Forget Big(ger) Data: It’s Time to Get Smart(er)


Share on LinkedIn


This is the first in a series on ways CX practitioners can move past focusing on the ‘bigness’ of data and get ahead with smart data techniques.

In my role at Convergys, I’m partially responsible for managing services delivery including selecting techniques and technologies for use in supporting the analytics practice. Although big data hit the crest of its marketing wave about a year ago, I still get emails every week about how to leverage it: Ways to store big data; ways to analyze it; emerging technologies that promise the world. Dealing effectively with big data has tremendous value for a huge array of business problems, but the hype has become just as huge. I’ve been asked by clients and prospects more than once, “Can’t you just use big data to analyze the voice of the customer and tell me what to do? I don’t care what you actually analyze … just make it big data.” Well, absolutely we will … if that’s the right approach to the problem.

So this got me thinking. We’re collecting, generating and disseminating more information than at any other point in history. Data are getting bigger every day with increases in social traffic, the growth of the Internet of Things (IoT) and a collective “data everywhere” mentality. There’s more data than most of us know how to use and how to analyze. So how do we turn surveys and other information about customers into better experiences for them?

In my opinion, the field of customer experience (CX) doesn’t require big data to improve interactions. We don’t need dozens of servers running in parallel analyzing real-time geo-location data from a mobile app to understand what customers want. Instead, we need to use data we already have readily at hand or that we can easily acquire. Being smarter about data that currently exists in your company enables a targeted focus on smart data and allows you to be more selective in informing CX investments.

“Listening” Today

Too many times, companies invest in collecting hundreds of thousands of voice of the customer (VOC) surveys yet are left wondering, “Now what?” To answer that question, we’ve used three strategies to augment VOC data and enable smarter business actions. These strategies vary from fairly simple to highly complex and include:

  1. Add a new flavor of qualitative research into existing VOC programs to get more from them;
  2. Use unstructured data from across the company to get at why customers say what they do on surveys; and
  3. Build a linkage ecosystem that shows how upstream process ties to customer attitudes and how attitudes affect downstream behaviors and financial outcomes.

Most CX researchers agree that to truly hear what customers are saying, you need a 360-degree measurement platform. Interestingly, this is a marked departure from just 10-15 years ago. If you look back at the history of VOC research, satisfaction and loyalty measurement programs used to be episodic. Companies would run an annual or semi-annual program touching on multiple areas of the business and would go from there. But the industry quickly learned this approach can fail to generate enough customer feedback to know how you’re doing in each key touchpoint and service experience along the way.

That’s where the evolution of listening in a 360-degree fashion came from. Today’s best-in-class measurement programs measure VOC across the lifecycle, across service interactions, across contact channels and at key points in the journey. That’s a lot of surveys. And those surveys typically get integrated into a comprehensive customer view, reporting and analysis platform and closed-loop service recovery system – all to reach individual customers, change their perceptions and improve the business. At least that’s the concept and where most aspire to take their CX programs.

So what if you embark on that journey – you enable 360-degree measurement, you write the surveys, you launch a bunch of programs, you even build the framework and start reporting – and nothing changes? There are a lot of VOC programs that never reach their full potential. Instead, they end up with a nice set of dashboards and push reports that are exciting at first but quickly lose impact over time. When programs overly focus on measurement at the expense of action, senior leadership starts getting frustrated with the customer experience office. It’s never enough just to know your score and how it is distributed across lines of business or management teams; you have to know how to change.

To get to root cause, you need more data than exists in the answers to a survey. Let’s talk about three ways to do just that.

Tip #1: Less What, More Why

The first way to better inform survey results is to inject qualitative approaches into ongoing measurement. Surveys give just a taste of why a customer feels the way s/he does but often lack deep insight into root causes. For example, let’s say you look at your scores and one product is consistently rated a 5 on quality. That’s good, right? Not necessarily. It’s good on a 5-point scale; it’s not good on a 10-point scale. And what does a 5 really tell you about what to change (or not) about the product in question? Until you have more information, you’re pretty much flying blind. Even if you have a verbatim, you may only get a couple of lines of feedback – which isn’t much to go on.

The smart data approach to getting deeper is by talking to customers in-depth about a particular issue. Historically, this has been done with focus groups or individual interviews. But focus groups take time and travel and in most cases you’re not talking to the same person that provided the original scores. To answer this need, we developed a capability that leverages chat-based qualitative research while the customer is in the act of completing a survey. Invitations to a second part of the survey are triggered by any number of criteria – answers to a question, demographic profile, customer segment, anything important to the business at the time. After they accept an invite, customers start chatting with a trained moderator in a new window. These guided sessions last about 10 minutes and are focused on pre-defined topics with a goal of explicitly getting at the root cause behind the numbers.

One big benefit is that standing up these chat sessions is quick and easy; they only run for a couple of days. And since this is qualitative guidance not quantification, you don’t need more than 20-40 chats. Plus the approach is nimble; it’s easy to quickly ramp up and down.

What Happens When Behaviors and Attitudes Don’t Match?

We work with a large credit card issuer that noticed a counter-intuitive pattern in their VOC program. Results consistently showed customers saying they’d use their card more in the future – even though the same customers rated the card poorly across all other loyalty metrics. When we applied statistical modeling to the survey data, this group stayed an outlier; what the model predicted they would do based on attitudes wasn’t reflective of behavioral intentions. So what was going on?

We used ConverSat® (our name for this smart data qualitative approach) to conduct 40 online chat sessions exploring the drivers of cardholder’s specific future usage. When we analyzed the data, some interesting patterns emerged. For this particular group, it was all about behavioral economics; habits and rational motives drove usage patterns, not card satisfaction and emotional loyalty. In fact, these cardholders compartmentalized cards for use only in specific purchases. They also were extremely rewards savvy and allocated usage based on rewards and APR, which kept them using cards they didn’t necessarily like even when they had alternatives.

Armed with these insights, the card issuer got smarter about treating this group differently and deployed new early strategies to challenge spend patterns by:

  • Providing marketing materials to highlight new use cases,
  • Offering rewards in frequently used categories and
  • Creating incremental rewards in non-typical spend patterns.

Interventions to reach these customers and drive higher share of wallet were designed to appeal to a behavioral economic strategy rather than emotional loyalty. If they’d relied only on survey insights, they would never have known how to meet the needs of this customer sub-group.

Next Up

In the next two blogs in this series, I’ll expand the smart data concept to include ways of leveraging unstructured data and predictive modeling in ongoing CX strategy. I also recently conducted a webinar called, “Predictive Analytics: Evolving from Big Data to Smart Data,” so if you’d like to watch the recording, click here.

Read the rest of our Smart Data blog series:

Part II: We *heart* Unstructured Data.
Part III: Smart Data: Integration to Action

Image source: Thinkstock

John Georgesen, Ph.D.
John Georgesen, Ph.D., is Senior Director, Analytics at Concentrix. He specializes in designing customer experience (CX) programs that drive tangible improvements. With 20 years of applied experience, John is a recognized innovator in the field of customer experience management.