Social Media & Big Data:Implications for MarketersMarch 2013
What is Social Media?• Merriam Webster Online defines social media as “forms of electronic communication (as Web sites for social networking and microblogging) through which users create online communities to share information, ideas, personal messages, and other content (as videos)”
Total Users of Select Social Media Sites(March 2013)1,200,000,0001,000,000,000 800,000,000 600,000,000 400,000,000 200,000,000 -
Social Media Usage• Facebook ▫ 67% of online adults• LinkedIn ▫ 20% of online adults• Twitter ▫ 16% of online adults• Pinterest ▫ 15% of online adults
Facebook • Users ▫ 167 million unique visitors per month ▫ 500 million likes per day ▫ 24% aged 35-44 ▫ 58% women, 42% men ▫ 350 million users suffer from Facebook Addiction Syndrome • Ad policies ▫ Advertisers will be able to sync CRM database info with Facebook user info Brands will be able to more effectively target users without waiting for them to “like” the page Users can opt-out • Marketers are much more interested in data from Facebook interactions than less prevalent sites • Produces a “Gross national Happiness Index” through text mining words and phrases posted
LinkedIn • Largest professional social network • 2 new members sign up every second • 42% of users update their profile regularly • 65% Male, 35% Female • 82% of users are aware there are ads ▫ 60% have clicked • Corporate talent solutions are used by 85 Fortune 100 Companies
Twitter • Users ▫ Adults 18-29 ▫ African Americans ▫ Urban residents ▫ “The disproportionate African-American use of Twitter has fascinated culture commentators and scholars” • Ad policies ▫ Advertisers can target users based on broad categories ▫ Categories are not created from contents of tweets, but other actions and who the user follows
Pinterest • 12 million unique visitors per month • 79% women, 21% men • 29% of users aged 25-34 • Users have higher average income than Facebook & Twitter • Average time spent on Pinterest ▫ 1 hour 17 minutes
What is Big Data?• There is no pat definition for big data… In fact, big data can be relatively small, but represents a difficult processing-time issue. Basically, you’ve got big data whenever you exceed the capacity of a conventional relational database to handle it.” –Jim Davis, Senior VP/CMO at SAS ▫ Much of the data mined from social media can be considered big data, because incorporating it into current databases and CRM systems can be problematic.
Social Media Data Mining• Social media users share a considerable amount of information about themselves through their posts, likes, tweets, and connections. Social media data mining allows marketers to: ▫ Discover new niches ▫ Tailor advertisements to best meet the needs of smaller demographic groups ▫ Identify and/or predict buying patterns ▫ Manage customer issues before they become PR problems ▫ Conduct research to aid in the development of new products and services ▫ Conduct sentiment analysis
Sentiment Analysis• “Social media is the canary in the coal mine. It provides early warning of issues that can become major problems if they are not detected quickly.” –Catherin van Zuylen, VP of Products at Attensity
Sentiment Analysis• Sentiment analysis is “… one particular form of social media data mining, involv[ing] the application of a range of technologies to determine sentiments expressed within social media platforms about particular topics, in order to arrive at a measure of the ambient, or general sentiment”
Sentiment Analysis: How it works• Text mining ▫ Natural Language Processing Determines whether comments are positive, negative or neutral by analyzing word use, order, and combinations• Often done by third-parties ▫ Provide clients with: Insight on how to engage with their customers Community management services Raw data• Began as a score or grade for the business, however the new trend is to use the data in real time to deepen client/customer relationships
The good…• Companies want to know what people are saying about them• Consumers are trusting advertisements less, and peer recommendations more• Insight into customer opinions was previously unavailable on such a large scale ▫ Possible Outcomes Better customer service Quick resolution of customers’ problems Better products and services available to consumers Marketers can create better messages and identify the most efficient means of delivery Businesses can gain a deep understanding of their target audience’s psychographics
… and the bad• Slang, abbreviations, sarcasm are common on social networks, and are difficult to process• Studies have shown that it is difficult to extract sentiment from the things/ideas people tweet/post about most• The analysis may be inaccurate ▫ 70% accuracy is considered good ▫ Data may not be “clean”• Monetization of personal relationships• Social discrimination ▫ Less desirable demographic groups may be marginalized• Positivism problem ▫ People tend to give high ratings on many sites• General public is largely unaware of this practice• What data is considered public on social networks? When you join a social networking site, are you implicitly opting-in?
