CX Error Analyzer(CXエラーシナリオ分析)
Overview
CX Error Analyzer は、システムやアプリケーションのエラーシナリオを顧客体験(Customer Experience)の観点から体系的に分析・評価するスキルです。6軸の定量評価フレームワークにより、エラーが顧客体験に与える影響を可視化し、Impact vs Effort マトリクスで改善施策の優先順位を決定します。
When to Use
- エラーシナリオの顧客体験を評価するとき
- エラーメッセージの品質を改善するとき
- エラー改善の優先順位を付けるとき
- CX観点でのエラーハンドリング改善レポートを作成するとき
- エラー体験の定量的評価を行うとき
- 新機能リリース前にエラーシナリオのCXレビューを実施するとき
- サポートチケット削減のための改善施策を検討するとき
Prerequisites
- Error data: Access to error logs, support tickets, or documented error scenarios
- User journey map (optional): Understanding of user flows helps contextualize error impact
- Stakeholder access (optional): Ability to interview developers, support staff, or QA for error pattern insights
- Business metrics (optional): Support cost data, churn rates, or CSAT scores for ROI estimation
Output
This skill provides conversational guidance and analysis rather than file generation. Outputs include:
- Structured error scenario inventories with classifications
- Multi-axis CX evaluation scores with rationale
- Prioritized improvement recommendations (Quick Wins, Strategic Projects)
- Before/after error message comparisons
- ROI estimates and business impact projections
When formal documentation is needed, use Workflow 5 to generate a comprehensive report using the provided templates.
Workflows
Workflow 1: Error Scenario Inventory(エラーシナリオ棚卸し)
Enumerate and classify all error scenarios in the target system or feature.
- List all error scenarios in the system or feature under analysis
- Review source code, API documentation, and existing error logs
- Interview stakeholders (developers, support, QA) for known error patterns
- Examine user feedback and support ticket data for unreported errors
- Classify each error into one of 6 categories:
- Validation: Input format, required fields, range limits
- System: Server errors, database failures, memory issues
- Network: Timeout, connection loss, DNS failure
- Auth: Authentication failure, session expiry, permission denied
- Business Logic: Rule violations, state conflicts, limit exceeded
- External: Third-party API failures, payment gateway errors, integration issues
- Map each error to a user journey stage:
- Discovery (browsing, searching)
- Onboarding (registration, initial setup)
- Core Task (primary feature usage)
- Checkout (purchase, submission, completion)
- Support (help, account management)
- Assign error IDs: ERR-001, ERR-002, ERR-003...
- Capture current state for each error:
- Current error message text (exact wording)
- Current UX flow (what happens visually)
- Current recovery path (how the user proceeds)
Workflow 2: Multi-Axis CX Evaluation(多軸CX評価)
Evaluate each error scenario across 6 dimensions using a 1-5 scale.
- Load evaluation criteria from
references/cx_evaluation_methodology.md
- Evaluate each error scenario on 6 axes (1-5 scale each):
- Impact Severity (影響度): How much does this error disrupt the user's task?
- Frequency (頻度): How often do users encounter this error?
- Recovery Ease (復旧容易性): How easy is it for users to recover? (inverse: 5 = hard to recover)
- Message Quality (メッセージ品質): How helpful is the current error message? (inverse: 5 = poor message)
- Emotional Impact (感情的影響): How frustrated or anxious does the user feel?
- Business Cost (ビジネスコスト): What is the business impact (support tickets, churn, revenue loss)?
- Document rationale for each score
- Use
assets/cx_error_evaluation_template.md to record each evaluation
Workflow 3: CX Scoring(CXスコアリング)
Calculate weighted CX Scores and classify by severity tier.
Apply weights to each axis:
| Axis |
Weight |
| Impact Severity (影響度) |
25% |
| Frequency (頻度) |
20% |
| Recovery Ease (復旧容易性) |
15% |
| Message Quality (メッセージ品質) |
15% |
| Emotional Impact (感情的影響) |
10% |
| Business Cost (ビジネスコスト) |
15% |
Calculate weighted CX Score (1.0-5.0) for each scenario:
CX Score = (Impact * 0.25) + (Frequency * 0.20) + (Recovery * 0.15)
+ (Message * 0.15) + (Emotional * 0.10) + (Business * 0.15)
Higher score = worse customer experience.
