1. Influence of Friends On Our Purchases Consumer Behavior Experiment Rags Srinivasan @pricingright
2. Let Me First Thank These Great Folks Hiten Shah @hnshah Chris Hopf @pricing LinkedIn Grp Members @bouldernet
3. Do We Know the Influence of Our Friends on Our Purchases? There are many predictions and survey numbers There are popular metrics asking if we would recommend products to friends There is lot at stake – Marketing decisions are made on influence of friends
4. Hypothesis: We Do Not Know Biases in Measurements It’s a guess 37% to 75% Numbers reported based on surveys are suspect Responses are subject to anchoring biases
5. Experiment: Test Low and High Anchors There were two polls – respondents were randomly assigned to one In a previous survey it was found that people who are active in the social media reported 37% of their purchases are influenced by friends. Please state what percentage of your purchases are influenced by friends? In a previous survey it was found that people who are active in the social media reported 75% of their purchases are influenced by friends. Please state what percentage of your purchases are influenced by friends?
6. Population, Sampling and Tests People on Twitter and LinkedIn ~75 responses received for each poll Chi-Square test and Crosstab
7. The Raw Responses are … Low Anchor: 37% High Anchor: 75% These are so low and are ignored in the computation
8. Did a Chi-Square Test on Each Sample For the responses to be statistically significant i.e., not just random, Chi-Square should be greater than 6 (df = 2. p=0.05) Low Anchor: 37% High Anchor: 75% Computed Chi-Square= 1.32 Computed Chi-Square=0.14 What does this mean? Responses are not statistically significant. Any recommendation metric reported are not dependable
9. Did a Crosstab Between Samples For the responses to be statistically significant i.e., not just random, Chi-Square should be greater than 6 (df = 2. p=0.05) Computed Chi-Square= 0.57 What does this mean? Responses are not influenced by anchoring. It is just that people randomly pick between 0% to 50%
10. What are the Marketing Implications? Coming soon … Check out my other work on SlideShare Rags Srinivasan http://iterativepath.wordpress.com @pricingright