I was asked to give a two-hour lecture at the University of Washington to an executive MBA class on the topic of measurement and analytics in customer experience. I covered such topics as Big Data, data science and how they can help you improve customer understanding. I’d like to thank the instructor, Sasha Frljanic, for letting me share my work with an exceptional group students (they asked some really good questions!). In today’s post, I’d like to share my slides and provide a summary of my talk, highlighting content that I think is particularly important to customer experience and customer success professionals.
Customer Experience (CX) in a Big Data World
There has been an explosive growth in interest in Big Data and all things related. While customer experience (CX) pales in comparison to data-related topics (slide 7), customer pros could improve the profile of CX and Customer Success in their companies by incorporating big data principles (e.g., data science, machine learning) into their customer programs.
We live in a world now in which everything about our customers can be quantified. This quantification of everything that is coupled with processing power of computers gives us the opportunity to learn about our customers through statistical modeling. Companies who can get insights from their data will necessarily win. Integrating disparate data silos helps companies obtain a single view of the customers. Your goal is to know everything about each of your customers. When you do, your analytics will result in better predictive models for all customers and will lead to true CX personalization.
Data science is way to extract insights from your data. To extract this information, business leaders need to ensure they have employees with the requisite data science skills. These skills fall into three broad areas: 1) subject matter expertise, 2) technology and 3) statistics. Putting the science into data science, data practitioners employ the scientific method to help focus their efforts to answer specific questions (slide 16).
Optimal Customer Survey
The optimal customer relationship survey helps companies collect the right attitudinal data to help them make strategic decisions. This survey is based on decades of research and practical experience; the survey not only minimizes the number of questions, it also optimizes the insights companies are able to get from their customers. This annual customer survey contains, at most, 20 questions that fall into a few categories, including:
- Customer Loyalty – questions that ask about the likelihood of future behaviors
- Customer Experience – questions that ask about the satisfaction with general CX touch points
- Relative Performance – questions that ask about your ranking within your industry
Analytics of Survey Data
Getting value from the customer relationship survey data requires three types of analytics (slide 38):
Descriptive analytics tells you what happened in the paset. Predictive analytics tells you what will happen. Finally, prescriptive analytics tells you what actions you need to take to maximize opportunities/mitigate risk.
The combination of these three types of analytics results in the Loyalty Driver Matrix (slide 44) that helps executive leaders make better decisions about how best to improve customer loyalty that drive business growth. I illustrate the use of the Loyalty Driver Matrix and also provide exercises (with answers) where you can apply your knowledge of these types of analytics to the results of a survey (slide 47).
While the short survey (above) has been shown to provide reliable, valid and useful information about your customers, I’m of the mind today that even fewer questions are needed in a customer survey. Why? We have so much data about our customers today, we simply don’t need to include as many questions in our survey to understand their needs. The only two questions you need are:
- What one word best describes this company/product/service?
- If you were in charge of this company, what improvements, if any would you make?
Even though these two questions do not require a numerical rating by the customers, we can still calculate a customer sentiment score from the first question. The score is generated by an algorithm that assigns a sentiment value to the one-word response (slide 56). The responses to the second question, in conjunction with the sentiment value we calculated using the first question, helps us identify business areas that need improvement.
When it comes to the application of measurement principles and analytics approaches, business leaders need to collect the right data and then apply the right types of analytics to extract the insights they need to move their business forward. We live in a highly quantified world in which customer insights lie at the intersection of different variables across multiple data silos. From data that are intentionally collected (e.g., surveys) to data that are simply available to us (e.g., social media), we can learn a lot about our customers through the appropriate use of analytics. Moving your company (as well as your career) forward starts with an understanding and appreciation of data science skills and practices.