Optimizing Usage Analysis During Implementation Of Social Media Systems
0 likes681 views
This document discusses optimizing social media usage analysis during implementation through sentiment analysis. It notes that while companies are using social media, many are not benefiting fully as conventional usage monitoring focuses only on metrics and ignores user sentiment. Sentiment analysis of user comments could provide more insights into issues and influence user sentiment. Detecting "bursts" of activity through text analytics may help understand events driving sentiment changes and allow companies to better predict and respond to user feedback.
Optimizing Usage Analysis During Implementation Of Social Media Systems
- 1. Optimizing Usage Analysis during Implementation of Social Media Systems A specific use-case for sentiment analysis Martin Rueckert SAP EMEA Skill & Knowledge Management Philip Raeth Ph.D. Student, European Business School
- 2. Social Media is a widespread phenomenon 2 Flowtown (2010) Facebook: 500M user Twitter: 114M user Google: more than 500M user
- 3. Many companies are using social media but many are not able to benefit from them Many companies already use social media (McKinsey, 2009) Not every company is able to benefit from social media potential (McKinsey, 2009) 3 (McKinsey, 2009) 51% 12% 37% bought social media plan to buy social media (software not yet looked at buying social media (software) 94% 87% 75% Internally At the interface with customers At the interface with suppliers Companies are using social media … 27% 29% 43% Employees Customers Suppliers Companies (have) … Companies realize usefulness effects of …
- 4. Companies are seeing small participation numbers while participation is the basis of the success! 4 “The diffusion describes the process of how the innovation is being published through various channels of the social system over a certain period of time.” Rogers (2003) Partofthepopulation 100% 50% Innovators 2,5% Early adopters 13,5% Early majority 34% Late majority 34% Laggards 16% How can the process of our own initiative be measured in that respect?
- 5. Success is being measured with usage rates but does that give sufficient insight? 5 Flowtown (2010) Facebook: 500M user Twitter: 114M user Google: more than 500M user
- 6. Usage of information systems is being visualized from three perspectives 6 The System as the element of usage Logins/Hits/Visits (frequency), Features used (amount of features) Processing time (duration). The user as the element of usage (cognitive elements measured ie. through surveys) Sentiments during usage State/Level of focus The Task the system is used in/for tasks supported by the system How is the system being used? Conventional system monitoring focuses on measuring system usage and ignores other perspectives!
- 7. Today: Conventional usage monitoring Automatic, quantitative measuring Permanent, automatic usage monitoring of access rates Top level Communities Groups Registered users - logged in at least one time. Unique active users – Has at least one participation point Groups nominated for archiving – no activities since 3 month Removed Groups – Groups deleted by Group owner after nomination Closed Groups – no activities since 4 months All measures that are currently being monitored for social media at SAP: 0 1000 2000 3000 4000 5000 6000 Active Participants 0 200 400 600 800 1000 1200 1400 1600 1800 2000 Groups
- 8. Issue: Details remain invisible! Are conventional, quantitative measurements precise enough? Problems Certain events provoke a burst of activity Correlation of burst with event? → difficult (using conventional, quantitative methods) Extrapolation of burst with high precision? → difficult (without knowing details of event) Text- Analytics can detect events, problems and other relevant semantic concepts. → ability to detect the type of event → ability to predict subsequent events based on historical knowledge of similar events “Buzz” / “Burst” detection 0 10 20 30 40 50 60 70 80 90 100 Closed Groups 0 50 100 150 200 250 Groups nominated for archiving „What happened?“
- 9. Alternative today: manual user surveys Results of survey asking for opinions Early-Adopters survey during SAP Social Media Pilot 0 5 10 15 1. By using Jive SBS it takes less time than before to find relevant information. 2. Communication with experts and other teams is easy. 3. Everyone in our team does exactly know what the others are doing to achieve our team goals. 4. Everyone in our team is always informed about tasks, timelines and progress. 5. Team members can easily collaborate and share their knowledge. No Yes Small amount of users surveyed (n=13) because of significant costs (Time to create survey, to explain questions, fill in answers, consolidate answers) No continuous monitoring of user sentiment possible over a significant time span High precision of answers, low error rate and low recall Highlights
- 10. Alternative today: manual user surveys Results of usage frequency survey Early-Adopters survey during the SAP Social Media Pilot Small amount of users surveyed (n=13) because of significant costs (Time to create survey, to explain questions, fill in answers, consolidate answers) No continuous monitoring of user sentiment possible over a significant time span High precision of answers, low error rate and low recall Highlights 0 1 2 3 4 5 6 7 8 9 Create Online document Upload files Discussion Forums Group Blogs Personal Blog Post Poll Tags RSS Jitter/status update Announcement Connect to people Bookmarks Private Messages Comments daily weekly more than once a month once a month less than once a month not used
- 11. Determine whether people are positive or negative on a certain topic and to what degree Support root cause analysis and rapid refinements to strategy Understand and explain the why behind metrics such as Net Promoter Scores Tap into buzz about brand Is it positive or negative? Does it change in response to market events? Highlights 11 Basics: What is “Voice of the User/Customer”? Sentiment Analysis & Buzz Detection