Issues in Social Media Data Mining• While there have been few cases of the use of data from social media sites for illegal or unethical purposes, many in the industry believe it is more of a matter of when, not if.• Companies mining data from social media sites are also very secretive about what they do, and how they do it. This is partially due to the fact that it is a new frontier and they do not want to give away trade secrets, however the opacity makes some experts nervous.
Issues in Social Media Data Mining• Privacy issues ▫ Thoughts and feelings shared become part of a “vast market research project” ▫ Data is readily available through social networks and aps. The more data, the more of a chance for problems Employees leaking customer information Hackers• Mobile ▫ Over 50% of Americans own a smartphone Aps have location data and access to address books in phone Companies can predict where users will be throughout the day Companies know who your friends, family, and coworkers are• Ethics ▫ How are companies obtaining their data? ▫ Do consumers know they are being tracked?• Legislation ▫ US Consumer Privacy Bill of Rights Federal Trade Commission legislation ▫ Loose framework ▫ Opt-out and privacy notices ▫ European Union “Do not track” policy Consumers must opt-in
Data Applications• The best way to analyze data mined from social media is to use a combination of computational and manual methods. Analytics programs can be used to clean and help code large datasets. Human coders are then used to check for accuracy, as computers cannot pick up on contextual clues, sarcasm, or humor.• Some programs that can be used for mining big data include: ▫ Apache Hadoop ▫ Apache HBase ▫ Apache Hive ▫ Cassandra ▫ Cloudera ▫ Greenplum ▫ Hadoop Distributed File System ▫ Hortonworks ▫ MapReduce ▫ MongoDB ▫ NoSQL
Best Practices in Social Media DataMining• Employ a combination of computer technology and human analysis ▫ Even sophisticated programs have difficulty extracting meaningful insight because of the prevalence of slang, abbreviations, humor, and sarcasm on social media sites• Ethical collection of data ▫ US policy is a very loose framework ▫ Collect only data that is considered public ▫ Use a reputable third party company for social media data mining to avoid ethical issues• Do not collect personally identifiable information ▫ Unless it will be used to resolve customer issues• Focus on using data from big groups to create psychographic profiles and uncover the general sentiment• React quickly to customer problems• “Listen” to what customers are saying to provide better products and services ▫ As opposed to monitoring to keep control of the online conversation
Companies offering Social Media DataMining Services• 33Across• Attensity• Get.It• McKinsey Global• Media6Degrees• Pentaho• PHD• Place1Q• SAS• Skyhook• WiseWindow
Social Media Data Mining for Marketing• Mining and analyzing the vast amount of information available on social networks will be a win-win situation for marketers and consumers if ethical issues can be avoided. Currently, companies who mine big data average 6% higher productivity than those that do not.• Marketers can: ▫ “Listen in” on online conversations to get a better understanding of: Who their customers are What products they want and need What advertising messages are most effective What channels are most effective ▫ Change their messages or target groups in real-time ▫ Help manage PR issues through quick customer service ▫ Aid in the development of new products and services
