Understanding the customer in the world before Web 2.0 was a relatively straightforward task – but the emergence of Web 2.0 has brought an explosion of social channels including blogs, wikis, forums, social networks and an array of social media. Today, customers have a loud and clear voice where they openly share ideas, perceptions, and problems about products and companies. They create trusted communities and powerful, influential constituencies. The voice of the customer (VOC) is therefore expressed in different forms through different channels. These channels are insight rich, with a wealth of untapped customer intelligence. Traditional technologies are unable to access or decipher the unstructured content upon which today’s customer conversations and insights are built.

While the ability to connect with and understand customers has grown more complex, the need for customer intimacy has never been more critical. In the 2010 IBM CEO Study called Capitalizing on Complexity; one of four primary findings states that the most successful organizations co-create products and services with customers and integrate customers into core processes. This finding will drive companies to focus on VOC efforts across traditional functional areas including: Sales, Marketing, Advertising, Public Relations, Call Center and Product development. Voice of the Customer (VOC) is a time-tested business concept that has gained new life with this need for customer intimacy and the explosion of user generated content. These efforts will increasingly focus on gaining a collective view of all customer insight by analyzing a broad range of channels where customers openly voice their opinion.

Historically, VOC efforts were managed in silos and therefore sub-optimal. To handle specific issues quickly, organizations have turned to predictive analytics to focus on who is most likely to buy or what type of customer is most likely to churn. They have focused on data mining to understand past buying behaviours, and analyzed the results of surveys, email, Call Center notes, warranty claims or social media, by using a specific text mining application or service geared toward that specific customer channel. However, optimal results are only obtained when organizations take a holistic view of all customer channels and manage voice of the customer initiatives as part of a comprehensive program. When managed this way, and addressed by an information management platform, Voice of the Customer programs deliver considerable benefits.

Effective VOC efforts are data driven with clear goals and expectations. It complements and extends traditional CRM by capturing customer requirements and feedback in order to provide best in class products and services. This process is proactive and constantly innovative in an attempt to capture the changing requirements of customers over time. Tapping insight about customer requirements is complicated by sheer volume and the unstructured nature of a high percentage of customer data. Further, the market is inundated with software that attempts to tag, sort, search, organize and manage much of this unstructured data. But discovering the actual facts in this data is a challenge that frustrates most companies.

The only available means most companies have to understand unstructured data is to have humans read it – but humans are poorly suited to read volumes of text records. Online discussions and volumes of customer emails simply cannot be processed into useful, actionable insight without the capabilities offered by automated solutions. As a result, the unstructured data collected from customers is often not used at all. Valuable insight is lost, as the reason for an event or opinion is often revealed in text, as are potential marketing opportunities or even early warning on issues.

Emerging solutions will take the form of analytic applications – examples include customer retention, new product launch, customer acquisition, innovation, brand and reputation management, quality and safety management, customer service and marketing effectiveness. These applications deliver business outcomes associated with several critical customer related business challenges:

• The need for better and faster innovation
• A better understanding of where to focus the product team and associated development dollars
• A better and earlier understanding of the success or failure of product launches
• Ineffective and costly sales, marketing and customer service
• Sub optimal call center and product life cycle processes
• Customer churn
• Trouble acquiring new customers and growing share of wallet from existing customers
• High customer expectations
• Mitigating brand risk with a better understanding of brand and reputation impact
• An early warning of emerging product issues or defects in the market
• The identification of fraudulent warranty claims
• A better understanding of the actionable details behind the NetPromoter Score™ (NPS)
• A better understanding of competitors, the market and pricing problems

To address these challenges, unstructured data must be gathered from multiple sources and fused together with relevant structured data to create a 360 degree view of the customer. Examples of relevant unstructured data:

• Contact center transcripts and notes
• Blogs and other Social Media
• Surveys
• News and other articles
• Complaints via web forms
• Email
• Online forums
• Chat sessions
• Customer service notes
• Warranty notes
• Instant messaging logs
• Point of service notes
• Service requests via web forms
• Wikis
• Focus groups transcripts
• Defect reports
• Trial tests
• Repair notes

This fused data is then fed into the analysis process where multiple analytic methods are utilized. Out-of-the box natural language processing is supplemented by pre-defined dictionaries to broaden the analysis of unstructured data. These predefined dictionaries are established to focus on specific use cases (e.g. call center optimization, social media analysis, brand and reputation management, etc.). A key component of analytic applications is parts of speech analysis, which allows machine-based learning to accelerate dictionary creation. This is a great improvement over rules-based solutions that take months if not years to deploy. Other examples of analytic methods applied to VOC challenges are:

Predictive Models

• Predicting who is most likely to buy
• Predicting which customers are most likely to churn, etc.

 Text Analytics

• Survey analysis
• Email analysis
• Call-center notes analysis
• Warranty claims analysis
• Social media analysis
• Documents
• Other

Data Mining

• Analyzing prior marketing campaigns
• Churn analysis, etc.

 Statistical Analysis

• Mean, median, mode
• Quartiles
• Range
• Variance
• Skew
• Standard deviation
• Distribution
• Dispersion

Visual Analytics

• Dashboards
• Evolving approaches to visualizing large volumes of data

 Reporting and Analysis

• Reports
• Planning
• Ad Hoc analysis
• Drill anywhere

The resulting insight from this holistic program enables outcomes that address each of the stated business challenges. Some of these outcomes are:

• Potential customer churn is identified and proactively managed with new offers presented to retain those at risk of churn
• Prospects with a high probability to buy are identified. This translates to new customer acquisition and cross-sell /up-sell opportunities that increase share-of-wallet from existing customers
• Brand risk is proactively managed by monitoring dialog as it happens and identifying areas where damage control is required
• New product and services are introduced and market entry accelerated through new and innovative product ideas that are quickly found through customer feedback
• Customer service and satisfaction is improved and the cost of service reduced as customer issues are found regardless of interaction channel
• Sales and marketing expense are reduce through improved campaign conversion rates and new product adoption rates
• Expensive recalls are eliminated while reducing warranty costs through analysis of data regarding repairs and service requests, providing an early warning on quality and safety issues