of 20

A case study based on hypothetical data of Facebook pages of past U.S. presidents.
Published on: Mar 4, 2016
Published in: Social Media      Technology      Business

#### Transcripts - Presidents on Facebook

• 1. PRESIDENTS ON FACEBOOK L2 CASE STUDY QIUYUN TAN COLUMBIA UNIVERSITY
• 2. Imagine the past US presidents are on Facebook…
• 3. Question: What are the interaction rates for each President’s Facebook page?
• 4. No.1 Adams: 4.57 Factors Influencing Interaction Rates No.2 Washington: 4.18 •Pages with relatively small fan bases tend to have high No.3 Jackson: 3.29 interaction rates. • Having large numbers of comments and likes does not necessarily lead to high interaction rates. • Among the most “interactive” pages, some have extremely high interaction rate for one post, some have average relatively high interaction rates for all the posts.
• 5. Variables Activities of Page Audience Reaction 1) # Page Likes 2) # Fan Comments 3) # Fan Likes Used to calculate interaction rates
• 6. Definition  Conceptual Definition: Engagement rate (interaction rate) is the percentage of people who saw a post that liked, shared, clicked or commented on it. —— Facebook  Mathematical Definition: Number of Comments + Likes Interaction Rate = Total Fan Number
• 7. Assumptions    The fan numbers grow in a linear pattern between date 1, Apr 25, 2011 and date 2, Jun 20, 2011. The five posts of each president are equally spread between the two dates. Thus, we have: Posts Post 1 Post 2 Post 3 Post 4 Post 5 Time 11-4-25 11-5-9 11-5-23 11-6-6 11-6-20
• 8. Total Fan Number • According to the previous assumptions, we can calculate the fan numbers of the ten presidents when each post is put up. • The fan numbers of all the pages increase over the course of five posts. • At date 1, Van Buren (6,780,116), Polk (4,103,477), and Harrison (1,531,658) have significantly larger fan numbers than the other presidents (ranging from 9550 – 356,006).
• 9. Number of Comments and Likes 1. Van Buren, Harrison, and Washington all have one post that is liked and commented by far more people than the others posts. 2. The other presidents’ pages have approximately consistent number of comments and likes.
• 10. Interaction Rates Rank President 1 2 3 4 5 6 7 8 9 10 Adams Washington Jackson Tyler Monroe Jefferson Madison Polk Van Buren Harrison Average Interaction Rates 4.57 4.19 3.29 3.14 3.04 2.43 1.34 1.19 0.52 0.51 1. Washington, Adams, and Tyler rank high because of their extremely high interaction rates of one post. 2. Jackson and Monroe rank high because they have relatively high interaction rates for all the posts.
• 11. Relationship between Interaction Rates and Initial Fan Numbers There is a negative relationship between the size of initial fan numbers and the interaction rates, with a relatively large coefficient of determination.
• 12. Relationship between Interaction Rates and Sum of Comments and Likes There is a negative relationship between the sum of comments and likes and the interaction rates, but the coefficient of determination is smaller.
• 13. Summary • To sum up, the observations from previous calculation and linear regression show two factors that affect interaction rates: 1. Audience base Pages with relatively small fan bases tend to have high interaction rates. 2. Sum of Comments and Likes Having large numbers of comments and likes does not necessarily lead to high interaction rates.
• 14. Summary There are two scenarios where pages achieve high interaction rates: 1. Extremely high interaction rate on one post 2. Relatively high interaction rates on average
• 15. Features    None of the pages used feature 4. The ranking of Polk and Madison shows that, by not using any features, they fall into the lower half of interaction rates among the ten pages. There is not enough information to decide how the use of the other features affect the interaction rates.
• 16. Posts Per Week • Linear Regression shows that the coefficient of determination is very small (R2=0.00355). • Therefore, there is not an obvious linear relationship between the posts per week and the interaction rates.
• 17. Appendix • Calculation of Fan Numbers: • Step 1: # Likes of Date 2 - # Likes of Date 1 • Step 2: Divide the difference by 5 • Step 3: Add the result of step 3 for each post
• 18. • Fan Numbers of Ten Presidents Over Time: Post Post 1 11-4-25 Post 2 11-5-9 Post 3 11-5-23 Post 4 11-6-6 Post 5 11-6-20 Fan Base Increase between any two posts Jackson 55970 61203.75 66437.5 71671.25 76905 5233.75 Van Buren 6780116 7511840.75 7755749 243908.25 Washington 356006 367699 379392 391085 402778 11693 Jefferson 13975 14404.25 14833.5 15262.75 15692 429.25 Tyler 11298 11608.25 11918.5 12228.75 12539 310.25 Monroe 27625 28144 28663 29182 29701 519 Harrison 1531658 1559418 1587178 1614938 1642698 27760 Adams 9550 9720 9890 10060 10230 170 Polk 4103477 4171861.5 4240246 4308630.5 4377015 68384.5 Madison 258332 260763 263194 265625 268056 2431 7024024.25 7267932.5
• 19. • Linear Regression of Posts Per Week and Interaction Rates:
• 20. THE END BY QIUYUN TAN