Predictive Analytics to Improve Your Customer Retention in Contact Centers

Even a small reduction in customer churn leads to a considerable impact on the bottom-line by helping companies retain their valuable customers. Further, proactive engagement also leads to increased customer satisfaction levels. Dynamic and proactive customer retention strategies based on predictive analysis go a long way to increasing the effectiveness of your customer retention drive by enabling enterprise-wide actions.

A loyal, paying customer is the best success any type of business can have. Everybody knows this fact. However, when we start getting into nitty-gritty of retaining best customers, it starts getting complicated. Customer retention is often driven by overall customer experience, which itself is shaped by several distinct experiences a customer has throughout the customer journey. Most importantly, these customer experiences are not about rational experiences a customer has. How the customer feels after an interaction or a call to the call center probably defines the customer’s experience a lot more than what standard KPIs are being reported.

So, how can companies ensure that their customers feel valued and keep using their product or service?

A Contact Center is Critical in Shaping Customer Experience

What drives the customer sentiments towards a business? Our research using sentiment analysis of social media mentions how different brands confirmed that overall positive mentions of a company show a strong correlation to the positive mentions of “customer service” of that company. The following graph shows this correlation for various industries –

How to Ensure that Contact Centers Drive Customer Retention

The answer to this question lies in the interactions that happen at the contact center. The contact center is the most visible and approachable part of any business for a customer. Over the course of the customer’s journey, he or she approaches the contact center multiple times, for multiple reasons, and via a multitude of channels like emails, website chats, customer service calls and so on. Through all these interactions, customers voice their concerns, share their experiences, give their feedback or register complaints. Further, the customers also expect to get a satisfactory response when they contact the business.

All these interactions are a treasure trove of data about the needs and wants of the customers, their behaviors, pain points and expectations. If mined properly, these interactions reveal very crucial insights about the intention and behavior pattern of customers.

Typical Retention Approach – Performance Management

The typical retention program companies deploy are reactive in nature. When encountered with a customer who wants to cancel their services, the call is typically routed to a ‘save desk’ which creates an offer for the customer and tries to win them back. By managing the performance of the save desk agents with various training and coaching tools, contact centers aim to keep customer retention numbers high. However a segment of customers can never be saved with this approach if they already signed up with a competitor before the save desk attempts to retain them.

Typical Customer Retention


If we inspect typical activities that happen prior to a customer deciding to cancel, typically a customer calls to get their issue resolved from the service side of the contact center. Over a period of time, some issues are still left unresolved, resulting in dissatisfied customers. Often even these disgruntled customers give the company a chance, and express their dissatisfaction in their calls and expect quick resolution to their issue. However, if the issues this customer is facing are unresolved after a few calls, he or she finally decides to cancel the services and switch to a competitor. Until this point, call centers make no differentiation between this customer and others. When an actual cancellation request surfaces, the customer is then routed to the save desk. This team is responsible for convincing the customer to stay with the company by offering special discounts and deals, which is often a very expensive way to retain customers.

Besides the cost of retention issues, it might also be too late to ensure that the customer is retained at the time the save desk makes an offer. The focus here is on trying to minimize the customer churn, which is a reactive and less effective approach because the customer has already decided to cancel prior to this, and in some cases, they might’ve already signed up with another service provider.

A Better Approach –Predictive Customer Retention

Applying predictive analytics to mining contact center interactions uncovers new opportunities to approach customer retention effectively.

Predictive Customer Retention

Identifying a dissatisfied customer

All the interactions a customer has with a contact center have enough clues about their satisfaction or dissatisfaction. Typically businesses fail to leverage these interactions as a source of customer intelligence. Predictive analytics enables identifying these clues and categorizes the customer as ‘satisfied’ or ‘dissatisfied’ based on the their previous interactions, and then assigns a propensity to churn score.

The typical indicators of an at-risk customer are:

  • Large number of calls
  • Expression of dissatisfaction
  • Competitor mentions and comparisons
  • Enquiry of alternative price plans
  • Customer experience issues

Using interaction analytics and predictive churn modelling, a churn score can be created for each customer based on dissatisfaction identifiers like the ones mentioned above. When the churn score is above a pre-defined value, the customer is tagged and then various recovery efforts can be made either within the contact center as part of future interactions or, even better, a proactive outreach to the customer to resolve their issue.

