Online Social Media as a Publicity Platform Kai-Lung Hui, PhD Department of Information Systems, Business Statistics, and Operations Management Hong Kong University of Science & Technology

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

1

Nature of Online Social Media • Web 2.0 – often user-driven • Social multiplier effect – People directly affect others’ behavior through communication (aka “word of mouth”) – Friendship networks – A person’s behavior conveys useful information to others (aka “observational learning”) – Herding (“let me follow others”) Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

2

Why Publicity in Online Social Media? • A new channel of reaching consumers – Wide reach – Often cheaper; so, good for niche products

• Less biased – cf. advertising and sales promotions

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

3

Wide Reach – China • • • • • •

TV households: Number of magazines: Number of newspapers: Number of publishing houses: Internet subscribers: Broadband penetration:

Source: CMM Intelligence

Spanish Chamber of Commerce, March 28, 2012

396 million 9,468 1,938 578 256 million 85%

How to advertise? Which channel to select, and effectiveness? (c) K.L. Hui, 2012

4

Good for Niche Products (1) Distribution of music album sales, and the 80/20 rule

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

5

Good for Niche Products (2) On music blogs

Source: Dewan and Ramaprasad (2012)

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

6

Motivation to Go Online – China

So, people do respond to publicity efforts

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

7

Motivation to Go Online – China

These sound very promising!

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

8

Motivation to Go Online – China

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

9

Key Outstanding Issues • To what extent does the interaction in online social media help a business? • Is the effect similar to traditional media? • How can a company make use of online social media to promote its business? Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

10

Impact of Online Social Media: Some Evidence • Dewan and Ramaprasad (2012) – How blogs on The Hype Machine affected music album sales on Amazon.com

• Music sampling is positively associated with both blog and music popularity • The effect is stronger for: – newly released songs – the “long tail” than the “body” Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

11

The Long Tail

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

12

Impact of Online Social Media: Some Evidence • Rui et al. (2012) – How Tweets affect movie sales

• The chatters on Tweeter mattered – Positive Tweeter word of mouth (WOM) increased movie sales; negative WOM decreased movie sales

• Impact of WOM on movie sales is stronger for: – influential users (measured by number of followers) than less influential users

• The authors of the Tweets also mattered! Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

13

Author Effect

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

14

Impact of Online Social Media: Some Evidence • Wang et al. (2010) – How media publicity and online WOM affected movie sales in China

• Chinese consumers are more likely to rely on interpersonal communications by social media than U.S. consumers – Both media publicity on Sina.com (新浪网) and online WOM on Baidu Tieba (百度贴吧) increased the probability for consumers to watch a movie – Online WOM on Baidu Tieba (百度贴吧) increased the probability for consumers to follow others Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

15

Impact of Online Social Media: Some Evidence • Movie sales in China

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

16

Problems and Challenges • How do we know that publicity campaigns on online social media helped? – Album sales and blog posts – Movie revenue and tweets –…

• Alternative explanations – Advertising or general media publicity in the same period – Other unobserved attributes (e.g., popular artists) – Reverse causation Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

17

More Indicative Methods • Field experiments – Better control of other factors, such as artist popularity or other publicity – Allow for focused testing of interested effects (e.g., friendship influences) – The findings would apply directly to interested companies (and probably businesses in the same or similar industries) Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

18

Friendship Influence: An Experiment • Setting: – SMS seeding service in Guangzhou, China – Messages range from jokes, greeting message, news, to household tips

• Sample: – 300,000 users of the company, and their friends – None of these users were “friends” in our preexperimental period Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

19

Experimental Design

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

20

Findings (Stage 1) Average reply rates

Joke Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

21

Findings (Stage 1) Average forward rates

Nothing

Spanish Chamber of Commerce, March 28, 2012

Friend greeting

Friend joke

(c) K.L. Hui, 2012

System greeting

System joke

22

Alternative Explanations • Maybe the friends have merely helped us identify the more active users • How to eliminate user characteristics and identify friendship influences? – Ans: Repeat experiment in Stage 2 with treatments reversed

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

23

Findings (Stages 1 and 2) VARIABLES period Other inbound messages Seeded Seeded * Joke Seeded * Friend Seeded * Joke * Friend

