What changes are in store for the contact center in terms of analytics? This article series explores some of the major trends that will impact contact centers of all types over the next ten years. Read the first article.
In the 1970’s, marketing operations, customer service teams and collection and recovery operations relied on direct mail campaigns to achieve their goals. The direct mail campaigns were conducted by either “blanket bombing” customers with direct mail or mailing only to a targeted list of customers. In the 1980’s this evolved into single-agent telephone contact attempts, and included having marketers use the public phonebook to create their calling lists. In the 1990’s marketing and operations became more technologically efficient and began using "robo-dialers" and predictive dialers that allowed agents to reach a greater number of respondents. For the inbound call center, the advent of inbound interactive applications and IVRs meant that a greater level of automation could be used. The 1990’s and early 2000’s was a great time for call center technology development; however, all of the technology that was created didn’t make the contact center any "smarter." In the natural evolution of the contact center, the time is now for the development of an intelligent contact center that is able to take all of this additional technology and use it in a more effective manner.
Defining the Client and Strategy
The first essential step in developing an analytical approach is to be sure that all members of the team, across all functions, have a clear understanding of who the company believes that its clients are and what its strategy is. On the surface this sounds easy enough, however, have you ever really considered who your client really is, or what the strategy for your company is?
If your operation is focused on Loyalty, how is success measured? For many operations, this means cutting down defection from the product or "churn." Then, it would be important to understand what the drivers for churn are for the client base and then to understand what can be done to mitigate that churn. To have that level of understanding is only part of the puzzle. The rest of the puzzle comes from understanding what levers you have to pull and how to use your operational arm in the contact center to affect that change in churn behavior. It also has a significant amount to do with understanding what contact channels to use and when to use them. These are only some of the variables that need to be considered when developing a retention plan.
Sticking with the Loyalty example, Figure 1 offers some ideas of just a few of the variables that a Loyalty Strategy Team and Operational Team may consider, and provides some good examples of how to get a fairly comprehensive view of the client. Once a client view is obtained, segmentation of the client base can then be achieved that further predicts the likelihood of a client churning or defecting from your product. With this segmentation model, the Strategy team will be able to determine what offers to provide to the contact center for each segmentation as well as determine which contact channel to utilize. Different contact channels can be used based on risk, as well as different level of offers.
Once you fully understand who the client is and what strategies are required to address the client’s needs, the next step would be to analyze the results of contact efforts. By determining the correct channel for the correct segmentation, then adding that to the correct strategy to apply for the client’s needs will give your loyalty efforts a significant advantage when matching the needs to reduce churn with the maintain or reduce cost while you do so. This example of analytics for a loyalty team is one example of successfully understanding who your client is, as well as understanding what the strategic intent of the call center is.
Understanding the Analytical Process
There are two specific ways of understanding the analytical process. The first is to have a full grasp of the Scientific Method that we all learned in primary school (and have probably since forgotten). The second is to develop an analytical lifecycle. No model can be put in place permanently and they need to be “living” strategies that change as the variables change. For success in any analytical process, it cannot be approached as a "fire and forget" procedure - it should be reviewed at regular intervals. There is no magic formula as to when these reviews should take place; rather, it should follow along with the terms of the method.
The Scientific Method
The Scientific Method of approaching the analytical process is a simple one, and this process still holds true when it comes to analytics and a structured analytical approach. Figure 2 demonstrates the flow of the scientific method. Many of the same processes hold true when approaching the development of an analytical process for business intelligence and call centers.
Once you understand your client and have developed a strategic approach, it is important to start developing segmentation models. The process shown in Figure 2 is the same step-by-step approach that a strategist/analyst would take to come up with a model. Developing a standard, methodical approach that can be repeated and referenced throughout the process mitigates the number of mistakes and biases that are encountered. It is important that when developing these approaches to also build a segmentation model where a control group is also identified with similar relationship values so as to further ensure that any potential biases are accounted for.
Then, the experiment becomes taking the already defined variables and applying different techniques to them. Going back to the loyalty example, choosing a segment of the cohort that show similar characteristics and then applying treatments to all of them - such as contact strategies coupled with offers - will generate a variety of results. Those results can then be analyzed and strategies and offers refined.
As the experiment is repeated and strategies are refined, models will start to develop. These models should give a clear indication of what the available “levers” are to pull to achieve a result. The constant challenge of the strategy team and their analytical group is that there will always be environmental changes that impact any or all of the variables chosen. For this reason, a sound analytical governance process should be developed. In other words, tests on the established models and segmentations should be conducted at regular intervals.
This discussion on the analytical approach is the foundation by which the other articles will be built on. It is important to remember that this series is meant to provide means by which call center professionals and stakeholders can open discussions around how analytics can help their business and their operations. The examples given may be for one industry, but the concepts can be broadly applied.
In the next article in this series we will discuss Voice Analytics and how new approaches in this area will help shape the strategic and operational models of many different and diverse contact centers.
Rob Archambault is an ICMI-certified call center professional at Archnau Communications and Consulting. firstname.lastname@example.org. @archnau_comm. Robert Archambault on Google+.