Sentiment Analyzer:
Extract Customer Sentiments From Social Media
By Ramanand Janardhanan
Every day, through social networks and blogs, consumers are expressing their
opinions about products, services and brands. While accessing this content
may be easy, analyzing it is a more daunting task. Due to the data's
unstructured format, manual processing is both costly and time-consuming.
The result for businesses: slow or ineffective response to market trends and
consumer needs.
BFS Innovations Sentiment Analyzer applies novel algorithms to quickly
identify meaningful opinions from online content. Businesses can monitor
consumer opinions from social networks, blogs, surveys and news articles. By
doing so, they can respond in kind—influencing customer perception of their
brands, and affecting buying behavior.
Sentiment Analyzer can also help businesses:
- Get closer to customers through better listening
- Identify customer opinions and learn their needs earlier
- Correlate customer sentiment with events
- Stay ahead of competition
- Quickly understand influence
- Improve targeting of PR campaigns
- Extract sentiments through linguistic and statistical analysis
- Find customer wish lists and problems
- Identify trends over time
- Analyze Voice of the Customer surveys, forums, reviews, news and more
Jump links
Innovation and Sentiment Analyzer
Businesses routinely use quantitative research to understand prevalent
market trends and customer needs. But qualitative content (via free-form
text) is more difficult to aggregate. Add the deluge of user-generated
commentary on websites, and businesses can easily be overwhelmed by data.
The emerging field of sentiment analysis makes it easier to navigate this
maze, by automatically identifying and analyzing online opinion content
regarding specific products and services.
What customers say about you
Using state-of-the-art, natural language-processing algorithms, BFS
Innovations Sentiment Analyzer focuses on opinions that come directly from
consumers in undiluted form. Are customers saying good things or bad? This
tool significantly reduces the time to identify bouquets and brickbats, to
pinpoint positive and negative trends, and to identify dissatisfied customers.
Wishful thinking come true
Sentiment Analyzer includes a unique "wish list" finder that lets businesses
tap into a virtual consumer suggestion box. Using an innovative in-house
algorithm, it analyzes user comments to detect possible preferences. When
these comments are routed to a product/service manager, they can augment
a product roadmap and help identify latent needs.
Powering Social CRM
Sentiment Analyzer powers Cognizant's new Social CRM solution, which
enables businesses to use social networks to locate leads and trends. Across
the industry, the solution is being viewed as an exciting new development for
marketing and sales professionals.
The dramatic increase in social networks has prompted businesses to adopt
new ways of understanding and engaging existing and potential customers.
Organizations are exploring how to exploit processes like analyzing sentiment.
Among the questions being asked: Which are the right sources to heed? And
how can one identify and decipher value from raw data?
The solution: mapping opinion results to business goals, such as customer
acquisition, retention, brand health, and user-powered innovation. Since so
much social content is unstructured, conventional analysis tools don't apply—leading to new tools such as Sentiment Analysis.
Linking opinions to metrics
Executives from industries such as retail banking, consumer goods, and
publishing are using Sentiment Analyzer to monitor and rate reviews from
public on line sources. The result: a clear picture, both good and bad, of
products and services as seen by talkative users—in the form of opinion
scores that can be connected to metrics such as customer satisfaction and net
promoter scores.
Figure 1 - Sentiment ratings of various service aspects
Voice of the customer
One Sentiment Analyzer user, a financial software firm, uses the tool as a "Voice of the Customer" analyzer, extracting customer survey data to pinpoint
problems and suggestions. By doing so, it has reduced the time and cost of
processing data, and helped its product managers focus on addressing
complaints and analyzing trends.
Social CRM
With Web 2.0 systems in place, CRM systems have evolved to help companies
foster deeper customer communication and involvement. Cognizant's Social
CRM solution can fine-tune CRM 2.0 strategies—using Sentiment Analyzer to
actively listen to and analyze chatter, then joining the online conversation in a
controlled fashion.
Though many companies want to move to a new CRM strategy, they're still
seeking guidance on how to smartly apply it . Cognizant's CRM practice
supplements Sentiment Analyzer with appropriate systems and best practices
for using it meaningfully.
Implications for the Future of Work
Sentiment Analysis has begun to help marketing, sales and product development
departments find more customers, enrich the customer experience, and respond
faster to complaints. By monitoring channels such as Twitter, customer support staff
can proactively detect and solve an expressed problem without waiting for it to be
formally reported.
A shift in culture
As business operations go increasingly online, organizations need to understand the
tone, tenor and boundaries of new media. For example, while traditional CRM tools
capture everything from names to numbers and addresses, new leads identified via
Social CRM may contain nothing more than an online handle and rough demographic
profile. As a result, sales staff must often engage anonymous prospects via email or
forum—while respecting the culture of specific networks.
One example: Interactions on professional network sites like LinkedIn are more
formal than those on Twitter. Another: The condensed, 140-character nature of
Twitter "tweets" affects language tools such as Sentiment Analyzer, since Twitter
language is so distinct from that of blogs and news sites.
A shift in decision-making
Though the science behind Sentiment Analysis is increasing the accuracy of its
results, the business offerings around the tool are still maturing. One key
requirement is to link customer sentiment scores and reports with business and
social media metrics. For instance, decision makers may wish to understand how to
interpret a recent negative trend. Does it affect overall customer satisfaction scores?
How likely is customer attrition? Should an issue be fixed sooner than later?
Making sentiment more useful
Key decisions are also dependent on factors such as the influence and reliability of a
source (Was a critical blog post planted by a rival?). As industry acceptance of these
new methods grows, Sentiment Analysis is helping tackle related issues.
Figure 2 - Actual opinion-words about various product aspects
The BFS Innovations Sentiment Analyzer is a content processing engine that
analyzes consumers' online texts for opinions about brands, products and services.
By doing so, it helps businesses listen to and understand their customers better—and
identify problems, preferences and wish lists.
Sentiment Analyzer can take the following forms:
- Product/service review analyzer
- Voice of the Customer analyzer
- News monitor
A core component in Cognizant's new Social CRM offering, Sentiment Analyzer
enables companies to incorporate online channels in their CRM systems, to analyze
trends and identify new leads.
Figure 3 - Sample wish list extracted from public reviews on the iPod