Proactive customer retention strategy

The next time a customer calls, the call is automatically routed to expert agents with higher skills and empowerment on the basis of churn scores. With their skills in empowerment, empathy, resolution abilities and communication skills, the agent stands a better chance of offering a satisfactory solution to the customer. Note that this effort takes place even before the customer decides to cancel the service.

As you can see, the customer retention approach becomes proactive instead of reactive with the application of predictive analytics. The proactive retention efforts aim at identifying a dissatisfied customer and offering a timely solution to his or her concerns before it is too late. These programs can be implemented with very little additional cost, with intelligent segmentation and routing within the call centers based on risk scores.

Improving customer experience

Big data analysis of the customer interactions has another benefit – fine tuning operations to enhance customer experience. With the approaches discussed above, not only do customer churn scores go down, but other contact center KPIs such as satisfaction scores, first contact resolution scores, etc., typically increase as well resulting in a good return on investment (ROI).

In addition to improving contact center specific KPIs, this same intelligence from contact centers can identify the drivers of customer experience and the root causes behind them, as most customer experience issues lead to a contact center interaction. These insights into the customer experience help the contact centers to collaborate with other business units with clear action plans to enhance the overall customer experience across many other touch points without involving contact centers.

(The original article was published on

You can also download PDF version of this article here

The Power of Text Analytics-Driven Segmentation

Big Data is discussed everywhere today. Big data IS everywhere today. The idea of big data is compelling. In many cases big data is overkill. Although big data can help in identifying WHAT trends exist, the main question will always remain- WHY. Small data enters the picture here. What’s small data? Small data connects people with timely, meaningful insights (derived from big data and/or “local” sources) . Often visually organized and packaged, small data is accessible, understandable, and actionable for everyday tasks.

Data-driven Segmentation

Data-driven segmentation derives small data from big data by relying on the ability to create and identify actionable data. This data demonstrates the worth of a customer/prospect to a company, how often they interact with the brand, their purchasing habits and preferences, their possessions and, most importantly, their feedback. Organizations armed with this information can drive the message more acutely and create an exponential effect on marketing efforts as well as improve their customer experience tremendously.

Segmenting the data by any number of factors and parameters allows you to listen to your customers effectively. When you regularly review your VOC program in a way that incorporates data segmentation according to customer journey topics, you give yourself the ability to refine and optimize the strategies.

Restaurants and Text Analytics

We recently carried out a research on the restaurant industry using twitter data. We employed the following methodology:

  • Gather twitter data of the restaurants (TGI Fridays and Chili’s) for a period of 90 days (Q1, 2015).
  • Apply Voziq’s unique categorization technique to filter out noise and segments data into useful categories – Data Segmentation
  • Use Voziq’s visualization techniques to correlate these categories (data segments)
  • Also Using Voziq’s sentiment analytics engine to get sentiments for user comments.
  • Compare and spot the trends.

1.3 Picture2


What did we observe?

  • Through Sentiment Analysis:
  1. In general, people like the quality and taste of food in both the restaurants
  2. Customer service has not fared well.
  • Through Category Correlation:
  1. Customer Service issues center around wait time and staff behavior in both the restaurants.

Now these are pretty specific and targeted problem definitions. The purpose of data analysis in business is to hunt for the correlations that identify issues and drive sales. Once you find these correlations, you can determine what to emphasize in the future for each segment. The more you segment your data, the more opportunities you have for correlating segments with better accuracy, timing, format and content.

Text Analytics-Driven Segmentation Drives ROI

To compete in today’s digitally-driven world, businesses across all industries must achieve a 360° view of the customer in order to successfully market to their specific needs. This in turn drives revenue growth. Such a comprehensive vantage point rises out of integrating data across all platforms and from all available sources to realize true data-driven segmentation. Investing in data unification solutions to achieve data-driven segmentation and a superior customer experience, regardless of the device used to access information, is the next step toward building a better VOC engine and achieving the ROI.


Image Credit: Flickr


Customer Experience with Text Analytics

Why Market Analytics Trumps Market Research?

‘How can I please my customers?’ is a million-dollar question. Companies struggle to understand how their customers make purchase decisions regarding the products and services offered by the company. As most studies have established, the decision-making by a customer is mostly irrational. There may not be any logic to it.  Due to the explosion of social media, messaging apps etc., customer journey has become an extremely complex phenomenon. Customer expectations evolve too rapidly due to the access to information and social influence. A lot of these factors, apart from the actual product or service, affect customer decisions.

