Recently Cheryl Contee, our CEO at Fission Strategy and the Co-Founder of Attentive.ly, spoke on a webinar panel about Sentiment Analysis. Over 700 attendees registered to join. The speakers included 2 other experts in social analytics and the discussion was moderated by Social Media Today. There was a lot to learn not just from the panelists, but also from those interacting with the event on Twitter. Here is a summary of our discussion about this innovative approach to listening to sentiments within the digital world.

What is Sentiment Analysis (SA)? It is not your traditional method of analyzing potential customer sentiments towards words, issues, and organizations. However it is now becoming an adjunct to more traditional avenues, because it gives quick insight. SA is the study of consumer sentiments towards a product, word, or issue. This analysis is connected to conversations within social networks online.

Frank Cotignola, one of the speakers from Mondelez International Consumer Insight and Strategy Group, stressed the importance of understanding the challenges and weaknesses of Sentiment Analysis. Without its proper place within traditional analysis, we run the risk of “overselling it”, he said. One important challenge that he mentioned, is the technology’s inability to detect sarcasm, humor, and the overall context and nuances of the language used on social networks. This creates a tricky space with a necessity for human interpretation of the analytics.

However, as Cheryl said, Sentiment Analysis is one of the best tools that we have to deeply connect to people, because we can respond to them emotionally through the information this analysis gathers. She used Attentive.ly’s work with Rep. Mark Takano (D-CA) as a case study. You can read the full case study on Attentive.ly’s blog.

Keeping a Member of Congress’ office running smoothly is a busy tech operation, with CRMs, social media tools and an active comms team, both at home and in Washington DC. Sound familiar? In order to effectively listen and engage with constituents, they developed outreach goals based on 1) the top issues in the state, 2) increasing followers and 3) connecting with influencers.

Because the Congressman’s communications team tagged key issues based on demographic data such as transportation or immigration, this allowed them to send multi-channel and specific content via social networks & email, including near the launch of Rep. Takano’s very public transportation plan. The result? Active social listening with SA allowed the Congressman to respond personally on Twitter and Facebook when constituents talked about the announcement of his transportation plan.He targeted those who demonstrated an active interest on an issue, so it wasn't a spam or blast response to all constituents. This is one of specific strengths of Sentiment Analysis using boolean search techniques.
 

Cheryl noted that as much as there is always error and inaccuracy within any analytical method, it is possible to use keyword analysis, inspired by political news headlines for example, which can yield a 60% accuracy using keyword analysis and even up to 85% when natural language processing. Cheryl stated that you can reduce errors or off-topic noise by being as precise as possible, using well-designed boolean queries and using human intelligence that understand the nature of your brand and audience. She advised:  “You need a human mediator to interpret what the tech is saying to reduce the error rate and to pick out the messages that are most important.” She also mentioned that in the future, through more sophisticated technology, more sophisticated users, and even faster data crunching tools, the true promise of SA to enhance communications will be realized.

Eric, the VP of Tracx mentioned that despite inaccuracy, larger companies are trusting this method to gain insight into their employees’ sentiments about company culture and efficiency.

One of the ways that Fission and Attentively employs SA in their social analysis is through very close partnership with their clients, who understand their brand better than anyone. Together they can experiment and find new ways to make the technology work for them. Perhaps the most innovative way that Fission and Attentive.ly are using SA for clients is through interpretation of the results to identify key influencers within social communities. Just by identifying these influencers alone, Attentive.ly was able to help the Congressman gain over 150 new followers in a very short amount of time because the influencers called their followers to action.

It looks as if SA may be the go-to method of analysis for organizations and businesses in markets that are rapidly changing and where connecting to an audience in a timely manner is urgent. It is even more important for causes and issues that require a large following in order to make a difference through action — a key part of Fission Strategy’s work with clients. All of the speakers agreed that with the improvement of queries, the broadening of starting sources, and with great client / technology interaction, the future of Sentiment Analysis looks very bright.