# 統計的根拠を組み込んだストーリーテリングの実例 import pandas as pd import numpy as np import matplotlib.pyplot as plt import scipy.stats as stats # 日本語フォント設定 plt.rcParams[‘font.sans-serif’] = [‘DejaVu Sans’] print(“=” * 80) print(“【Case Study: E-commerce Website Redesign Proposal】”) print(“【統計的根拠を組み込んだプレゼンテーション】”) print(“=” * 80) print() # ストーリー構造 print(“【STORY STRUCTURE】”) print() print(“1. Context: Why this matters”) print(“2. Data: What we found (with statistical evidence)”) print(“3. Insight: What it means”) print(“4. Action: What we should do”) print(“5. Impact: What we’ll achieve”) print() print(“=” * 80) print() # 1. CONTEXT(文脈) print(“【1. CONTEXT: The Business Challenge】”) print() print(“Our e-commerce conversion rate has plateaued at 2.3%”) print(“while industry leaders achieve 4-5%.”) print() print(“Key Question: Can website redesign improve conversion?”) print() print(“Stakeholder Concern:”) print(” • Redesign cost: ¥50 million”) print(” • Risk of losing familiar customers”) print(” • Unclear ROI”) print() print(“Our Approach:”) print(” • Conducted A/B test with new design (2,000 users)”) print(” • Ran for 4 weeks to ensure statistical validity”) print(” • Measured conversion rate, revenue, and user engagement”) print() print(“=” * 80) print() # 2. DATA(データと統計的根拠) print(“【2. DATA: A/B Test Results with Statistical Evidence】”) print() # A/Bテストデータ np.random.seed(42) # Control group (現行デザイン) n_control = 1000 conversions_control = 23 # 2.3% revenue_control = np.random.normal(15000, 3000, conversions_control) # Treatment group (新デザイン) n_treatment = 1000 conversions_treatment = 38 # 3.8% revenue_treatment = np.random.normal(16500, 3200, conversions_treatment) # 基本統計 control_rate = conversions_control / n_control treatment_rate = conversions_treatment / n_treatment lift = (treatment_rate – control_rate) / control_rate * 100 print(f”Control (Current Design):”) print(f” • Sample size: {n_control:,} users”) print(f” • Conversions: {conversions_control} ({control_rate:.1%})”) print(f” • Avg revenue per conversion: ¥{revenue_control.mean():,.0f}”) print() print(f”Treatment (New Design):”) print(f” • Sample size: {n_treatment:,} users”) print(f” • Conversions: {conversions_treatment} ({treatment_rate:.1%})”) print(f” • Avg revenue per conversion: ¥{revenue_treatment.mean():,.0f}”) print() print(f”Lift: {lift:.1f}% improvement in conversion rate”) print() # 統計的検定 print(“=” * 80) print(“【STATISTICAL VALIDATION】”) print(“=” * 80) print() # 2標本比率検定 pooled_prob = (conversions_control + conversions_treatment) / (n_control + n_treatment) se = np.sqrt(pooled_prob * (1 – pooled_prob) * (1/n_control + 1/n_treatment)) z_score = (treatment_rate – control_rate) / se p_value = 2 * (1 – stats.norm.cdf(abs(z_score))) print(“Two-Proportion Z-Test:”) print(f” • Z-score: {z_score:.3f}”) print(f” • P-value: {p_value:.4f}”) print(f” • Significance level: α = 0.05″) print() if p_value < 0.05: print("✓ STATISTICALLY SIGNIFICANT (p < 0.05)") print(" → We can be 95% confident the improvement is real, not random.") else: print("✗ NOT statistically significant") print() # 信頼区間 ci_lower = (treatment_rate - control_rate) - 1.96 * se ci_upper = (treatment_rate - control_rate) + 1.96 * se print("95% Confidence Interval for the difference:") print(f" • Lower bound: {ci_lower:.1%}") print(f" • Upper bound: {ci_upper:.1%}") print() print(f" → We are 95% confident the true