Data Quality Impacts Customer Churn

Five Ways Your Data Could Be Hurting Customer Retention

According to a now classic article in the Harvard Business Review, depending on the study and what industry you’re in, acquiring a new customer is anywhere from five to 25 times more expensive than retaining an existing one. Further, the authors point to research that shows increasing customer retention rates by just 5% increases profits by 25%-95%!

Holding on to the right customers is certainly a smart move. That’s why so many companies invest in customer retention.

Managing churn is imperative for growth

Churn is when your customers leave you, usually for a competitor. Managing churn is incredibly important, first from a revenue standpoint. Losing a customer means you’ve also lost the revenue from that customer. Second, depending on their reasons for leaving, departing customers can cause damage to your company’s reputation.

Often, the reason that customer churn occurs is that companies do a poor job predicting customers’ needs and designing marketing and sales initiatives to present potential solutions well ahead of renewal time. There are five overarching reasons that may be why these predictions miss the mark:

  1. Wrong types of data. In order to forecast what your customers will need three months, six months, or 12 months from now, you need to be capturing and collecting the right kinds of data. For example: What are they buying? How are they using it? How often? What services are they paying for that they’re not using as much as expected? What industries are they in? Who are their customers? How many employees do they have, and how many of those employees use your products or services?

    The specific inputs you need to capture will vary, but in the end, you need data that is going to help you understand how your customer interacts with your product and company, and that has the potential to really change the way you approach your customers.

  2. "Is “Bob’s Drilling” a dental practice or a contractor in the oil industry? The answer will certainly impact how you engage with the company."
  3. Not enough data. Having good data on just a small percentage of your customers is a step in the right direction, but it’s not enough to make really good forecasts. To do that, you need to have a sufficient volume of data to be able to create control and test groups for your forecasts, to evaluate how your models fit against historical trends, and to be able to identify statistically significant differences between different cohorts of customers. Without enough data, you increase your risk of making decisions based on a sample of data that isn’t representative of the whole.
  4. Incorrect data. You might have identified the right metrics you need to collect and may have data on every single row of your customer tables, but if the data about those customers isn’t correct, all bets are off. If you collect your data in-house, it’s important to regularly evaluate the accuracy of the data your team is collecting, through independent verification or automated processes that check for typical data issues. If you’re gathering your data from a third party, you need to make sure they have those checks in place. Is “Bob’s Drilling” a dental practice or a contractor in the oil industry? The answer will certainly impact how you engage with the company.
  5. Old data. Data goes stale over time. In many ways, companies are living, animated entities that change daily. There’s new leadership, new employees, new locations, new industries. If you think that a customer has 40 employees but it’s recently gone on a hiring spree and now numbers more than 300, you’ll be at a disadvantage. Your sales and marketing efforts are going to focus on the wrong things, and your customer may see you as out of step with its trajectory. Keeping your data up to date will eliminate that issue.

    Of course, it’s easy to get overzealous with this. While companies do change a lot, not every company will undergo a meaningful change every day. If you just updated a CEO’s name two days ago it’s probably still good. Two years ago, though? You may want to confirm it. Find a cadence that works for your needs and try to make sure you’re updating data to fit that rhythm.

  6. Bad or outdated models. Models are tough: tough to build, tough to test, tough to deploy. It takes a lot of effort to keep predictive models accurate enough for meaningful forecasting. Even if the data that feeds a model is pristine, it takes a special skill set to properly develop and interpret predictive models. Make sure you have that skill set available, whether you’re the one developing or the one interpreting your models.

And remember: a predictive model isn’t a static tool; it can and must adapt to new information, as well as to market threats. Your competitors aren’t going to just rest on their laurels, so your models need to change to stay ahead of the others’ ability to make good predictions.

Addressing these five areas can help you develop an intrinsic strategy for holding on to existing customers, relieving the strain on your bottom line, and building long-term relationships.