Enterprise feedback management (EFM) and predictive analytics are coming together to help organizations discover what their customers are saying about them out on the Web - and to gain new insights that will help them make better, more informed business decisions. At the same time both EFM and predictive analytics are being shaken by the advent of social media.

 

To get a handle on these trends TMCnet recently interviewed Heena Jethwa, predictive analytics strategist for SPSS, an IBM (News - Alert) Company. Here are excerpts from our email interview:

 

TMCnet: Outline the connection between EFM/customer feedback and predictive analytics

 

Heena Jethwa: Predictive analytics informs and directs decision making by applying a combination of advanced analytics and decision optimization to an organization's structured and unstructured data, with the objective of improving business processes and optimizing decisions. Predictive analytics is a valuable and complimentary approach to EFM/customer feedback. [This allows] organizations to combine and analyze transactional, attitudinal and demographic data to gain a full understanding of a customer's wants and needs, thus creating a true sense of customer intimacy.

 

Customer intimacy is often confused with customer experience management (CEM), EFM and voice of the customer (VoC) programs. In practice, all are building blocks to achieving greater customer intimacy and produce valuable customer data. Predictive analytics uses this information to predict future actions and outcomes to make more informed decisions, attract and retain customers, grow revenue, reduce fraud and mitigate risk.

 

Most organizations start their customer-focused activities with either a predictive analytics or an EFM approach. In many cases, both may be undertaken to a certain degree, but each is deployed within different siloed departments. For example, an organization's CRM department might use predictive analytics based on existing CRM data, while the customer insight/market research department focuses on more attitudinal data gathered from surveys or focus groups. Both approaches are very valuable; however, the power of customer intimacy emerges through transforming the siloed philosophy to combine both practices, encompassing all the data across the entire enterprise and, thus, combining these valuable data elements to drive more actionable and accurate insight to determine what customers will want next.

 

Predictive analytics offers numerous advantages for organizations that recognize the inherent value locked within their existing enterprise data. Strategically, predictive analytics provides a quantitative foundation for rapidly identifying, objectively evaluating and confidently pursuing new market opportunities. Tactically, predictive analytics identifies precisely whom to target, how to reach them, when to make contact, and what messages should be communicated.

 

Customer intimacy can produce powerful results, allowing organizations to differentiate themselves in today´s competitive marketplace by increasing profit margins and developing more rewarding customer relationships. Understanding customers and gaining deep insight and foresight can generate numerous direct benefits, including more efficient customer acquisition, recurring cross-sell and up-sell opportunities and increased loyalty and retention. All of which will maximize actual and potential customer value. Latent benefits of an effective customer intimacy solution enable organizations to be more effective in their product innovation. [It accomplishes this by] quickly delivering on the specific product features and functions customers want and on operational excellence by optimizing internal decisions to better meet the needs of various growing segments of their customers.

 

VoC, EFM and CEM form the building blocks that lead to customer intimacy. Predictive analytics uses these data sources to create truer one-on-one relationships, as well as predict future actions and outcomes to make more informed decisions, attract and retain customers, grow revenue, reduce fraud and mitigate risk. By establishing a comprehensive customer data set, including transactional, interaction, attitudinal and demographic customer data, not only are organizations able to deliver personalized customer interaction, but this information can also have a direct impact on the operational and product strategy.

 

TMCnet: What trends are you seeing in EFM and what impact is this having on predicting the outcomes of future customer behavior?

 

HJ: The trends with EFM include:

 

--Using feedback to become predictive and actionable

--Enabling point-of-interaction employees to use feedback to determine the up-sell, cross-sell or retention offers in real time

--Closing the feedback loop

--Including wider data sources, such as unstructured data (e.g. social media, blogs, contact center notes)
--Making Feedback Actionable

 

The issue with feedback today for many organizations is making the data actionable. A growing trend is that they collect feedback, but fail to act on it or do so in a timely fashion. The problem lies in knowing how to tap into that insight and, just as importantly, how to use that keen insight to improve business processes. Predictive analytics delivers the solution, allowing both analysts and business users to easily and quickly turn all of their data sources - including survey results created with data collection software - into positive results.

 

Organizations that combine the power of predictive analytics and EFM can directly impact customer behavior. For example, an organization that believes a customer is likely to leave can determine why and implement specific actions to prevent that defection from occurring. [They would be] basing the retention offer on the existing and potential value of the customer, as well as the customer profile. Through the use of predictive analytics, the retention offer becomes proactive, as opposed to reactive, and that analysis can be compiled based on previous similar customer engagements and behaviors. This proactive, personalized approach will not only prevent customers from leaving, but could turn potentially dissatisfied customers into promoters.

 

Empowering Point-of-Interaction Employees

 

Another trend is that organizations are turning to predictive analytics to help frontline employees make real-time decisions at the point of interaction. Based on the complete, holistic view of the customer, resulting from the collected attitudinal, transactional, demographic and interaction data, organizations can automate and optimize the thousands of tactical decisions that are made on a daily basis. This takes the ambiguity out of what decisions should be made and ensures consistency in the way which to interact with customers. It also empowers more employees, such as call center agents, to make those critical decisions - decisions that are cost effective, yet also based on what is best for that individual customer.

 

Closing the Feedback Loop

 

Many organizations collect feedback, but now a key step in this process involves closing the loop and telling customers how their input will be used - emphasizing the importance of their viewpoints and insight, and stressing how valued it is by the business. For example, collecting feedback on how to enhance a particular product and then sharing with the customer the changes that were made based on that feedback is very empowering. This closing-the-loop practice makes customers more likely to want to be involved as they can see and understand the benefits of the feedback they contribute. Ultimately, this gives customers the confidence to provide more accurate and more detailed information.

