Slide 1: Introduction

• Title: “Hypothesis Testing: Making Smarter Business Decisions”
• “Ever wondered if a new ad boosts sales or if customers prefer your product? Hypothesis testing has the answers!”
• “Today: What it is, how it works, and business examples with data.”

Slide 2: What is Hypothesis Testing?

• Definition: A statistical tool to make decisions about a business problem using data.
• Key Idea: Test a “hypothesis” (business assumption) to see if it’s true.
• Two Hypotheses:
  • Null (H₀): “Nothing’s changed” (e.g., no impact).
  • Alternative (H₁): “Something’s different” (e.g., an improvement).

Slide 3: The Steps

1. State the Hypotheses: Define H₀ and H₁.
2. Set Significance Level (α): Risk of error (usually 0.05).
3. Collect Data: Gather sample data.
4. Make a Decision: Use a test statistic or p-value.
5. Conclusion: Accept or reject H₀.

Slide 4: Key Terms

• P-value: Chance of seeing your data if H₀ is true. Small (< 0.05) → Reject H₀.
• Test Statistic: A number from your data (e.g., z-score, t-score).
• Errors:
  • Type I: Wrongly rejecting H₀.
  • Type II: Missing a real effect.

Slide 5: Example 1 - New Ad Campaign

• Scenario: Sales rise from $10,000 to $11,000 weekly after a new ad (n = 50, sd = $1,500). Did the ad work?
• Hypotheses:
  • H₀: No effect (mean = $10,000).
  • H₁: Increases sales (mean > $10,000).
• Test: t = 4.71, critical t ≈ 1.68 (α = 0.05).
• Result: Reject H₀—ad likely worked.

Slide 6: Example 2 - Product Pricing

• Scenario: Price drop of $5, sales rise from 100 to 110 units weekly (n = 30, sd = 15). Does it help?
• Hypotheses:
  • H₀: No effect (mean = 100).
  • H₁: Increases sales (mean > 100).
• Test: t = 3.65, critical t ≈ 1.7 (α = 0.05).
• Result: Reject H₀—price drop likely worked.

Slide 7: Example 3 - Customer Satisfaction

• Scenario: Loyalty program, satisfaction rises from 75% to 80% (n = 200, sd = 20%). Did it improve?
• Hypotheses:
  • H₀: No effect (proportion = 0.75).
  • H₁: Improves satisfaction (proportion > 0.75).
• Test: z = 1.63, critical z ≈ 1.645 (α = 0.05).
• Result: Fail to reject H₀—not enough evidence.

Slide 8: Why It Matters

• Marketing: Test campaign effectiveness.
• Operations: Evaluate process changes.
• Product: Assess customer preferences.
• “Turn data into dollars!”

Slide 9: Pitfalls

• P-value Misuse: Small p-value ≠ big impact.
• Sample Size: Too small = unreliable; too big = over-sensitive.
• Context: Statistical ≠ practical significance.

Slide 10: Recap

• Helps businesses decide with data.
• Steps: Hypotheses → Significance → Data → Decision → Conclusion.
• Examples: Ads (worked), pricing (worked), satisfaction (inconclusive).
• “Your data-driven crystal ball!”

Slide 11: Questions & Practice

• Question: “If a website redesign boosts session time from 2 to 2.5 minutes (n = 40), what’s H₀ and H₁?”
• Answer:
  • H₀: No effect (mean = 2 min).
  • H₁: Increases time (mean > 2 min).
• “Questions? Let’s test one together!”