by Hanna Rocks

Technology and data are revolutionizing CX. As opposed to simply measuring and responding to customers who have provided direct feedback, CX teams now can assess, track, and grow every customer account by utilizing machine learning and artificial intelligence to analyze the full breadth of the data their customers generate. 

This kind of data-driven business optimization is quickly becoming a question of survival and presents CX leaders with an extraordinary opportunity. 

Data mining for surprising customer insights isn’t new. Back in 2014, the dating site OK Cupid found that there was a high correlation between a couple’s long-term romantic success and their matching answers to three seemingly random questions: Do you like horror movies? Have you traveled alone in a foreign country for fun? Would you like to ditch it all and live on a sailboat?

In recent years, technology has matured into a powerful tool that can drive better business outcomes. A leading job website had long worked to increase the number of applicants every listing received, under the assumption that there was a direct correlation between more applications and greater customer satisfaction. Direct survey feedback appeared to confirm that belief. 

When CX executives looked at a wider set of data, however, it turned out that satisfaction (and retention) didn’t keep growing with additional applicants–in fact, it began to decline beyond about five. The CX team led a company-wide shift to focus on other metrics. It was a massive success: The money invested in engineering resources for this initiative returned 100x in revenue. 

You need to think beyond the survey—way beyond the survey. Your company is already collecting a universe of data, and the tools are now readily available for savvy CX leaders to turn that data into an improved customer experience as well as a stronger bottom line.

At commonFont we have created a framework to help teams at some of the largest companies in the country make this leap. Outlined below are the five steps to help your company understand, and better serve, your entire customer base.

1. Broaden Your Data Horizon

The first step is a data audit. What data does your company collect? Which department controls what? How can you access it? 

The question of access often creates a challenge–siloed departments, security protocols, and overstretched IT departments can make it hard to connect with what you need. If your company has centralized data in a warehouse, lake, or lakehouse, the best option is usually to access it within that secure environment. Data storage and management services like Snowflake, AWS or Databricks now offer powerful tools with high-speed data processing capabilities even for semi-structured or unstructured data, which now dwarf the volume of structured data companies collect. 

In-situ data analysis may require fresh skillsets on your CX team or partnerships with data stakeholders, but it is often the most robust solution. This is because it doesn’t require the creation or maintenance of data pipelines, there is no duplication of data, and data owners don’t need to continually track their potentially sensitive data in other environments.

2. Make Sense of Your Data

Your customer leaves a detailed digital data trail as they interact with your company across website, mobile, phone, and in-person visits. But what does that data mean for customer outcomes? The best approach is to start with a hypothesis and then let machine learning algorithms uncover the subtle and hidden correlations hidden deep in the data. For example, a recent client tracked product usage and service tickets to see how that correlated with contract renewals. While the initial hypothesis was that the increased service requests indicated an unhappy customer at risk of churn, the relationship turned out to be far more complex, including the timing and type of these requests, the type and frequency of product usage, and host of other considerations.

In this expanded data universe, the direct-feedback data that CX teams have traditionally collected becomes simultaneously less and more important. Less important because you have an ocean of new data to consider—but more important because this direct feedback can not only guide you to a good hypothesis but can then confirm or call into question what your data analysis reveals. Negative feedback about some of the suggested ideas to reduce churn in the previous example helped guide the client to ideas that worked better.

3. Choose Your Metric(s)

Every company has Key Performance Indicators (KPIs) that it tracks to assess customer health and loyalty. Customer spend can be a good proxy for satisfaction, though the relationship is rarely perfect. NPS is widely used as a broad measure, as is CSAT for specific interactions. A churn probability rating is particularly relevant and popular for subscription revenue companies. The metrics that you start with should be directly tied to the improved business results, so that you can evaluate actions relative to the impact they have on that metric.

4. Test, Iterate, and Follow the Data

With data access secured and metrics selected, it’s time to let artificial intelligence get to work. There are thousands of different AI tools that you could use, and it can take some time to see which one yields the most effective insights for your business. 

Most importantly, keep an open mind as you test both the tools you’re using and the data you are analyzing. Sometimes the correlations in the data are unexpectedly powerful and offer a fresh perspective about how your customers feel about your business. This is why it’s best to cast as wide a data net as possible and let your AI tool identify trends and patterns. 

For example, ServiceNow recently built a model that it found best predicted customer churn when it ingested 500 separate operational variables. The system began to correctly identify at-risk accounts and suggest ideas on how to retain them. Surprisingly, it found that accounts that had increasing sales were statistically as at-risk as ones that were stable or shrinking.

5. Assess Impact and Take Action

Your goal is to improve business outcomes. Once you have a data-driven thesis about what actions might do that, you can provide a personal phone call, a special offer, or an extra service to a subset of customers and then measure the result compared to a customer group that didn’t receive that treatment. It’s important to have a structured and statistically valid approach to work through these permutations, and our team here at commonFont developed a system that we will dive into in a future article.

The Future of CX

CX departments have the opportunity to drive success for their organizations in ways that they never could before. Your organization needs a new CX paradigm in order to successfully navigate the changing business world. The great news is that the tools and approach required are now available to you. Here at commonFont, we are helping companies make that transition. There has never been a better time to leverage all of your customer data to improve experiences and drive better business results.