Getting to Know Your Customers
Author:David Woodhead and Nina Weina Jiang
HOW TO GET REAL INSIGHT FROM YOUR OMNICHANNEL CONTACT CENTRE
Many companies have invested in sophisticated contact centres to allow their customers to reach them through whichever contact channel suits the customers. But no matter what the contact channel may be — phone, email, Web or social — customers just don’t ask questions in a structured way. They often make multiple enquiries at the same time. They may ask a question and complain or compliment in the same interaction.
These unstructured interactions take place every day, all over the world. And they represent masses of under-utilised insight into what your customers need and how they behave.
Getting at that insight can be tricky. Today, two-thirds of call centres still use a random selection process for evaluating agent calls, while just 5 percent have a process for collaborating with agents to pinpoint the calls they believe are the most insightful, according to a recent survey of more than 100 contact centre professionals conducted by Sift Media.
As a result, while many contact centres are data-rich environments in many ways, these unstructured interactions are infrequently categorised or evaluated, and hence have little meaning until you start looking at the verbatim free-text information contained in emails, tweets or contact centre agent notes. When you apply natural language processing to mine the requests, requirements, sentiment and emotion contained in these unstructured narratives, a wealth of new insight becomes available.
AN INDUSTRY EXAMPLE
Water industry regulator Ofwat continually challenges water companies to improve customer service to their more than 50 million consumers in England and Wales. The Ofwat service incentive mechanism (SIM) has been designed to motivate water companies to improve the quality of service and value for money for consumers. Quality is measured through customer satisfaction scores that seek to show how well the water companies are serving customers by getting things right the first time and resolving complaints quickly and effectively.
Significant consequences can occur if water companies’ SIM scores fall below par.
First, there is an emerging competitive issue. Starting in 2017, the water market will be gradually opened to competition in the same way the gas and electricity utilities are now. This means that consumers and business customers will be able to vote with their feet in a new retail marketplace.
Second, but most significant today, is that Ofwat, in its regulatory role, has the power to determine the level of profit that water utilities are allowed to retain in any fiscal period. This has the effect of placing a direct financial cost of up to 12 percent of revenues on the water business if they are at the bottom of the SIM tables.
TACKLING THE ISSUE
CSC has been working with one of the largest water utilities in the United Kingdom. When we started, managers at the utility said they knew their SIM scores needed to be improved. The problem, however, was that they didn’t know what customers were really asking for or where to focus their resources to achieve the greatest benefit.
CSC immediately noticed that although the contact centre records had been creating an excellent source of insight, this source was difficult to work with. The contact centre system asked agents to record the name of the caller, select a single call-reason code, and enter free-text notes to describe the problem and its resolution.
Each call could have only a single reason code, and with more than 60 codes to choose from, many of which were duplicated, it was almost impossible for agents to tag the call correctly. In many cases, a caller was enquiring about multiple issues in a single call — one of which might be addressed while the other required a callback. So there was little insight available.
CSC addressed this with a data scientist-led customer intelligence project, which made sense of the notes and, for the first time, provided a genuine insight into the narrative of each call.
Download the paper to learn more.