When size matters Is social media data really that BIG

• 1. When size matters:is social media data really that BIG?Olha BondarenkoSocial Media Architect | Philips IT | March 2013

2. Olha Bondarenko, Social Media Architect@Philips IT Olha Bondarenko | Social Media architect | Philips IT | March 2013 3. TodaySocial media & big data: what makesit big and why using it?Social media types of data andapplications for businessMeasuring social: the good,the bad and the ugly Olha Bondarenko | Social Media architect | Philips IT | March 2013 3 4. Your takeawaysUnderstand what kind of data is available in socialmedia space and how it can be usedThink of relevant applications of data from social mediaspace for reaching your business objectivesImage: office.microsoft.comOlha Bondarenko | Social Media architect | Philips IT | March 2013 4 5. Socialmedia& bigdataImage: office.microsoft.comOlha Bondarenko | Social Media architect | Philips IT | March 2013 5 6. These impressivenumbers have tobe translated intobusinessopportunities &revenuesSource: Big data: The next frontier forinnovation, competition, andproductivityMcKinsey Global Institute, May 2011 Olha Bondarenko | Social Media architect | Philips IT | March 2013 6 7. What makes data BIG: IBM viewSource: Analytics: The real-world use of big data, IBM Institute for Business Value, 2012 Olha Bondarenko | Social Media architect | Philips IT | March 2013 7 8. Does social media data qualify as a big datasource?Respondents with active big data efforts were asked which platformcomponents are currently either in pilot or integrated into thearchitecture. Each data point was collected independently. Totalrespondents for each data point range from 297 to 351.Source: Analytics: The real-world use of big data, IBM Institute for Business Value, 2012 Olha Bondarenko | Social Media architect | Philips IT | March 2013 8 9. What makes social media data big? Social media data Huge, but scalable inNeeds effort to structure & Requires the business relevant areas standardizemodel to adapt Very high, but offers greatbusiness opportunitiesBased on:: Analytics: The real-world use of big data, IBM Institute for Business Value, 2012 Olha Bondarenko | Social Media architect | Philips IT | March 2013 9 10. Social media data for business use: someexamples Social media data +Other data sources Competitive intelligence Innovation & co-creationInfluencers engagement Issue prevention Campaign evaluation Product development Crisis preventionCustomer journeys Brand managementMI metricsOlha Bondarenko | Social Media architect | Philips IT | March 2013 10 11. Example: social media data use @Dell http://www.slideshare.net/dellsocialmedia/idc-sadler-feb2012Dell converted an early 2005 social media crisis into a holistic strategyOlha Bondarenko | Social Media architect | Philips IT | March 2013 11 12. Socialdatatypes &theirapplicationsImage: office.microsoft.comOlha Bondarenko | Social Media architect | Philips IT | March 2013 13. Three types of data available from social media 1. Linkage* 2. Profile3. Message*The linkage behavior of the Information about theContent published on various network, important nodes, participants of the network,platforms, from 140-charater-communities, links, evolving either provided by them or cryptic tweets to lengthy regions* deduced opinion blogs*Social Network Data Analytics. Charu C. Aggarwal, Ed. Springer Science+Business Media, LLC 2011 Chapter 1. An Introduction to Social Network Data Analytics, pp. 5, Images: office.microsoft.comOlha Bondarenko | Social Media architect | Philips IT | March 201313 14. 1. Linkage data: the secrets of the net 1. Linkage* 2. Profile3. Message*The linkage behavior of the network, important nodes,communities, links, evolving regions**Social Network Data Analytics. Charu C. Aggarwal, Ed. Springer Science+Business Media, LLC 2011 Chapter 1. An Introduction to Social Network Data Analytics, pp. 5, Images: office.microsoft.comOlha Bondarenko | Social Media architect | Philips IT | March 201314 15. Police uses linkage data to understand thestructure of a gang & identify missing members [] the social network analysis alsoidentified "six other vital players of whichImage: office.microsoft.com the police were unaware."Sources: http://www.core77.com/blog/technology/visualizing_criminal_networks_to_help_police_solve_crime_22462.asp ,http://www.zdnet.com/ten-examples-of-extracting-value-from-social-media-using-big-data-7000007192/#photo Olha Bondarenko | Social Media architect | Philips IT | March 201315 16. Less exciting world of daily business, two(anonymous) Forrester examples for linkage data Increased chances to cross-sell/upsell: A telcotaps into the Facebooksocial groups to market friends-and-family plans.Image: office.microsoft.comPreventing chain cancellations: A credit card company retainscustomers by understandingImage: office.microsoft.comsocial relationships. Source: The Big Deal About Big Data For Customer Engagementby Sanchit Gogia, Forrester, June 1, 201216Olha Bondarenko | Social Media architect | Philips IT | March 2013 17. 