improvement is between") print(f" {ci_lower:.1%} and {ci_upper:.1%}") print() # 効果サイズ(Cohen's h) h = 2 * (np.arcsin(np.sqrt(treatment_rate)) - np.arcsin(np.sqrt(control_rate))) print(f"Effect Size (Cohen's h): {h:.3f}") if abs(h) < 0.2: effect_interpretation = "Small effect" elif abs(h) < 0.5: effect_interpretation = "Medium effect" else: effect_interpretation = "Large effect" print(f" → {effect_interpretation}") print() print("=" * 80) print() # 3. INSIGHT(洞察) print("【3. INSIGHT: What This Means for Business】") print() print("Key Findings:") print() print(f"1. Conversion Rate Improvement: +{lift:.1f}%") print(f" • Current: {control_rate:.1%}") print(f" • With new design: {treatment_rate:.1%}") print(f" • Statistical confidence: 95% (p={p_value:.4f})") print() print("2. Revenue Impact Analysis:") monthly_visitors = 50000 current_conversions = monthly_visitors * control_rate new_conversions = monthly_visitors * treatment_rate additional_conversions = new_conversions - current_conversions avg_order_value = 15500 monthly_revenue_increase = additional_conversions * avg_order_value annual_revenue_increase = monthly_revenue_increase * 12 print(f" • Current monthly conversions: {current_conversions:.0f}") print(f" • Expected with new design: {new_conversions:.0f}") print(f" • Additional conversions: +{additional_conversions:.0f}/month") print() print(f" • Monthly revenue increase: ¥{monthly_revenue_increase:,.0f}") print(f" • Annual revenue increase: ¥{annual_revenue_increase:,.0f}") print() print("3. User Behavior Insights:") print(" • Time on site increased by 25%") print(" • Cart abandonment decreased from 68% to 52%") print(" • Mobile conversion improved most (+85%)") print() print("=" * 80) print() # 4. ACTION(提案) print("【4. ACTION: Our Recommendation】") print() print("RECOMMENDATION: Implement the new design") print() print("Investment Required:") print(f" • Design implementation: ¥50,000,000") print(f" • Expected annual return: ¥{annual_revenue_increase:,.0f}") print(f" • Payback period: {50000000/annual_revenue_increase*12:.1f} months") print(f" • ROI (Year 1): {(annual_revenue_increase-50000000)/50000000*100:.0f}%") print() print("Implementation Plan:") print() print("Phase 1 (Month 1-2): Preparation") print(" • Finalize design specifications") print(" • Set up tracking and analytics") print(" • Prepare rollback plan") print() print("Phase 2 (Month 3): Gradual Rollout") print(" • Week 1: 10% of traffic") print(" • Week 2: 25% of traffic") print(" • Week 3: 50% of traffic") print(" • Week 4: 100% if metrics stable") print() print("Phase 3 (Month 4-6): Optimization") print(" • Monitor performance daily") print(" • A/B test individual components") print(" • Continuous improvement") print() print("Risk Mitigation:") print(" • Gradual rollout allows quick rollback") print(" • Keep current design as backup") print(" • 24/7 monitoring during transition") print() print("=" * 80) print() # 5. IMPACT(期待される効果) print("【5. IMPACT: Expected Outcomes】") print() print("Financial Impact (Year 1):") print(f" • Revenue increase: ¥{annual_revenue_increase:,.0f}") print(f" • Investment: ¥50,000,000") print(f" • Net benefit: ¥{annual_revenue_increase-50000000:,.0f}") print() print("Strategic Impact:") print(" • Improved customer experience") print(" • Competitive advantage in UX") print(" • Foundation for future optimization") print(" • Increased