms, training, and structural levers without mandating a single fix.1011## What this skill does not do1213- **Does not log governance decisions** — route continue/escalate calls to `/pmo:decision-log` when installed.14- **Does not label people as blockers** — describes behaviors and incentives, not character.15- **Does not recommend punitive action by default** — escalation is an option, not the default response.1617## Preconditions1819| Input | If missing |20|---|---|21| Observed resistance (quotes, behaviors, metrics) | Ask for specific examples |22| Stakeholder segment affected | Ask who; link to map if available |23| Practice profile resistance taxonomy | Use loss/competence/habit/politics defaults |2425## Provisional mode2627With anecdote only: hypothesize root causes; flag `[review]` — need corroboration before major escalation.2829## Trust spine3031- **Confidence bands** (`hypothesis-driven-analysis`):32 - **High:** Multiple independent observations; root cause tied to named loss or incentive.33 - **Medium:** Single source or manager report; hypotheses labeled.34 - **Low:** Vague "they don't like it" — ask for behaviors before prescribing.35- **Failure modes:**36 - **Analytical rigor:** Distinguish active sabotage from passive workarounds from capacity overload.37 - **Strategic advice vs. support:** Present response options; leadership owns escalation and trade-offs.3839## Workflow40411. **Read practice profile** — resistance taxonomy, escalation thresholds, sponsor model.42432. **Document observed resistance** — behaviors (skipping training, shadow IT, public criticism, quiet non-use), not labels ("toxic," "blocker").44453. **Classify root cause** (may be multiple):46 - **Loss** — status, autonomy, relationships, job security.47 - **Competence** — fear of failing at new process; insufficient skill.48 - **Habit** — old way is faster under pressure; no muscle memory yet.49 - **Politics** — incentive misalignment; another leader's priority conflicts.50 - **Capacity** — change stacked on BAU with no time budget — often misread as attitude.51524. **Test hypotheses** — what evidence would confirm or falsify each cause; what question to ask the segment or manager.53545. **Response options** (pick subset relevant — do not recommend all):55 - **Sponsor** — visible removal of conflicting priority, personal ask to holdout influencer.56 - **Comms** — reframe WIIFM, address loss explicitly, manager talking points.57 - **Training** — job aid, coaching, safe practice environment.58 - **Structural** — metric change, role clarity, decision-rights fix (hand off to operating-model if needed).59 - **Timing** — narrow scope, pilot, defer — with trade-off stated.60 - **Escalation** — per profile threshold; what decision is needed from steering.61626. **Flag if resistance is rational** — when the new process is genuinely worse for the segment, say so `[review]` — comms cannot fix a bad design.6364## Output format6566```67SEGMENT: [name]68OBSERVED BEHAVIORS: [specific, sourced]6970ROOT CAUSE ANALYSIS:71 Loss: [evidence or "not primary"]72 Competence: [...]73 Habit: [...]74 Politics: [...]75 Capacity: [...]7677PRIMARY HYPOTHESIS: [cause] — [mechanism]78FALSIFICATION TEST: [question or evidence to gather]7980RESPONSE OPTIONS:81 Sponsor: [option or "N/A"]82 Comms: [...]83 Training: [...]84 Structural: [...]85 Timing: [...]86 Escalation: [decision needed, or "below threshold"]8788RATIONAL RESISTANCE FLAG: [yes — design issue | no]89```9091## Worked example9293**Input:** "Sales won't use the CRM."9495**Expected output:** Split competence (slow on mobile) vs habit (spreadsheet faster); primary hypothesis = habit under quota pressure; sponsor option = remove parallel reporting ask; not "mandatory training again" as sole fix.9697## Quality checks before delivering9899- [ ] Behaviors documented, not personality labels100- [ ] Multiple causes considered101- [ ] Response options include trade-offs102- [ ] Rational resistance checked103104## Propose profile update105106When a stable convention surfaces during this run (thresholds, naming, tone, output format, or recurring corrections), **propose a profile update**: show the exact diff against `~/.claude/plugins/config/claude-for-strategy/change-management/CLAUDE.md` (org-wide facts go to `org-profile.md`), ask for confirmation, and write only on yes. Only `/change-management:practice-setup` auto-applies a full profile write.107108## Outputs109110Follows plugin `CLAUDE.md` § Outputs. Next: `sponsor-roadmap`, `communications-plan`, or `/pmo:decision-log` for escalation.
daddia