by Seth Grimes
by Jan Wiebe
One of social media monitoring's benefits is that the results reveal trends in opinions that are not uncovered by ongoing customer satisfaction survey research. Recently, many companies have moved toward mining social media data for customer opinions and in parallel, reduce the amount of time and money spent on customer satisfaction survey research. This approach unfortunately may reduce the insight a company has on customer satisfaction. This presentation will explore the problems with only mining social media data, and will demonstrate the benefits of combining survey research with social media monitoring. We will show how to properly perform the the data integration and then classify the combined results to attain the holy grail of all customer satisfaction research: find out negative sentiment about the company that wasn't previously known and, if left unexplored, will impact the bottom line.
by Rich Brown
Sophisticated trading and investment firms, seeking a competitive edge from managing 'big data,' need to make sense of all the unstructured news and opinion distributed every day over, not only traditional news sources, but financial news sites, chat rooms, blogs and other social media channels as well. With modern economic and financial theory increasingly looking towards behavioural and crowd-based models to explain cyclical booms and busts, the ability to quantitatively assess broad market sentiment can be a valuable indicator to help guide successful trading and investment strategies. The world of 'big data' is a new market data frontier that can provide trading firms with a real edge by helping to model behavioural economics. News and Social Media Sentiment tools allow traders to quantify market sentiment from the millions of public and premium sources of financial information available via the Web, turning a lot of noise into actionable insights.
There are three forms of power: knowledge, position and influence. In the social space, position is irrelevant. Knowledge and influence is paramount. How that comes across is captured in sentiment. Sentiment is defined as thoughts, views, or attitudes, especially ones based mainly on emotion instead of reason. Sentiment drives a person' s consideration set and thereby influences their decisions. Too often people focus on the 'poles', e.g. the positive or negative sentiment and miss the opportunity capture by neutral sentiment, aka the 'swing voter'. Taking the time to consider this swing sentiment will in turn shape strategies, products, offers and messaging.
As social web becomes increasingly mainstream as a customer engagement channel, it grabs a larger share of interest and intrigue from customer experience practitioners and sentiment analysis technologist. Banafsheh Ghassemi will discuss practical opportunities and blind spots this trend delivers in a multi-channel world.
by Srini Bharadwaj
Hear how Crimson Hexagon's social media monitoring and analytics software helps organizations use insights gained from the deep research and analysis of online social media conversations to directly impact a business's bottom line.
by Andera Gadeib
Cases from research practice will demonstrate a text mining process that works towards meta-ontologies, differentiating 9 categories, comprising among others functional and emotion attributes. In 2011 we went a step further, splitting all emotions found into 8 dimensions generating a graphic called Emotions Radar. The main two axes range from pleasantness to unpleasantness and from engagement to disengagement. In between strong/weak positive and negative associations are covered. The system builds on visual analysis and has proven to capture emotions quite well - going beyond positive/negative sentiment.
Examples shown will be a data set from a 360 deg ad effectiveness study, showing that TV or print generate quite different emotions engaging consumers much less than a recommendation by friends for example. Two other practical cases will be shown:
1. A data set from an online innovation project on oral care for kids and dogs. The ideas are brainstormed with panelists and analyzed looking at the ontology.
2. Emotions Radar analysis of a Social Media project searching for new product ideas in the area of vacuum cleaners where we have been surprised how emotional consumers can get.
* How the Media Uses Sentiment Analysis: A Perspective from the Wall Street Journal
Ryan Sager, the Wall Street Journal
* Capturing Sentiment via Customer Intelligence
Can we capture customer sentiment from enterprise feedback received? How reasonably? Can we glean any insight from that feedback? Would this help in building the strongest customer relationship? The presentation will answer these questions, how Fidelity analyzed sentiment from customer feedback as part of our customer intelligence paradigm.
Sobhan Hota, Fidelity Investments
* "How Can I Listen If I'm Talking?": The Power Of Social Media Listening
Social media listening research has emerged as a powerful alternative to more traditional, asking research. Through a number of examples you will find out not only how you can research important brand topics but how you can provide in-depth insights into new product development, segment analysis, and broader topics that traditionally you might not have the funds to research. Using a mixture of paid and widely-available tools you'll learn how to use this cutting-edge research method for your important research questions.
Frank Cotignola, Kraft Foods
by Seth McGuire
by Jeff Catlin
* Human, Hand-Crafted Brand Sentiment: Crowdsourcing for Quality Data Management
For years enterprises requiring crowdsourcing data quality verification could not support the manual authentication processes associated with big data sets. And, projects with smaller numbers of records still required customers to sort, qualify and make sense of task outputs, producing little ROI. Today's crowdsourcing tools provide a quality and manageability layer for businesses engaged in crowdsourcing efforts, producing meaningful outputs that may be leveraged immediately for business improvements and growth.
Max will discuss how artificial intelligence is best suited for crowdsourcing and 'next' practices for revolutionizing marketing, brand and advertising campaigns, providing businesses with competitive advantages.
Max Yankelevich, CrowdControl Software
* Using the Crowd to Tell You What Customers Are Thinking
With access to more than 3 million Contributors worldwide processing more than 5 man years of work on a daily basis, the crowd is a powerful tool if you know what to do with it. Many sentiment analysis tools are limited and do not provide the level of detail that a brand needs to really understand what the customers are thinking, saying, feeling, etc. about it. With the introduction of crowdsourcing, sentiment can be evaluated faster, more efficiently, and with scalability that traditional means of gathering this information lack. Biewald brings insight and experience as a thought leader in the crowdsourcing space who will talk about use cases and best practices to highlight the power of the crowd.
Zack Kass, CrowdFlower
* Training the Cloud with the Crowd
The SproutLoop - Dialogue Earth collaboration focuses on inferring public opinion from Twitter in a nuanced and scalable manner. Our strategy is built upon a foundation of domain knowledge, be it about weather, global warming, or brand management. We are driven to understand the key questions, and then how to infer reliable answers to these questions by "listening in" to the Twitterverse. Our method is based on developing a "survey" that, when paired with a batch of tweets, results in tweets being reliably sorted into various bins, such as "user had a positive mood about current weather" or "user had an unfavorable DQ Blizzard experience." The method is scalable because we rely on crowd-sourced workers with minimal training, who can bin tweets using our survey. We have processed 1+ million crowd sourced sentiment judgments. This crowdsourced sentiment data can then be leveraged as training for predictive models. These Niche/Domain specific predictive models can be integrated for real-time analysis of Twitter streams. The result is an approach that can yield rich insights from tweets, with options for presenting data geographically over time. A sentiment analysis case study will be presented that used the cloud technologies: Twitter API, CrowdFlower and Google Prediction API.
Kevin Cocco, SproutLoop.com
by Ronen Feldman
Researchers and practitioners are developing and applying new methods to determine how Web users feel about everything: products and politicians, medical drugs, friends and family, scientific articles and celebrities. In this talk we will cover recent advances that combine information extractionwith sentiment analysis to improve accuracy in assessing sentiment about specific entities. We will present several real world applications ofsentiment analysis that utilize unsupervised learning and semantic clustering.
by Carol Rozwell, Romi Mahajan, Bing Liu and Leslie Barrett
Sentiment analysis exploits academic and industrial research to enable business solutions that harvest and make use of opinion, attitude, and emotion in the broad array of human communications. Progress is constant, with much room for further advances: Subject matter for what should prove to be an illuminating discussion of sentiment analysis Innovation, featuring panelists who are themselves exemplary Innovators.
8th May 2012