mobile market share") print() print("Success Metrics (to track):") print(" • Conversion rate ≥ 3.5%") print(" • Revenue growth ≥ 15%") print(" • Customer satisfaction (NPS) +10pts") print(" • Mobile conversion rate +50%") print() print("=" * 80) print() print("【CONCLUSION】") print("=" * 80) print() print("The data strongly supports implementing the new design:") print() print("✓ Statistically significant improvement (p < 0.05)") print("✓ Substantial business impact (¥{:,.0f} annual revenue)".format(annual_revenue_increase)) print("✓ Strong ROI ({:.0f}% in Year 1)".format((annual_revenue_increase-50000000)/50000000*100)) print("✓ Quick payback ({:.1f} months)".format(50000000/annual_revenue_increase*12)) print("✓ Low risk with gradual rollout strategy") print() print("RECOMMENDATION: Proceed with implementation")

# プレゼンテーション用の可視化 import matplotlib.pyplot as plt import numpy as np import scipy.stats as stats # データ設定 n_control, n_treatment = 1000, 1000 control_rate, treatment_rate = 0.023, 0.038 lift = (treatment_rate – control_rate) / control_rate * 100 monthly_visitors = 50000 avg_order_value = 15500 monthly_revenue_increase = (monthly_visitors * treatment_rate – monthly_visitors * control_rate) * avg_order_value annual_revenue_increase = monthly_revenue_increase * 12 # 統計値計算 pooled_prob = (23 + 38) / (n_control + n_treatment) se = np.sqrt(pooled_prob * (1 – pooled_prob) * (1/n_control + 1/n_treatment)) z_score = (treatment_rate – control_rate) / se p_value = 2 * (1 – stats.norm.cdf(abs(z_score))) ci_lower = (treatment_rate – control_rate) – 1.96 * se ci_upper = (treatment_rate – control_rate) + 1.96 * se # 可視化 fig = plt.figure(figsize=(16, 12)) # 1. A/Bテスト結果の比較 ax1 = plt.subplot(3, 3, 1) groups = [‘Current\nDesign’, ‘New\nDesign’] rates = [control_rate * 100, treatment_rate * 100] colors = [‘#95a5a6’, ‘#667eea’] bars = ax1.bar(groups, rates, color=colors, alpha=0.8, edgecolor=’black’, linewidth=2) ax1.set_ylabel(‘Conversion Rate (%)’, fontsize=11, fontweight=’bold’) ax1.set_title(‘A/B Test Results\nConversion Rate Comparison’, fontsize=12, fontweight=’bold’) ax1.set_ylim([0, 5]) ax1.grid(axis=’y’, alpha=0.3) # 数値表示 for bar, rate in zip(bars, rates): height = bar.get_height() ax1.text(bar.get_x() + bar.get_width()/2., height, f'{rate:.1f}%’, ha=’center’, va=’bottom’, fontsize=14, fontweight=’bold’) # 改善率表示 ax1.annotate(f’+{lift:.1f}%’, xy=(1, treatment_rate*100), xytext=(0.5, 4.5), arrowprops=dict(arrowstyle=’->’, color=’red’, lw=2), fontsize=12, fontweight=’bold’, color=’red’, ha=’center’) # 2. 統計的有意性の可視化 ax2 = plt.subplot(3, 3, 2) x = np.linspace(-4, 4, 1000) y = stats.norm.pdf(x, 0, 1) ax2.plot(x, y, ‘k-‘, linewidth=2, label=’Null Distribution’) ax2.fill_between(x[x <= -1.96], y[x <= -1.96], alpha=0.3, color='red', label='Rejection Region') ax2.fill_between(x[x >= 1.96], y[x >= 1.96], alpha=0.3, color=’red’) ax2.axvline(z_score, color=’blue’, linestyle=’–‘, linewidth=2, label=f’Our Z-score: {z_score:.2f}’) ax2.set_xlabel(‘Z-score’, fontsize=11) ax2.set_ylabel(‘Probability Density’, fontsize=11) ax2.set_title(f’Statistical Significance Test\np-value = {p_value:.4f}’, fontsize=12, fontweight=’bold’) ax2.legend(fontsize=9) ax2.grid(alpha=0.3) # 3. 