This presentation was given to Techgirlz annual entrepreneurial summer camp for middle school girls. This presentation was part of a 2-hour workshop on product design, testing, and refinement.
6. The Scientific Method
1. Ask a question
2. Do background research
3. Construct a hypothesis
4. Test your hypothesis
5. Analyze your data and draw a conclusion
6. Report the results
19. Book Funnel Analysis
• English speakers on earth – 100%
• Read a business book – 45%
• Discover the product ad – 30%
• Click the product ad – 15%
• Click the “Add to cart” button – 10%
• Start the checkout process – 8%
• Enter their credit card information – 7%
• Enter their shipping information – 6%
• Buy the book – 5%
23. Budget Challenge: $???
• Given
– We have 10 names to test
– We will test each one on 10 different people
– It costs $1 for each test (i.e. to buy an ad)
• How much money do we need?
35. Sample Test Results
• Twitter converts 2% of people
• Facebook converts 10% of people
36. A/B Testing Execution
• Given 3,000 people
• A/B test for the first 300 (10% of them)
– 6 different campaigns with 50 tests each
• Then test final 2,700 people with best campaign.
• Results
– Worst campaign (Twitter) converted 2%
– Best campaign (Facebook) converted 10%
– Difference between 54 and 270 people.
– For a $50 product, that is $10,800!
37. Budget Challenge: Implementation
• Implement Twitter login: $10
• Implement Facebook login: $15
• Implement LinkedIn login: $20
• But we only have $10!?! What do we do?
39. Group Work
• Assignment [50 minutes]
– List 1-2 of your funnels, then “guess and test”
– Identify a step to optimize
– List your hypothesis and test(s)
– Test them on other group members, collect currency
• Group Presentations [30 minutes]
– Steps in the funnel(s)
– Optimization step(s)
– List your hypothesis and test(s)
– Summarize the results
Question: Why does it have two names?
Fake/exciting answer: Maybe they are trying to see what sells better?
Real/boring answer: Girl scouts doesn’t impose naming standards on its multiple bakers.
One of the most important components is the title!
Review consequences of guessing wrong: Expensive. Life or death.
Question: How would you figure out the right name?
Question: How do scientists get answers to scientific questions?
Sorry, but you still need science outside of science class!
Question: Who remembers the scientific method?
My attempt at creating a simple slogan
Question: How to validate your hypothesis?
All products need communication to understand prospective customers.
Communication is necessary to collect currency.
Questions, comments, and concerns should be addressed.
How to get validation? Collect currency from a given sample of people.
Question: What are the different forms of currency?
Question: What is another way of receiving validation?
Question: What is another way of receiving validation?
Question: What is another way of receiving validation?
Question: What is another way of receiving validation?
Clicks
Quick recap
List out all of the hypotheses we want to test
Verbal: Surveys asking if they will buy.
Signature / LOI: Asking customers to sign a form certifying that they will buy once it’s available
Pre-pay: Asking customers to buy them ahead of time
Question: IMPORTANT!!! WHO IS TOTALLY CONFUSED FROM THIS EXAMPLE?
Question: Would you do anything different for naming a book? Venue. Typically purchased online whereas girl scout cookies are not.
Question: How can you test something like this online?
Every product has a funnel like this.
We are going to optimize the “Click the product advertisement step”
Some might call this a WORKFLOW
Advanced topics: Conversion Optimization, Conversion Funnel Analysis
We are optimizing the “click the product ad”
The heart of conversion analysis, and what marketers spend a lot of their time doing
Question: Super simple 3rd grade math question for no really good reason
Discuss being convinced. Gut feeling vs true data-driven.
Sample size, confidence levels, etc
Tim Ferris actually used search advertising to test the book title
Surveys asking if they will buy.
Asking customers to sign a form certifying that they will buy once it’s available
Asking customers to buy them ahead of time
Question: What is different?
Question: Try to guess their hypothesis
Question: What is different?
Question: Try to guess their hypothesis
Every product has a funnel like this.
Some might call this a WORKFLOW
Advanced topics: Conversion Optimization, Conversion Funnel Analysis
Create a username and password
“Borrow” the login capability from another product
Explain how this is done
Question: How can we test these?
Question: Discuss the hypothesis for the “create a username” step in the funnel analysis slide.
Describe the word “convert”
Bring it back to the funnel.
Question: Explain this in detail.
Emphasize: This is just 1 campaign optimization!
Really make sure people understand.
Question: What do we do? How do we not build the functionality, and still learn from the experience.
Scope Kills!!! -- Should mention the expense required for implementation.
Solution: Have the buttons just collect the number of clicks, but redirect to the Twitter login with a message saying “Sorry, facebook isn’t implemented yet”