(1) GR: M2 -0.207*** (0.006) 0.009 (0.006) -0.009 (0.008) -0.011 (0.013) 0.019 (0.081) -0.124 (0.142)

(2) GR: Other -2.931*** (0.052) 0.275** (0.122) -0.086* (0.048) -0.089 (0.080) 1.187** (0.472) -3.777*** (1.259)

(3) GR: All -3.138*** (0.054) 0.283** (0.126) -0.094* (0.051) -0.100 (0.084) 1.206** (0.497) -3.901*** (1.278)

Greeting from friend

7.950*** (0.876) -12.857*** (0.958) -0.138*** (0.054) -0.161*** (0.061) 3.658*** (0.044)

8.336*** (0.928) -13.628*** (1.028) -0.156*** (0.057) -0.180*** (0.064) 3.922*** (0.046)

600000 0.033 300000

600000 0.034 300000

3.793*** (0.036)

4.064*** (0.037)

600000 0.004 300000

600000 0.023 300000

600000 0.024 300000

600000 0.006 300000

Joke from system

Commerce, March 28, 2012

(6) Actual: All -2.813*** (0.051) 0.300** (0.133)

0.270*** (0.004)

Greeting from system

Observations R-squared Number of focal users Spanish Chamber of

(5) Actual: Other -2.625*** (0.049) 0.289** (0.129)

0.387*** (0.141) -0.771*** (0.193) -0.018* (0.010) -0.019* (0.010) 0.265*** (0.006)

Joke from friend

Constant

(4) Actual: M2 -0.188*** (0.006) 0.010* (0.006)

(c) K.L. Hui, 2012

24

Viral Marketing • Viral index V = N * Cr – N: Number of friends that a user forward the message to – Cr: Conversion rate

• Objective: V > 1

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

25

Viral Marketing

Source: Kalyanam et al. (2007) Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

26

Example – Plaxo

Source: Kalyanam et al. (2007)

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

27

Who Should We Approach and Affect? • Common belief – the “influentials” (opinion leaders), or hyper influentials • Findings (Watts and Dodds 2007)

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

28

Implications • Except for hyper influential users, general “opinion leaders” may not help more than having a large network of users • So, the objective should be to reach out to as many friends in a network as possible!

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

29

Social Media in China

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

30

Some Weibo Patterns • Please refer to: http://blog.sina.com.cn/s/blog_672349b3010 0sjg3.html

American Chamber of Commerce, March 21, 2012

(c) K.L. Hui, 2012

31

Overall Picture • Online social media is attractive as a publicity platform because: – It is relatively “cheap” – It has wide reach to consumers – It helps a firm leverage on friendship networks, or others’ influences

• Key managerial issues: – Does it really have an impact on driving sales? – Is the effect similar to traditional media? – What should firms pay attention to when using online social media as a publicity platform? Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

32

On Impacts • Abundant studies that show a positive impact – – – –

Blogs helped music sales Tweets helped movie sales Artists’ posts in MySpace affected album sales Blogs helped movie sales in China

• Chinese seem to particularly rely on interpersonal communications in online social media • But, we have to be careful about causality – some of these findings could be spurious Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

33

Nature of the Impacts • The impact of online social media could be asymmetric across products – Perhaps more helpful for niche products, the socalled “long tail” – For mainstream products, the benefit is less apparent

• Authorship also matters – Hyper-influential authors may particularly help! Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

34

Friendship Effects • There seems to be a robust influence due to “friends” – Promise of viral marketing

• As long as the referral or advice comes from a friend, whether the friend is influential or not may not make a huge difference – V = N * Cr

Spanish Chamber of Commerce, March 28, 2012

The traditional focus (c) K.L. Hui, 2012

Maybe we should focus on the network size instead … 35

Finding the Right Networks in China • Blogs, instant messaging, and BBS are all good • People particularly like to share and seek advice, and they listen to others more than the company/”authority” • Time profile of access is important – Wednesday and Thursday – Lunch and late evening hours – Catching the wave is important! Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

36

Questions/Comments • Contact information: – Kai-Lung Hui, Dept of ISOM, HKUST, Hong Kong – Email: [email protected]

Spanish Chamber of Commerce, March 28, 2012

(c) K.L. Hui, 2012

37