信頼区間の可視化 ax3 = plt.subplot(3, 3, 3) difference = (treatment_rate – control_rate) * 100 ci_lower_pct = ci_lower * 100 ci_upper_pct = ci_upper * 100 ax3.errorbar([1], [difference], yerr=[[difference – ci_lower_pct], [ci_upper_pct – difference]], fmt=’o’, markersize=12, color=’#667eea’, ecolor=’#667eea’, capsize=10, capthick=3, linewidth=3) ax3.axhline(y=0, color=’red’, linestyle=’–‘, linewidth=2, alpha=0.5, label=’No difference’) ax3.set_xlim([0.5, 1.5]) ax3.set_ylabel(‘Difference in Conversion Rate (%)’, fontsize=11) ax3.set_title(‘95% Confidence Interval’, fontsize=12, fontweight=’bold’) ax3.set_xticks([1]) ax3.set_xticklabels([‘Treatment – Control’]) ax3.grid(axis=’y’, alpha=0.3) ax3.legend(fontsize=9) # 4. 月次収益インパクト ax4 = plt.subplot(3, 3, 4) months = range(1, 13) current_monthly = [monthly_visitors * control_rate * avg_order_value / 1000000] * 12 new_monthly = [monthly_visitors * treatment_rate * avg_order_value / 1000000] * 12 ax4.plot(months, current_monthly, ‘o–‘, label=’Current Design’, color=’#95a5a6′, linewidth=2, markersize=8) ax4.plot(months, new_monthly, ‘o-‘, label=’New Design’, color=’#667eea’, linewidth=2, markersize=8) ax4.fill_between(months, current_monthly, new_monthly, alpha=0.2, color=’green’) ax4.set_xlabel(‘Month’, fontsize=11) ax4.set_ylabel(‘Revenue (Million JPY)’, fontsize=11) ax4.set_title(‘Monthly Revenue Projection’, fontsize=12, fontweight=’bold’) ax4.legend(fontsize=10) ax4.grid(alpha=0.3) # 5. 累積収益差 ax5 = plt.subplot(3, 3, 5) cumulative_benefit = [(i+1) * monthly_revenue_increase / 1000000 for i in range(12)] ax5.plot(months, cumulative_benefit, ‘o-‘, label=’Cumulative Benefit’, color=’#27ae60′, linewidth=3, markersize=8) ax5.axhline(y=50, color=’red’, linestyle=’–‘, linewidth=2, label=’Investment (¥50M)’) ax5.fill_between(months, 0, cumulative_benefit, alpha=0.2, color=’green’) # 損益分岐点 breakeven_month = 50 / (monthly_revenue_increase / 1000000) ax5.axvline(x=breakeven_month, color=’orange’, linestyle=’:’, linewidth=2, label=f’Breakeven ({breakeven_month:.1f} months)’) ax5.set_xlabel(‘Month’, fontsize=11) ax5.set_ylabel(‘Amount (Million JPY)’, fontsize=11) ax5.set_title(‘Cumulative Benefit vs Investment’, fontsize=12, fontweight=’bold’) ax5.legend(fontsize=9) ax5.grid(alpha=0.3) # 6. ROIの推移 ax6 = plt.subplot(3, 3, 6) roi_values = [((i+1) * monthly_revenue_increase – 50000000) / 50000000 * 100 for i in range(12)] bars = ax6.bar(months, roi_values, color=[‘red’ if r < 0 else 'green' for r in roi_values], alpha=0.7, edgecolor='black') ax6.axhline(y=0, color='black', linewidth=1) ax6.set_xlabel('Month', fontsize=11) ax6.set_ylabel('ROI (%)', fontsize=11) ax6.set_title('Return on Investment Over Time', fontsize=12, fontweight='bold') ax6.grid(axis='y', alpha=0.3) # 7. コンバージョンファネル比較 ax7 = plt.subplot(3, 3, 7) funnel_stages = ['Visitors', 'Product\nViews', 'Add to\nCart', 'Checkout', 'Purchase'] current_funnel = [100, 60, 30, 10, 2.3] new_funnel = [100, 70, 42, 15, 3.8] x_pos = np.arange(len(funnel_stages)) width = 0.35 bars1 = ax7.bar(x_pos - width/2, current_funnel, width, label='Current', color='#95a5a6', alpha=0.7) bars2 = ax7.bar(x_pos + width/2, new_funnel, width, label='New Design', color='#667eea', alpha=0.7) ax7.set_ylabel('Percentage (%)', fontsize=11) ax7.set_title('Conversion Funnel Comparison', fontsize=12, fontweight='bold') ax7.set_xticks(x_pos) ax7.set_xticklabels(funnel_stages, fontsize=9) ax7.legend(fontsize=10) ax7.grid(axis='y', alpha=0.3) # 8. デバイス別改善 ax8 = plt.subplot(3, 3, 8) devices = ['Desktop', 'Mobile', 'Tablet'] improvement = [14.3, 83.3, 25.0] bars = ax8.bar(devices, improvement, color=['#667eea', '#e74c3c', '#f39c12'], alpha=0.7, edgecolor='black') ax8.set_ylabel('Improvement (%)', fontsize=11) ax8.set_title('Conversion Rate Improvement by Device', fontsize=12, fontweight='bold') ax8.grid(axis='y', alpha=0.3) # 数値表示 for bar, imp in zip(bars, improvement): height = bar.get_height() ax8.text(bar.get_x() + bar.get_width()/2., height, f'+{imp:.0f}%', ha='center', va='bottom', fontsize=11, fontweight='bold') # 9. シナリオ分析 ax9 = plt.subplot(3, 3, 9) scenarios = ['Pessimistic\n(90%)', 'Base Case', 'Optimistic\n(120%)'] returns = [annual_revenue_increase * 0.9 / 1000000, annual_revenue_increase / 1000000, annual_revenue_increase * 1.2 / 1000000] colors_scenario = ['#e74c3c', '#f39c12', '#27ae60'] bars = ax9.bar(range(len(scenarios)), returns, color=colors_scenario, alpha=0.7, edgecolor='black', linewidth=2) ax9.axhline(y=50, color='red', linestyle='--', linewidth=2, label='Investment (¥50M)', alpha=0.7) ax9.set_ylabel('Annual Return (Million JPY)', fontsize=11) ax9.set_title('Scenario Analysis', fontsize=12, fontweight='bold') ax9.set_xticks(range(len(scenarios))) ax9.set_xticklabels(scenarios, fontsize=9) ax9.legend(fontsize=9) ax9.grid(axis='y', alpha=0.3) plt.tight_layout() plt.savefig('storytelling_presentation.png', dpi=150, bbox_inches='tight') plt.show()

# 反論対処を組み込んだプレゼンテーション print(“=” * 80) print(“【Addressing Potential Objections】”) print(“=” * 80) print() objections = { “Objection 1”: { “concern”: “Sample size is too small”, “data_response”: “Power analysis shows 80% power to detect 1.5% difference”, “evidence”: “n=2,000 exceeds minimum required n=1,500”, “conclusion”: “Sample size is statistically sufficient” }, “Objection 2”: { “concern”: “Could be seasonal effect”, “data_response”: “A/B test ran simultaneously – both groups affected equally”, “evidence”: “Year-over-year comparison shows no seasonal anomaly”, “conclusion”: “Seasonal effects are controlled” }, “Objection 3”: { “concern”: “Effect may be short-term (novelty bias)”, “data_response”: “4-week duration captures initial novelty decay”, “evidence”: “Week 4 conversion still 3.6% (vs Week 1: 4.0%)”, “conclusion”: “Effect is durable, with planned long-term monitoring” }, “Objection 4”: { “concern”: “Investment cost is too high”, “data_response”: “ROI analysis shows 100%+ return in Year 1”, “evidence”: “Payback period: 2.8 months”, “conclusion”: “Investment will be recovered quickly” }, “Objection 5”: { “concern”: “Implementation risk is high”, “data_response”: “Gradual rollout minimizes risk”, “evidence”: “10% → 25% → 50% → 100% over 4 weeks”, “conclusion”: “Rollback possible within 24 hours at any stage” } } for obj_name, details in objections.items(): print(f”【{obj_name}】”) print(f” Concern: \”{details[‘concern’]}\””) print() print(f” Our Response:”) print(f” • Data: {details[‘data_response’]}”) print(f” • Evidence: {details[‘evidence’]}”) print(f” • Conclusion: {details[‘conclusion’]}”) print() print(“=” * 80) print() print(“【Summary: Why This Proposal is Low-Risk】”) print() print(“✓ Statistically rigorous testing (p < 0.05)") print("✓ Adequate sample size (n=2,000)") print("✓ Controlled for external factors") print("✓ Strong financial returns (ROI > 100%)”) print(“✓ Gradual rollout with rollback capability”) print() print(“All major concerns have been addressed with data.”)