2. Profile data: valuable, sensitive & uncertain 1. Linkage* 2. Profile3. Message*Information about the participants of the network, either provided by them ordeduced*Social Network Data Analytics. Charu C. Aggarwal, Ed. Springer Science+Business Media, LLC 2011 Chapter 1. An Introduction to Social Network Data Analytics, pp. 5, Images: office.microsoft.comOlha Bondarenko | Social Media architect | Philips IT | March 201317 18. GE uses social media (geolocation) as one of thedata sources to detect & locate power disruptions Source: GE Grid IQ brosuer http://www.zdnet.com/ten-examples-of-extracting-value-from-social-media-using-big-data-7000007192/#photoOlha Bondarenko | Social Media architect | Philips IT | March 2013 18 19. Dutch railway organization Prorail uses Twitter& geolocation to detect the snowfallsImages: office.microsoft.comhttp://blog.prorail.nl/twitcident-inzicht-in-sneeuwval-via-innovatieve-social-media-scanOlha Bondarenko | Social Media architect | Philips IT | March 2013 19 20. 3. Message data: the needle in a haystack 1. Linkage* 2. Profile3. Message* Content published on various platforms, from 140-charater-cryptic tweets to lengthyopinion blogs*Social Network Data Analytics. Charu C. Aggarwal, Ed. Springer Science+Business Media, LLC 2011 Chapter 1. An Introduction to Social Network Data Analytics, pp. 5, Images: office.microsoft.comOlha Bondarenko | Social Media architect | Philips IT | March 201320 21. A machine cantfully understand human talk yetImage: office.microsoft.comOlha Bondarenko | Social Media architect | Philips IT | March 2013 21 22. Example: text analysis of Amazon reviews byAttensitySource: Making Social Insights Actionable, An Attensity eBookOlha Bondarenko | Social Media architect | Philips IT | March 2013 22 23. Respecting privacy and obeying to legislations is of the outmost importance for PhilipsOlha Bondarenko | Social Media architect | Philips IT | March 2013 23Image: office.microsoft.com 24. Measuringsocial:the good,the bad& the uglyOlha Bondarenko | Social Media architect | Philips IT | March 2013Image: office.microsoft.com 25. The information value chain: making sense of chaos Image: office.microsoft.comOlha Bondarenko | Social Media architect | Philips IT | March 2013 25 26. Simple (numerical) metrics: measure the conversation Share of voice per competitorMentions per media type Trend over time per media typeSource: http://www.salesforcemarketingcloud.com/products/social-listening Olha Bondarenko | Social Media architect | Philips IT | March 201326 27. Example: Listen to the whispers - a simpleanalysis leading to a great business outcomeSource: Philips OneVoice Connect for GM&C, 2013Hackers start their conversation 24 hrs before the bad news hit mainstreamOlha Bondarenko | Social Media architect | Philips IT | March 2013 27 28. Understanding the sentiment ofa conversation isimportant but may appear more difficult and lessmeaningful than one hopesneutralImage: office.microsoft.comOlha Bondarenko | Social Media architect | Philips IT | March 201328 29. Certain topics, such as elections, render highvolume of emotional conversationSource: Image courtesy of Clive Roach, through HootsuiteOther speculative metrics examples: recommendation, purchase intent Olha Bondarenko | Social Media architect | Philips IT | March 2013 29 30. Influencers: having a focused impactful conversation, especially in the B2B spaceCommon as well asproprietary influencescores existIdentify the mostinfluential participantsof a conversationNot as simple as justcounting followers orfriendsSubject- and time-dependentScreenshot from: http://twittercounter.com/pages/100 Olha Bondarenko | Social Media architect | Philips IT | March 201330 31. Complex combinatory metrics: Synthesio SocialReputation ScoreSource: SynthesioOlha Bondarenko | Social Media architect | Philips IT | March 2013 31 32. Seeing the future through social: predicting fluspread, movie tickets sales & stock market4.4 million tweetsfrom 630,000 usersanalyzed.Claimed to predictwhen healthypeople will fall sickwith 90% accuracyup to eight days inadvance.Source: http://www.newscientist.com/blogs/onepercent/2012/07/ai-predicts-when-youre-about-t.html Olha Bondarenko | Social Media architect | Philips IT | March 2013 32 33. Your takeaways Image: office.microsoft.com Understand what kind of data is available in social media space and how it can be used Is social media data big? It may be! Three data types and a range of metrics exist The business value is indisputable; applications vary Think of relevant applications of data from social media space for reaching your business objectives Use todays examples for your inspiration Start exploring & connect to the team!Olha Bondarenko | Social Media architect | Philips IT | March 201333 34. Olha Bondarenko | Social Media architect | Philips IT | March 2013 34