Social media data and the customer-centric strategy
During the last five years, the evolution of social media and its influence on business has been unprecedented. The bottom line: the customer is no longer outside the organization; the customer is now shaping the organization. If you want to thrive in this economy as a product or service provider, evolving the business requires a customer-centric approach.
Applying a customer-centric strategy
What does being customer-centric mean? Providing a superior customer experience across an organization’s product and services, and providing the customer with indisputable value and satisfaction is the simplest form of being customer centric. One can argue that many organizations today adopt this approach, but the reality is only a handful have really embraced the concept and made gains. Others have tried and failed, and yet others probably offer more of a lip-service approach rather than a lifecycle approach to customer-centric evolution:
The biggest challenge for acquiring customers is the crucial step of creating a value quotient with the customer. In the past, this step was accomplished through multichannel marketing using mail, catalogs and coupons. That strategy was well suited for a product- or brand-driven customer strategy, but now the equation has shifted to a customer-driven product and brand creation strategy.
This fundamental shift has created the need to treat the customer as a stakeholder in the business. To get this approach correct, organizations need to understand the new customer, who is very savvy with social media and can be very influential about your products, services or both to a large network of people. To understand the social media impact and integration, you need to learn a few terms and a few important underlying concepts or surrounding themes:
- Crowdsourcing: First coined by Jeff Howe in the Wired magazine article, “The Rise of Crowdsourcing” (June 2006), crowdsourcing is a concept in which people form Internet communities based on shared interests. The basics of the concept revolve around degrees of separation among individuals that have been reduced greatly because of the Internet, and this reduction has created a virtual crowd. Traditional crowds have long been a powerful force in creating and paving ways for a brand and its associated products and services. The new crowd based on communities forms a very powerful vehicle that can be tapped by an organization to help drive creating its products and services. The net result of such an endeavor helps foster brand loyalty and increase the market presence from word-of-mouth marketing (WOMM).
- Word-of-mouth marketing: This kind of person-based marketing is not new. Before the mass commoditization of telephones and television, a marketplace emerged in every community, and it had a wide range of products and services. The vendors in this marketplace relied on a loyal set of customers, who would bring in new customers in the form of friends and family. Today, we call this WOMM. The difference today is that WOMM behavior happens on the Internet and in community forums, and shared-interest websites and personal websites such as Facebook, LinkedIn and Tumblr. This behavior is a key trend that needs to be measured.
- Long tail: As a statistical term, long tail is a disruption to the normal Pareto or Gaussian distribution, in which the larger population of the statistic rests in the tail. Chris Anderson popularized the long tail in an October 2004 Wired magazine article, in which he mentioned Amazon.com and Netflix as examples of businesses applying this strategy. The long-tail strategy is driven by volume of business at lower cost, resulting in higher profits. A number of organizations have since embraced this model.
To embrace the new type of customer, organizations need to understand these three concepts and apply them to their business models. Such an exercise helps establish a business case for creating the program popularly called voice of customer. This type of program creates a sense of stakeholders among the customer base and fosters a growth of community around the business—thus enabling the brand to succeed in new markets, in new products and new services areas. This model is not something that all organizations can benefit from, but conducting an experiment will always provide a basis for making the determination.
Several innovations have emerged during the past five years to benefit organizations: Apache Cassandra; Apache Hadoop and its Apache HBase, Apache Mahout, Apache Pig and R ecosystem; Google MapReduce; high-speed disk arrays; in-memory technologies; and second- and third-generation data warehouse appliances. All these put together in a solution architecture enable a technology platform for social media integration into the organization.
Bringing social media into the fold
We can consider a fictitious company to take a look at tapping into social media and how it can benefit any organization in a services industry. This fictitious organization decides to implement a voice of customer program that seeks to promote a better understanding of customers and their sentiments expressed in chat, email and phone conversations with call center representatives and in survey responses. To accomplish this program, the organization needs to follow these steps:
- Establish several contact points or listening posts to hear and understand the customers and their solutions and grievances. The first important step in understanding the customer is to listen to the customer.
- Extract the data from these listening posts and examine the trends expressed in the conversations.
- Integrate the result set into reporting and analytics engines through data integration.
- Visualize the trends and metrics from the data integration.
- Provide the data to relevant business users to derive the intelligence and understand the customer intelligence.
Fast-forward to the next step: The organization has implemented a technology solution platform that can provide rich insight into sentiment analysis. The software can capture speech and convert it to text, and further perform analysis of the data within the text to gather the sentiment of the conversation. Following this step, the software can categorize the conversation tone as positive or negative along with the associated keywords and trends that led to the inference. While this approach marks a huge step in connecting to the customer, it also represents the following unfortunate scenario.