 

Including Wider Data Sources

 

Feedback no longer only relates to surveys these days, as valuable data can also be found on social network sites, blogs and contact center notes, etc. This data from the Internet often is not always solicited, as customers are having direct conversations with other customers. Being able to gather and analyze this valuable data is important to ensure a wider view. Incorporating other sources of data also improves the reliability of the information, which then leads to more accurate predictive models that determine what actions are to be taken next.

 

TMCnet: Social media has rapidly emerged as a new customer channel and a vital source of feedback. Outline how it has affected EFM and in turn predictive analytics.

 

HJ: Around the world every day, people are talking and sharing their opinions through the estimated 1.6 million blog posts, on diverse social networking sites and in community message boards. That's a goldmine of information that organizations can tap into as they improve business performance with the use of predictive analytics. And it hasn't gone unnoticed. According to independent research company Forrester, 69 percent of retailers surveyed indicated they have already implemented social network pages (e.g., Facebook (News - Alert), MySpace.com) and 24 percent planned to implement/enhance them in 2009; 58 percent offer customer ratings and reviews, and 25 percent planned to implement/enhance them in 2009; and 54 percent employ microblogs such as Twitter and 28 percent planned to implement/enhance them in 2009.

 

Predictive analytics enables organizations to analyze that unstructured data -sentiments and relationships that are embedded in e-mail, contact center notes, RSS feeds, surveys and social media sources - and combine it with structured data, such as demographic and transactional data. [This gives a] complete understanding of their customers', employees' or constituents' future behavior. This more complete view of the customer leads to more accurate results, better predictive modeling and deeper insight. This allows organizations to become more effective when reducing customer churn, improving productivity, fighting crime, detecting fraud and increasing marketing campaign results.

 

The addition of social media outlets can have a huge impact on a company's EFM program. Social media is a very agile and dynamic environment and organizations can use it to extend their EFM practice by comparing internally collected feedback to what is being written on these sites. As a result, they can determine whether the social media data is fully representative or just reflect only a couple of unhappy customers.

 

Another benefit is that organizations can analyze how their products/services compare to the competition. Acting on feedback is essential, but a company must also validate and apply context to ensure the action taken is necessary and will actually impact customers in a positive way.

 

The problem organizations have with these unstructured forms of data is the ability to turn it into insight so they can find that "needle in the haystack." This is where they turn to predictive analytics, to harness all this data (unstructured or structured) and deliver insight and foresight. This valuable information can then be included in the creation of predictive models, providing further accuracy and moving toward a holistic view of the organization's customers. Time is of essence and making sense of the wealth of data that exists requires predictive analytics, including text mining of the unstructured data.

 

TMCnet: What changes are in the works with your solutions in response to these issues and developments?

 

HJ: IBM just announced IBM SPSS (News - Alert) Decision Management software that places the power of predictive analytics directly into the hands of business users to automatically deliver accurate, high-value, high-volume decisions at the appropriate point of customer interaction.

 

With its customized module for customer interactions, organizations can better retain customers, grow revenue and drive profits by creating a personalized experience for every inbound customer and prospect via call center, web, point-of-sale or e-mail. Forrester (News - Alert) estimates approximately 60 percent of companies evaluate their capabilities to be poor/below average for customer interaction management, and 62 percent cannot easily manage real-time scoring of customers.

 

For example if a high-value retail banking customer calls into the contact center to complain about a product or service, the new software may predict, based on the customer's data, that the individual is likely to churn. The information about the complaint, combined with the customer's history, can then be used to create a customized retention offer on the spot.

 

IBM SPSS Data Collection feedback management and survey research software continues to grow, focusing on gathering feedback across multiple touch points, making it easier for users to gather feedback and also enabling integration across the IBM SPSS portfolio to drive customer intimacy.

 

Another member of the IBM predictive analytics portfolio, IBM SPSS Modeler data mining and text analytics workbench, has been enhanced to address the influx of data now available through social media outlets. With Modeler's capabilities, organizations have the ability to monitor changes in consumer, constituent and employee attitudes, uncover deeper insights and then predict key factors that will drive future customer acquisition and retention campaigns. As an example, companies can now extract sentiment from the use of emoticons and slang terminology that people often use in describing their view toward a product or service.

 

For social media, in particular, IBM SPSS Modeler enables organizations to directly access text, web and survey data and integrate that data into predictive models for more comprehensive recommendations and better business decisions. It uses natural language processing to allow clients to pull key concepts, opinions and categories relevant to their business from these data sources to uncover deeper customer insights.

 

With IBM SPSS Modeler, organizations can:

--Create and evaluate sophisticated models easily and visually: SPSS Modeler provides a variety of pre-built algorithms to create models easily and intuitively. Users can quickly view models interactively and apply a variety of analysis and visualization techniques that help them understand and communicate the results of their analysis efforts.

 

--Greater flexibility for business and professional analysis: Firms can gain greater flexibility for business and professional analysis with Modeler's automatic data preparation and automatic modeling functionality, user obtain results quickly. Full integration with IBM SPSS Statistics enables them to use best-in-class statistical analysis and reporting to support their data mining efforts from a single interface.

 

--Obtain support for enterprise standards and technologies: In-database mining allows organizations to leverage their investments in operational databases fully while enterprise features such as password protection and single sign-on ensure adherence to corporate governance over data and models.