Classify by CX Score tier:
| Score Range |
Tier |
Action |
| 4.0-5.0 |
Critical CX Issue |
Immediate action required |
| 3.0-3.9 |
Significant CX Issue |
Plan improvement |
| 2.0-2.9 |
Moderate CX Issue |
Add to backlog |
| 1.0-1.9 |
Minor CX Issue |
Nice to have |
Rank all scenarios by CX Score descending
Refer to references/cx_metrics_reference.md for business impact estimation
Workflow 4: Improvement Priority Matrix(改善優先度マトリクス)
Plot errors on an Impact vs Effort matrix to determine improvement priorities.
Assess implementation effort for improving each error scenario:
- Low Effort: Message text change, minor UI update (< 1 day)
- Medium Effort: Recovery flow redesign, new validation logic (1-5 days)
- High Effort: Architectural change, new error handling infrastructure (> 5 days)
Plot on 2x2 matrix (Impact = CX Score, Effort = Implementation Difficulty):
High Impact ┌──────────────┬──────────────┐
│ Strategic │ Quick Wins │
│ Projects │ (Do First) │
│ │ │
├──────────────┼──────────────┤
│ Deprioritize │ Fill-Ins │
│ (Avoid) │ (Batch) │
Low Impact └──────────────┴──────────────┘
High Effort Low Effort
Identify Quick Wins (High Impact + Low Effort) for immediate action
Plan Strategic Projects (High Impact + High Effort) with proper roadmap
Suggest specific improvements for top priority items:
- Load
references/error_ux_best_practices.md for design patterns
- Draft improved error messages (before/after)
- Design improved recovery flows
- Estimate improvement impact using
references/cx_metrics_reference.md
Workflow 5: Report Generation(レポート出力)
Compile all analysis results into a comprehensive CX Error Analysis Report.
- Load report template from
assets/cx_error_report_template.md
- Populate all sections:
- Executive summary with key findings
- Complete error scenario inventory
- Multi-axis evaluation results for every scenario
- CX Score ranking
- Priority matrix visualization
- Quick Wins list with specific action items
- Detailed improvement proposals (top scenarios)
- Expected benefits (support cost reduction, CSAT improvement, churn reduction)
- Include quantitative estimates where possible:
- Use
references/cx_metrics_reference.md for ROI calculations
- Estimate support ticket reduction
- Project CSAT/CES improvement
- Review and finalize the report
Resources
| File |
Purpose |
When to Load |
references/cx_evaluation_methodology.md |
Detailed scoring criteria for all 6 evaluation axes, error classification taxonomy, and data collection guidance |
Workflow 2 (Multi-Axis CX Evaluation) |
references/error_ux_best_practices.md |
Error message design principles, recovery flow patterns, emotional design, accessibility guidelines, and good/bad examples |
Workflow 4 (Improvement Priority Matrix) when drafting improvements |
references/cx_metrics_reference.md |
CX metrics (CES, CSAT, NPS), support cost formulas, churn risk estimation, industry benchmarks, and ROI calculation |
Workflow 3 (CX Scoring) and Workflow 5 (Report Generation) |
assets/cx_error_report_template.md |
Complete CX Error Analysis Report template with all sections and placeholders |
Workflow 5 (Report Generation) |
assets/cx_error_evaluation_template.md |
Individual error scenario evaluation card template for per-scenario assessment |
Workflow 2 (Multi-Axis CX Evaluation) for each scenario |
Best Practices
Do's
- 実際のエラーメッセージテキストをそのまま記録する(推測しない)
- ユーザージャーニーの文脈でエラーを評価する(同じエラーでも決済時とブラウジング時で影響度が異なる)
- 改善提案には必ず「改善前→改善後」のメッセージ比較を含める
- ROI計算にサポートコスト削減とチャーンリスク低減の両方を含める
- Quick Winsを最優先で特定する(小さな改善でも即効性が高い)
Don'ts
- 技術的な正確性だけでエラーを評価しない(CX観点が主軸)
- 頻度データなしにスコアリングしない(推定でも頻度を見積もる)
- 感情的影響を過小評価しない(ユーザーの不安・焦りは離脱に直結する)
- 改善コストの見積もりを省略しない(Impact vs Effort マトリクスに必須)
Examples
例1: ECサイトのエラーCX分析
User: ECサイトのエラー体験を改善したい。主要なエラーシナリオを分析してください。
Claude:
1. [Workflow 1] エラー棚卸し → ERR-001(在庫切れ), ERR-002(決済失敗),
ERR-003(セッションタイムアウト), ERR-004(住所バリデーション), ...