The customer sentiment expressed in the conversation is not categorized based on the context. For example, the customer makes the following statement: “I have been very frustrated with a particular service offering and the number of times I had to follow up on it. I’m not going to engage in the pursuit any further because support is minimal. I’m very disappointed.”
In this situation, the sentiment analysis software helps verify that the sentiment is negative, the reason is minimal support and the customer is disappointed. What the business user is likely to miss here is the big picture that text mining and analytics look at: the customer is unhappy with certain services because the customer had to follow up and received minimal support. In this context, the customer wishes to cancel the service. This big picture is contextual in nature, but there are several soft links here:
- How many other services the customer uses and might cancel
- How many other people are in the customer’s network that the customer might influence
- How many other customers have expressed such concerns and canceled services
Unless these gaps are addressed, the value from the voice of customer initiative is deemed primitive.
The second listening post is that customers are likely to follow up the conversation with emails. For example, a customer writes an email with a Customer Service Feedback subject line and this information:
During the past 30 years, I have been a customer of your services. While the relationship has had its share of highs and lows, recently your customer service team has been performing very poorly. The response times have been lagging, a lack of urgency to close questions exists and the intent is to sell more services and not address problems. While I appreciate the self-service channels you have opened, this direct channel has deteriorated. Should this trend continue, I will be forced to consider other alternatives.
Several key problems and associated sentiments and comparisons are mentioned in this email. If the customer had written this email and then within a 30-day time period followed up with a call the organization, time was available to react and raise a potential attrition alert had the email been parsed to know the customer's intent to move on.
Why is this scenario important? If the customer has 50 friends who hear his story, chances are the services organization could experience a potential loss of all 50 customers. Or, over a period of time, loss of groups of customers could occur that may lead to revenue loss. But if the customer were to express the grievances in a social media forum, brand reputation is at stake and more customer attrition may possibly occur.
To increase more actionable insights, the organization needs to go beyond just sentiment analytics to integrate data across multiple channels including email and social media analytics. Not only can this approach bring better insights, but it also can provide the organization with the ability to predict and model customer behavior and be prepared to react in an enhanced fashion when such situations arise. Additionally, the data and analytics enable the business user community to better address its knowledge base and better aid its customer interactions.
Taking an integrated approach
A high-level integration approach can combine processing structured data from online transaction processing (OLTP) and operational data store (ODS) systems; processing extract, transform, load (ETL) rules; and processing unstructured data from sentiment, email and social media channels:
The advantage of combining the data from both sources is gaining a holistic view of the customer. Existing master data management (MDM) and metadata collections enable the linkage between the different types of data.
Enabling cross-sell and up-sell opportunities
Assume that the fictitious organization discussed previously has concluded a campaign to reach out to its customers about a new, integrated portfolio services plan. The campaign has resulted in several calls from the customer community to call center and business services teams. In this scenario, a need exists for real-time access to the customer, campaign, and social media and listening post data. This integrated data set can provide a clear roadmap for additional cross-sell opportunities. Such data can be analyzed and visualized in an all-in-one mashup that can be consumed by the business service executive or call center executive. This technique then results in an enhanced customer experience and drives a true customer-centric approach. The end result can be measured by gains in revenue for the organization.
When customers call, the system loads in their information and provides the following data:
- Customer LTV
- Last transaction date
- Last product purchased
- Last campaign responded to
- Customer stickiness
- Customer life events
- Customer cross-sell opportunity
- Customer social media affiliations and presence—as a traceable or generic customer behavior model
When presented to the business services executive or call center executive, this data forms a guiding portal for them to understand customers, their current situation and the relevance of their calls. They can then answer questions with a more focused, customer-centric approach, thereby providing an excellent customer experience. Even more data can be extracted from the content management, contracts and other financial data that helps provide an enriched customer experience.
Providing business benefits
The business benefits from the integration exercise include:
- A 360-degree view of the customer
- Revenue leakage identification and recovery
- Cross-sell and up-sell opportunities
- Enhanced customer connect
As shared in this article, the power of integrating social media data enables the evolution of a customer-centric approach to build business brands. Enhanced connection to the customer enables a better wallet share and creates a sense of importance for customers.
Join more than 1,500 analysts, business professionals, engineers and scientists for three days of in-depth learning and networking at Spark Summit East 2017, 7–9 February 2017, in Boston, Massachusetts.
Editor’s note: This article is offered for publication in association with the Big Data Seminar 2017, 16–17 November 2017, in New York City, at the Hotel Pennsylvania, and sponsored by Data Management Forum. Additional information is available in the Big Data Seminar flyer.