Since the very fate of your business is tied to the understanding of the customer, companies put a lot of effort and money into researching their customers, and are constantly on the lookout for more advanced and efficient methods of dealing with the deluge of data that is being generated every second.

This calls for a need to upgrade the techniques and methods of carrying out research to understand customer behavior as traditional market research fails to deal with the explosion of data – the opportunities as well as the complexities of it. Companies need to differentiate themselves in order to gain a competitive advantage over their rivals and also effectively streamline their businesses. In this such scenario, expanding the focus from market research to market analytics can be a game changer for businesses.

How to differentiate between the market research and market analytics?

McKinsey & Co. suggests that companies can be a part of the research revolution by paying attention to four key areas: 1) Leverage the Internet to rapidly obtain details about consumers; 2) Keep the limitations of focus groups in mind; 3) Learn how people shop; and 4) Link consumer attitudinal and behavioral data.

Based on the above, certain questions related to differentiation between market research from market analytics arise.

  1. How is the value derived from big data?
  2. How is data integrated to gain actionable insights?
  3. How to link consumer attitudinal and behavioral data?
  4. How is the immensely vast internet leveraged for real-time data and analysis?

Market research with its traditional methodologies has limitations and is unable to answer the challenges posed by the rapidly burgeoning data and the need to integrate customer data from various data sources and to analyze it in real time.

Traditional Market Research Approach

Traditional Market Research primarily consists of getting an overall feel of the customers by addressing strategic questions broadly raised by the management. The research, in that sense, is largely driven by what you already know. Market research  broadly answers questions related to business aspects like product awareness, availability, packaging, placement, pricing, competition within the market, general sentiment of the customers within the target audience, brand loyalty, brand positioning, customer satisfaction aggregate and segment awareness, message effectiveness, and advertising recall among others.

Thus, market research helps in painting a broad, holistic picture of the customers and the marketplace. This is done by studying a sample from the universe. This study is time bound, and analysis done using statistical packages like SPSS and SAS. The findings of the study are presented to the management with or without strategic recommendations based on objectives.

Technology-Driven Market Analytics

The lives of people are increasingly being shifted to the virtual world. Online Media reveals anything and everything that a company needs to know about its consumers and the market. People discuss, comment, critique and post about products, services and provide details based on their general experience. They also rely more on the publicly shared opinion than a company’s advertised claims. This makes online media the most powerful platform for businesses to leverage their immense potential. However, to tap into the vast resource base, the traditional research techniques prove to be incompetent. This is exactly where the strength of analytics comes into the picture. Companies need advanced market analytics capable of transcending the confines of primary marketing research.

How Market Analytics trumps Market Research

There are a couple of factors which distinguish market analytics from traditional market research –

  1. Technology-driven: Market analytics is technology driven. This means it’s fast, accurate and can process large volumes of data significantly faster
  2. Discovery-oriented: Market analytics is more geared towards discovery of customer insights and prediction of customer behavior
  3. Big samples: Owing to the vastness of online media, the sample size for this type of analytics is virtually unlimited
  4. Data integration: Market analytics also has the ability to integrate customer data from multiple sources like contact centers, customer surveys as well social media
  5. Real-time analytics: Market analytics generates and communicates the customer insights in real-time via custom dashboards, email, and smartphone alerts
  6. Text analytics: Market analytics lends you the ability to analyze text data at scale
  7. Predictive analytics: Market analytics lends you the ability to map customer journeys more accurately, predict customer behavior and take pre-emptive actions

All these aspects of market analytics lead to very accurate customer insights which lead to concrete actions that deliver a very high ROI.

5 Steps to Control Customer Churn

Effective Customer Churn Management Process

Controlling customer churn is vital for the success of any business. To improve customer retention and customer loyalty, companies need to first analyze customer churn and quantify its impact. This provides insight into the different customer groups that may need to be better addressed or need specific attention. Based on the industry vertical and the market, companies need to design a predictive churn model to identify potential customers who have a high probability of churn. This analysis can be the stepping-stone to improving customer loyalty and customer retention figures through marketing initiatives aimed at reducing the impact of customer churn.

Dr. Vasudeva Akula, CEO of VOZIQ, has elaborated on this topic in a LinkedIn post. Here is the link to the post – 5 Steps to Control Customer Churn.