2. [Workflow 2] 6軸評価 → ERR-002(決済失敗)が最高スコア4.2: 影響度5, 復旧困難4, 感情5
3. [Workflow 3] CXスコア計算 → ERR-002: 4.2(Critical), ERR-003: 3.5(Significant), ...
4. [Workflow 4] 優先度マトリクス → Quick Win: ERR-004(メッセージ改善のみ、工数Low)
5. [Workflow 5] レポート出力: 改善提案8件、推定サポートコスト削減20%
例2: SaaSログイン周りのエラーUX改善
User: ログイン・認証エラーのUXを改善したい。
Claude:
1. [Workflow 1] 棚卸し → ERR-001(パスワード誤り), ERR-002(アカウントロック),
ERR-003(2FA失敗), ERR-004(SSO連携エラー)
2. [Workflow 2] 各シナリオを評価 → ERR-002のメッセージ品質が5(最低: "Account locked"のみ)
3. [Workflow 4] Quick Win: ERR-002のメッセージ改善
Before: "Account locked. Contact support."
After: "セキュリティのためアカウントがロックされました。30分後に自動解除されます。
すぐにログインが必要な場合はパスワードリセットをお試しください。"
1---2name: cx-error-analyzer3description: エラーや例外シナリオを顧客体験(CX)の観点で体系的に分析し、改善優先度を付ける スキル。6軸評価(影響度・頻度・復旧容易性・メッセージ品質・感情的影響・ ビジネスコスト)でスコアリングし、Impact vs Effort マトリクスで改善施策を 優先順位付けする。Use when analyzing error scenarios from customer experience perspective, evaluating error message quality, prioritizing error UX improvements, or creating CX-focused error analysis reports. Triggers: "CXエラー分析", "error experience", "エラーUX改善", "customer experience error", "エラーメッセージ品質", "error scenario analysis", "顧客体験エラー評価"4---56# CX Error Analyzer(CXエラーシナリオ分析)78## Overview910CX Error Analyzer は、システムやアプリケーションのエラーシナリオを顧客体験(Customer Experience)の観点から体系的に分析・評価するスキルです。6軸の定量評価フレームワークにより、エラーが顧客体験に与える影響を可視化し、Impact vs Effort マトリクスで改善施策の優先順位を決定します。1112## When to Use1314- エラーシナリオの顧客体験を評価するとき15- エラーメッセージの品質を改善するとき16- エラー改善の優先順位を付けるとき17- CX観点でのエラーハンドリング改善レポートを作成するとき18- エラー体験の定量的評価を行うとき19- 新機能リリース前にエラーシナリオのCXレビューを実施するとき20- サポートチケット削減のための改善施策を検討するとき2122## Prerequisites2324- **Error data**: Access to error logs, support tickets, or documented error scenarios25- **User journey map** (optional): Understanding of user flows helps contextualize error impact26- **Stakeholder access** (optional): Ability to interview developers, support staff, or QA for error pattern insights27- **Business metrics** (optional): Support cost data, churn rates, or CSAT scores for ROI estimation2829## Output3031This skill provides **conversational guidance and analysis** rather than file generation. Outputs include:3233- Structured error scenario inventories with classifications34- Multi-axis CX evaluation scores with rationale35- Prioritized improvement recommendations (Quick Wins, Strategic Projects)36- Before/after error message comparisons37- ROI estimates and business impact projections3839When formal documentation is needed, use Workflow 5 to generate a comprehensive report using the provided templates.4041## Workflows4243### Workflow 1: Error Scenario Inventory(エラーシナリオ棚卸し)4445Enumerate and classify all error scenarios in the target system or feature.46471. **List all error scenarios** in the system or feature under analysis48 - Review source code, API documentation, and existing error logs49 - Interview stakeholders (developers, support, QA) for known error patterns50 - Examine user feedback and support ticket data for unreported errors512. **Classify each error** into one of 6 categories:52 - **Validation**: Input format, required fields, range limits53 - **System**: Server errors, database failures, memory issues54 - **Network**: Timeout, connection loss, DNS failure55 - **Auth**: Authentication failure, session expiry, permission denied56 - **Business Logic**: Rule violations, state conflicts, limit exceeded57 - **External**: Third-party API failures, payment gateway errors, integration issues583. **Map each error to a user journey stage**:59 - Discovery (browsing, searching)60 - Onboarding (registration, initial setup)61 - Core Task (primary feature usage)62 - Checkout (purchase, submission, completion)63 - Support (help, account management)644. **Assign error IDs**: ERR-001, ERR-002, ERR-003...655. **Capture current state** for each error:66 - Current error message text (exact wording)67 - Current UX flow (what happens visually)68 - Current recovery path (how the user proceeds)6970### Workflow 2: Multi-Axis CX Evaluation(多軸CX評価)7172Evaluate each error scenario across 6 dimensions using a 1-5 scale.73741. **Load evaluation criteria** from `references/cx_evaluation_methodology.md`752. **Evaluate each error scenario** on 6 axes (1-5 scale each):76 - **Impact Severity (影響度)**: How much does this error disrupt the user's task?77 - **Frequency (頻度)**: How often do users encounter this error?78 - **Recovery Ease (復旧容易性)**: How easy is it for users to recover? (inverse: 5 = hard to recover)79 - **Message Quality (メッセージ品質)**: How helpful is the current error message? (inverse: 5 = poor message)80 - **Emotional Impact (感情的影響)**: How frustrated or anxious does the user feel?81 - **Business Cost (ビジネスコスト)**: What is the business impact (support tickets, churn, revenue loss)?823. **Document rationale** for each score834. **Use `assets/cx_error_evaluation_template.md`** to record each evaluation8485### Workflow 3: CX Scoring(CXスコアリング)8687Calculate weighted CX Scores and classify by severity tier.88891. **Apply weights** to each axis:90 | Axis | Weight |91 |------|--------|92 | Impact Severity (影響度) | 25% |93 | Frequency (頻度) | 20% |94 | Recovery Ease (復旧容易性) | 15% |95 | Message Quality (メッセージ品質) | 15% |96 | Emotional Impact (感情的影響) | 10% |97 | Business Cost (ビジネスコスト) | 15% |98992. **Calculate weighted CX Score** (1.0-5.0) for each scenario:100 ```101 CX Score = (Impact * 0.25) + (Frequency * 0.20) + (Recovery * 0.15)102 + (Message * 0.15) + (Emotional * 0.10) + (Business * 0.15)103 ```104 Higher score = worse customer experience.1051063. **Classify by CX Score tier**:107 | Score Range | Tier | Action |108 |-------------|------|--------|109 | 4.0-5.0 | Critical CX Issue | Immediate action required |110 | 3.0-3.9 | Significant CX Issue | Plan improvement |111 | 2.0-2.9 | Moderate CX Issue | Add to backlog |112 | 1.0-1.9 | Minor CX Issue | Nice to have |1131144. **Rank all scenarios** by CX Score descending1155. **Refer to `references/cx_metrics_reference.md`** for business impact estimation116117### Workflow 4: Improvement Priority Matrix(改善優先度マトリクス)118119Plot errors on an Impact vs Effort matrix to determine improvement priorities.1201211. **Assess implementation effort** for improving each error scenario:122 - **Low Effort**: Message text change, minor UI update (< 1 day)123 - **Medium Effort**: Recovery flow redesign, new validation logic (1-5 days)124 - **High Effort**: Architectural change, new error handling infrastructure (> 5 days)1252. **Plot on 2x2 matrix** (Impact = CX Score, Effort = Implementation Difficulty):126127 ```128 High Impact ┌──────────────┬──────────────┐129 │ Strategic │ Quick Wins │130 │ Projects │ (Do First) │131 │ │ │132 ├──────────────┼──────────────┤133 │ Deprioritize │ Fill-Ins │134 │ (Avoid) │ (Batch) │135 Low Impact └──────────────┴──────────────┘136 High Effort Low Effort137 ```1381393. **Identify Quick Wins** (High Impact + Low Effort) for immediate action1404. **Plan Strategic Projects** (High Impact + High Effort) with proper roadmap1415. **Suggest specific improvements** for top priority items:142 - Load `references/error_ux_best_practices.md` for design patterns143 - Draft improved error messages (before/after)144 - Design improved recovery flows145 - Estimate improvement impact using `references/cx_metrics_reference.md`146147### Workflow 5: Report Generation(レポート出力)148149Compile all analysis results into a comprehensive CX Error Analysis Report.1501511. **Load report template** from `assets/cx_error_report_template.md`1522. **Populate all sections**:153 - Executive summary with key findings154 - Complete error scenario inventory155 - Multi-axis evaluation results for every scenario156 - CX Score ranking157 - Priority matrix visualization158 - Quick Wins list with specific action items159 - Detailed improvement proposals (top scenarios)160 - Expected benefits (support cost reduction, CSAT improvement, churn reduction)1613. **Include quantitative estimates** where possible:162 - Use `references/cx_metrics_reference.md` for ROI calculations163 - Estimate support ticket reduction164 - Project CSAT/CES improvement1654. **Review and finalize** the report166167## Resources168169| File | Purpose | When to Load |170|------|---------|--------------|171| `references/cx_evaluation_methodology.md` | Detailed scoring criteria for all 6 evaluation axes, error classification taxonomy, and data collection guidance | Workflow 2 (Multi-Axis CX Evaluation) |172| `references/error_ux_best_practices.md` | Error message design principles, recovery flow patterns, emotional design, accessibility guidelines, and good/bad examples | Workflow 4 (Improvement Priority Matrix) when drafting improvements |173| `references/cx_metrics_reference.md` | CX metrics (CES, CSAT, NPS), support cost formulas, churn risk estimation, industry benchmarks, and ROI calculation | Workflow 3 (CX Scoring) and Workflow 5 (Report Generation) |174| `assets/cx_error_report_template.md` | Complete CX Error Analysis Report template with all sections and placeholders | Workflow 5 (Report Generation) |175| `assets/cx_error_evaluation_template.md` | Individual error scenario evaluation card template for per-scenario assessment | Workflow 2 (Multi-Axis CX Evaluation) for each scenario |176177---178179## Best Practices180181### Do's182- 実際のエラーメッセージテキストをそのまま記録する(推測しない)183- ユーザージャーニーの文脈でエラーを評価する(同じエラーでも決済時とブラウジング時で影響度が異なる)184- 改善提案には必ず「改善前→改善後」のメッセージ比較を含める185- ROI計算にサポートコスト削減とチャーンリスク低減の両方を含める186- Quick Winsを最優先で特定する(小さな改善でも即効性が高い)187188### Don'ts189- 技術的な正確性だけでエラーを評価しない(CX観点が主軸)190- 頻度データなしにスコアリングしない(推定でも頻度を見積もる)191- 感情的影響を過小評価しない(ユーザーの不安・焦りは離脱に直結する)192- 改善コストの見積もりを省略しない(Impact vs Effort マトリクスに必須)193194---195196## Examples197198### 例1: ECサイトのエラーCX分析199200```201User: ECサイトのエラー体験を改善したい。主要なエラーシナリオを分析してください。202203Claude:2041. [Workflow 1] エラー棚卸し → ERR-001(在庫切れ), ERR-002(決済失敗),205 ERR-003(セッションタイムアウト), ERR-004(住所バリデーション), ...2062. [Workflow 2] 6軸評価 → ERR-002(決済失敗)が最高スコア4.2: 影響度5, 復旧困難4, 感情52073. [Workflow 3] CXスコア計算 → ERR-002: 4.2(Critical), ERR-003: 3.5(Significant), ...2084. [Workflow 4] 優先度マトリクス → Quick Win: ERR-004(メッセージ改善のみ、工数Low)2095. [Workflow 5] レポート出力: 改善提案8件、推定サポートコスト削減20%210```211212### 例2: SaaSログイン周りのエラーUX改善213214```215User: ログイン・認証エラーのUXを改善したい。216217Claude:2181. [Workflow 1] 棚卸し → ERR-001(パスワード誤り), ERR-002(アカウントロック),219 ERR-003(2FA失敗), ERR-004(SSO連携エラー)2202. [Workflow 2] 各シナリオを評価 → ERR-002のメッセージ品質が5(最低: "Account locked"のみ)2213. [Workflow 4] Quick Win: ERR-002のメッセージ改善222 Before: "Account locked. Contact support."223 After: "セキュリティのためアカウントがロックされました。30分後に自動解除されます。224 すぐにログインが必要な場合はパスワードリセットをお試しください。"225```