The key point behind customer service is customer satisfaction. Today most of the customer contact channels are built around a very basic understanding of what the entire customer base might need in general. This is particularly true in the case of the voice channel where the interaction sequence is very rigid, and thus the customer is constantly vying for that human interaction i.e. agents. The need of the hour is to customize the experience for individual customers, to ensure that they find what they are looking with minimal effort. With the right mix of analytics and prediction the required human touch can be added to the voice channel.
The key to customized interaction is data. Though customer data is readily available with everyone in the service industry, they fail to capitalize on that. The data that we are talking about is both the customer profile data & customer interaction data. This customized interaction on the basis of data is possible through Predictive analytics. Predictive analytics employs many techniques like statistics, data mining, modeling etc. However its application to the voice channel in the service industry requires a deeper understanding of the business.
There is one interesting story that was run by Forbes a few years back regarding the use of predictive analytics (http://www.forbes.com/sites/kashmirhill/2012/02/16/how-target-figured-out-a-teen-girl-was-pregnant-before-her-father-did/). Though the story is under contention, the summary of the story is that Target (an American Retailer) started a system where they analyzed the customer shopping pattern. Based on a pre-defined metric they determined if a particular family had a baby due, and sent them coupons related to baby items. One day an angry man walked into the store demanding why those coupons were sent to his daughter, post which Target apologized. Later on the same man turned up apologizing, stating that his daughter was indeed pregnant.
This is something that can be incorporated for different verticals in the service domain. For example, when a customer logs a complaint through other channels, he usually calls up the IVRS (Interactive Voice Response System) for a status update on that complaint. The reason why the customer is calling can thus be identified and instead of a menu, the customer will first hear the status of the complaint & the expected resolution time. For a utility company that is facing a wide area downtime, it is usually observed that customers call in to get an update on the resolution. In this case customers identified to be in that particular service area can be directly updated regarding the downtime & resolution. When a new service is launched contact centers face increased call volumes regarding activation & complaints. The same can be identified dynamically and placed in the IVRS so that agent resources are optimally utilized and calls end on the IVRS itself.
Another implementation is dynamic placement of menus in the IVRS call flow depending on the customer metric. Customer metric may include age, location etc. Definitive weightages will be assigned to each metric. And the placement of menus in the call flow will be modified in real time. The business aim of this logic is that maximum customer queries should be resolved on the IVRS itself.
Cross selling is another area in the voice channel where this can be implemented. It is a system which is widely used by a majority of the e-commerce vendors. Depending on your previous shopping history & the search history a list of similar items you may like are presented to the user. Additionally there are a number of correlated items, for example, if you buy a pillow you are most likely to buy a pillow cover as well. Similarly for a majority of service providers promotions can be customized based on the customer profile, what he has and what he is most likely to subscribe to. Based on his interactions on various channels and current trends, specific customer segment can be targeted for an outbound campaign. This will ensure a higher conversion and success rate for a particular campaign, which in turn translates to increased revenue.
Another important point is that this system leverages data from all channels such as – voice, SMS, web, E-mail etc. Analysis of data across all the channels ensures a unified experience for the customer where regardless of the channel utilized, the customer is presented with the same and updated response.
An important objective behind predictive analytics is customer retention and brand recognition. The consumer expects superior service and that extra personal touch in their interaction with the supplier. Both these criteria can be fulfilled with the help of predictive analytics. A very recent interaction with a multinational bank made me realize the value of that extra mile which makes all the difference. I had about 8 leaflets remaining in my cheque book. Usually to receive a new cheque book we need to either fill up a slip at the branch, or register via the IVRS or ATM. The aforementioned bank acted proactively and mailed me a new cheque book.
I think Proactive Intelligence built into the customer contact channel using predictive analysis is the way forward. Today most of the customers are using static IVRS flows that do not take into account the dynamic nature of the customer behavior and industry trends. By leveraging the same we can ensure that we provide that extra level of differentiation and value to various verticals and customers.