# 複数分析を統合したプレゼンテーション print(“=” * 80) print(“【Integrated Analysis: Review Promotion Strategy】”) print(“=” * 80) print() print(“【Three Independent Analyses Supporting One Conclusion】”) print() # 分析1: 回帰分析 print(“┌” + “─” * 60 + “┐”) print(“│ Analysis 1: Regression Analysis │”) print(“├” + “─” * 60 + “┤”) print(“│ │”) print(“│ Question: What drives sales? │”) print(“│ │”) print(“│ Findings: │”) print(“│ • Review count has highest impact (β = 0.42, p < 0.001) │") print("│ • Each review adds +3.2% sales on average │") print("│ • Effect is statistically highly significant │") print("│ │") print("│ Implication: Reviews directly drive sales │") print("└" + "─" * 60 + "┘") print() # 分析2: コホート分析 print("┌" + "─" * 60 + "┐") print("│ Analysis 2: Cohort Analysis │") print("├" + "─" * 60 + "┤") print("│ │") print("│ Question: Do reviewers become better customers? │") print("│ │") print("│ Findings: │") print("│ • Repeat rate of reviewers: 65% │") print("│ • Repeat rate of non-reviewers: 22% │") print("│ • Reviewers are 3x more likely to return │") print("│ │") print("│ Implication: Review writing creates loyalty │") print("└" + "─" * 60 + "┘") print() # 分析3: A/Bテスト print("┌" + "─" * 60 + "┐") print("│ Analysis 3: A/B Test │") print("├" + "─" * 60 + "┤") print("│ │") print("│ Question: Can we increase review submissions? │") print("│ │") print("│ Findings: │") print("│ • ¥500 coupon incentive tested │") print("│ • Review rate: 5% → 15% (3x increase) │") print("│ • Statistical significance: p < 0.001 │") print("│ │") print("│ Implication: Reviews can be incentivized effectively │") print("└" + "─" * 60 + "┘") print() # 統合結論 print("=" * 80) print("【Integrated Conclusion: Triple Validation】") print("=" * 80) print() print("Three independent analyses converge on one strategy:") print() print(" ┌─────────────────────────────────────────────────────────┐") print(" │ │") print(" │ RECOMMENDED: Implement Review Incentive Program │") print(" │ │") print(" │ Logic Chain: │") print(" │ │") print(" │ Incentive → More Reviews → More Sales + Loyalty │") print(" │ ↑ ↑ ↑ │") print(" │ (A/B Test) (Regression) (Cohort) │") print(" │ │") print(" │ All three analyses support this conclusion │") print(" │ │") print(" └─────────────────────────────────────────────────────────┘") print() # ビジネスインパクト print("【Projected Business Impact】") print() print("Current State:") print(" • Monthly purchases: 10,000") print(" • Review rate: 5% (500 reviews/month)") print(" • Monthly revenue: ¥100,000,000") print() print("After Implementation:") print(" • Review rate: 15% (1,500 reviews/month)") print(" • Additional reviews: +1,000/month") print(" • Sales increase: +3.2% per review = +32,000,000/month") print() print("ROI Calculation:") print(" • Annual revenue increase: ¥384,000,000") print(" • Annual cost (coupons + system): ¥38,000,000") print(" • Net annual benefit: ¥346,000,000") print(" • ROI: 910%")

# 時間別プレゼンテーション構成 print(“=” * 80) print(“【Presentation Formats by Time Allocation】”) print(“=” * 80) print() formats = { “1-Minute (Elevator Pitch)”: { “structure”: [ “Conclusion only (30 sec)”, “Key number (15 sec)”, “Ask (15 sec)” ], “example”: ‘”We should implement redesign. It will increase revenue by ¥139M/year with 100% ROI. I need your approval to proceed.”‘ }, “5-Minute (Executive Summary)”: { “structure”: [ “Conclusion (30 sec)”, “Key data points x3 (2 min)”, “Recommendation (1 min)”, “Ask + Q&A (1.5 min)” ], “example”: “Focus on business impact and ROI. Technical details in appendix.” }, “15-Minute (Standard)”: { “structure”: [ “Context (2 min)”, “Methodology overview (2 min)”, “Results (5 min)”, “Recommendation + ROI (4 min)”, “Q&A (2 min)” ], “example”: “Balance between detail and clarity. Show key visualizations.” }, “30-Minute (Deep Dive)”: { “structure”: [ “Full context (3 min)”, “Detailed methodology (5 min)”, “Comprehensive results (10 min)”, “Recommendation + scenarios (7 min)”, “Implementation plan (3 min)”, “Q&A (2 min)” ], “example”: “Include statistical validation, alternative approaches, risk analysis.” } } for format_name, details in formats.items(): print(f”【{format_name}】”) print() print(“Structure:”) for i, item in enumerate(details[‘structure’], 1): print(f” {i}. {item}”) print() print(f”Note: {details[‘example’]}”) print() print(“-” * 60) print() print(“【Key Principle】”) print() print(“Always prepare multiple versions!”) print(” • You never know how much time you’ll actually get”) print(” • Mark ‘CORE’ slides that must be shown”) print(” • Have detailed slides ready for Q&A”)