- 12. 12 Mr. Jones is very unhappy with Green Insurance Co.; the main issue is that the offer for his totaled vehicle is too low. He states that Green offered him $1250.00 but his car is worth anywhere from $2,500 to $4,500 based on Kelly Blue Book and Auto Trader. He has spoken with Jenny J. about this, and an independent adjustor (555-222-1234) will assess the vehicle. The second issue is that John Smith is not returning his calls. Jones has also spoken to supervisor Jane Doe and his e-mail has been escalated to headquarters.... PERSON Mr. JonesJones; Jenny J.; John Smith; Jane Doe ORGANIZATION Green Insurance Co.Green; Kelly Blue Book, Auto Trader CURRENCY 1250 USD, 2500 USD, 4500 USD PHONE 555-222-1234 SENTIMENT_NEGATIVE Mr. Jones is very unhappy with Green Insurance Co. PROBLEM the offer for his totaled vehicle is too low John Smith is not returning his calls SERVICE Jones has also spoken to supervisor Jane Doe his e-mail had been escalated to headquarters… Basics: What is sentiment analysis? Text Analytics explained
- 13. 13 Automatic Sentiment analysis in Social Media Systems Screenshot: sample user comment from social media
- 14. © SAP 2008 / Page 14 Automatic analysis of frequency of comments where the topic is social media (over several months) Alternative Monitoring - Highlights
- 15. © SAP 2008 / Page 15 Detailed sentiment analysis using text analytics Detailed detection of issues Enables influencing user sentiment Highlights
- 16. © SAP 2008 / Page 16 High level of detail in detected sentiments Highlights
- 17. Additional analytics possebilities: Burst detection in content and activity stream 17 (Kleinberg, 2002) (Kleinberg, 2002) Highlights Burst of activity in communication streams are usually traceable back to a discrete event The activities of a chain of events increase with respect to intensity and frequency It is possible, to segment large text corpora into activity oriented sub-patterns It is possible to categorize such sub-patters and assign labels using keyword annotation Adding the time/event dimension, social media usage and user sentiment can be correlated
- 18. Outlook: prediction of events Underlining the importance of „early adopters“ sentiment 18 (Gruhl et al. 2005)(Gruhl et al, 2005) Highlights Sales rank = Expression of user sentiment Count and quality of user comments correlate with bursts of user sentiment (marketing effect) It is possible to identify documents that predict to-be user sentiment It is possible to automate such identification processes It is possible to use social media for these predictions Mentions/Sales Rank Diagram: “The Lance Armstrong Performance Program” Mentions/Sales Rank Diagram : “What not to wear”
- 19. Summary 19 Highlights The success of a social media roll-out is largely depending on the acceptance of early adopters: Requires very good support of this user-type and their issues Requires to predict and identify bursts of activity and sentiment trends Measuring conventional system usage can be insufficient: Missing details/depth of details No direct deduction of counter measures from system usage No continuous measuring of user sentiment Text analysis can help monitor user sentiment: Continuous measurement possible Highlight issues or requests Detection of “Bursts of Activity” can help understand significance of user sentiment and take appropriate counter measures i.e.. through clustering of issues using burst detection Make success of marketing measures visible/predictable
- 20. Questions? Martin Rueckert SAP Philip Raeth European Business School
- 21. 21 No part of this publication may be reproduced or transmitted in any form or for any purpose without the express permission of SAP AG. The information contained herein may be changed without prior notice. Some software products marketed by SAP AG and its distributors contain proprietary software components of other software vendors. SAP, R/3, xApps, xApp, SAP NetWeaver, Duet, SAP Business ByDesign, ByDesign, PartnerEdge and other SAP products and services mentioned herein as well as their respective logos are trademarks or registered trademarks of SAP AG in Germany and in several other countries all over the world. Business Objects and the Business Objects logo, BusinessObjects, Crystal Reports, Crystal Decisions, Web Intelligence, Xcelsius and other Business Objects products and services mentioned herein as well as their respective logos are trademarks or registered trademarks of Business Objects S.A. in the United States and in several other countries. Business Objects is an SAP Company. All other product and service names mentioned and associated logos displayed are the trademarks of their respective companies. Data contained in this document serves informational purposes only. National product specifications may vary. The information in this document is proprietary to SAP. No part of this document may be reproduced, copied, or transmitted in any form or for any purpose without the express prior written permission of SAP AG. This document is a preliminary version and not subject to your license agreement or any other agreement with SAP. This document contains only intended strategies, developments, and functionalities of the SAP® product and is not intended to be binding upon SAP to any particular course of business, product strategy, and/or development. Please note that this document is subject to change and may be changed by SAP at any time without notice. SAP assumes no responsibility for errors or omissions in this document. SAP does not warrant the accuracy or completeness of the information, text, graphics, links, or other items contained within this material. This document is provided without a warranty of any kind, either express or implied, including but not limited to the implied warranties of merchantability, fitness for a particular purpose, or non-infringement. SAP shall have no liability for damages of any kind including without limitation direct, special, indirect, or consequential damages that may result from the use of these materials. This limitation shall not apply in cases of intent or gross negligence. The statutory liability for personal injury and defective products is not affected. SAP has no control over the information that you may access through the use of hot links contained in these materials and does not endorse your use of third-party Web pages nor provide any warranty whatsoever relating to third-party Web pages. Copyright 2010 SAP AG All Rights Reserved