References• Barton, Dominic. "My, what big data you have." Canadian Business. 85.13 (2012): 14. Web. 17 Mar. 2013.• Brenner, Joanna. "Pew Internet: Social Networking (full detail)." Pew Internet. Pew Internet, 14 Feb 2013. Web. 17 Mar 2013. http://pewinternet.org/Commentary/2012/March/Pew-Internet-Social-Networking-full-detail.aspx.• Delo, Cotton. "Startups Mining Social Data take on Facebook." Advertising Age. 09 Apr 2012: 3. Web. 17 Mar. 2013.• Delo, Cotton. "You are big brother (but that isnt so bad)." Advertising Age. 23 Apr 2012: 1-19. Web. 17 Mar. 2013.• Every Stat You’ll Ever Want About LinkedIn (Infographic). 2012. Digital Marketing Ramblings… The Latest Digital Marketing Tips, Trends and Technology. Web. 17 Mar 2013. <http://expandedramblings.com/index.php/every-stat-youll-ever-want-about-linkedin-infographic/>.• Facebook Logo. N.d. Blogspot.com. Web. 17 Mar 2013. <http://3.bp.blogspot.com/- KNqO9JuXUN8/Ti2b1LHRquI/AAAAAAAAAIU/L6k8Wlzxj9k/s1600/logo_facebook.png>.• Facebook vs Twitter vs Pinterest – 2013 Statistics [Infographic]. 2013. Envision Media 360. Web. 17 Mar 2013. <http://www.envisionmedia360.com/infographics/facebook-vs-twitter-vs-pinterest-2013-statistics-infographic-719>.• Ferenstein, Gregory. "Fresh Stats On Social Networks: Pinterest Catches Up With Twitter, Digital Divide Shrinks." Tech Crunch. N.p., 17 Feb 2013. Web. 17 Mar 2013. <http://techcrunch.com/2013/02/17/social-media-statistics-2012/>.• Giles, Jim. "Text Mining." New Scientist. 14 May 2011: 34. Web. 17 Mar. 2013.• Greengard, Samuel. "Advertising Gets Personal." Communications of the ACM. 55.8 (2012): 18-20. Web. 17 Mar. 2013.• Kennedy, Helen. "Perspectives on Sentiment Analysis." Journal of Broadcasting & Electronic Media. 56.4 (2012): 435-450. Web. 17 Mar. 2013.• Lamont, Judith. "Big data has big implications for knowledge management." KM World. Apr 2012: 8-10. Web. 17 Mar. 2013.• Lamont, Judith. "Customer sentiment analysis: A shift to customer service." KM World. Feb 2013: 8-9. Web. 17 Mar. 2013.• Learmonth, Michael. "In pursuit of revenue, social networks ramp up ad targeting." Advertising Age. 10 Sep 2012: 20. Web. 17 Mar. 2013.• Lewis, Seth C., Rodrigo Zamith, and Alfred Hermida. "Content Analysis in an Era of Big Data: A Hybrid Approach to Computational and Manual Methods." Journal of Broadcasting & Electronic Media. 57.1 (2013): 34-52. Web. 17 Mar. 2013.• LinkedIn Logo. N.d. Mediameasurement.com. Web. 17 Mar 2013. <http://www.mediameasurement.com/mobile-social-networking-rises-by-44/linkedin-logo-008/>.• Mims, Christopher. "Mining the Mobile Life." Scientific American. 307.6 (2012): 42-43. Web. 17 Mar. 2013.• Moore, Andy. "Whats Different Now?" KM World. Oct 2012: n. page. Web. 17 Mar. 2013.• Moss, Rick. "All you need is love (and Facebook)." USA Today 14 Feb 2013, News, 8. Web. 17 Mar. 2013.• Pinterest Logo. N.d. mediaups.com. Web. 17 Mar 2013. <http://www.mediaups.com/wp-content/uploads/2013/02/Pinterest-logo.png>.• Smith, Craig. "(March 2013) How Many People Use the Top Social Media, Apps & Services?" Digital Marketing Ramblings… The Latest Digital Marketing Tips, Trends and Technology. N.p., 02 Mar 2013. Web. 17 Mar 2013. <http://expandedramblings.com/index.php/resource-how-many-people-use-the-top-social-media/>.• “Social Media.” Merriam Webster Online, Merriam Webster, n.d. Web. 17 Mar 2013.• Twitter Logo. N.d. Biomedicalimaging.org. Web. 17 Mar 2013. <http://www.biomedicalimaging.org/2012/images/logos/Twitter_logo.jpg>.