Calculated Loyalty: How to Leverage Data to Combat Customer Churn
Every savvy business owner knows that keeping an existing customer is more profitable than bringing on a new one. But how many actually devise successful plans to encourage repeat business?
While magazines, newspapers and streaming websites rely on subscribers to drive revenue, the importance of subscribers isn’t limited to media companies. Car-sharing companies, mobile phone providers and a number of other industries can also carve out a significant revenue stream through subscribers.
To reverse churn and improve subscriber rates, it’s important to understand what prompts customers to purchase and why they stop — especially if your company has the potential for repeat customer interactions, operates under an annuity-like structure or offers loyalty rewards. Armed with this knowledge, you can more effectively market your products and services to drive revenue.
Leveraging Rich Customer Data
In the context of subscription services, understanding why customers leave is called churn analysis. Although customer satisfaction surveys are helpful, companies often take a statistical approach known as survival analysis, which studies the events leading up to a subscriber’s exit. For example, how long did the individual remain a customer, and what factors influenced his or her decision?
Survival analysis is essentially a model of human behavior, so to paint a complete picture, you need to capture as many details as possible. You’ll have to gather information on each individual’s subscription activity — e.g., when a subscriber joined; the product (or products) included in the subscription; the length of the subscription; and any renewals, cancellations or churn events that occurred.
However, while this information will describe what happened, it rarely sufficiently explains why the subscriber joined, much less why the renewal, cancellation or churn actually occurred. To address these questions, you need more granular information.
Uncovering the Source of Customer Churn
Once you’ve accumulated rich customer data, it’s time to analyze the right details, identify the low-cost, high-benefit means to increasing subscription revenues, and alter your offering accordingly.
However, it’s important to know what to look for in the data and which aspects should be analyzed. You’ll need to account for customer-specific data, loyalty program data, transactional history, loyalty program history and peer effect data (e.g., social media data, customer surveys, hyperlocal sociodemographic data, referral patterns, neighborhood shopping patterns, etc.).
You can find most of this data through your company’s accounting system. However, when marketing and accounting functions are segregated, the information is often managed separately. You can also leverage customer service records to map out customers’ experiences and learn which touchpoints prompt customers to drop off or abandon your service altogether.
External sources can also provide useful insights, especially when conducting a competitive analysis. Look to local, regional and national economic trends to control for variations in employment, income or consumer satisfaction. An effective churn analysis will identify all sources of relevant information and combine them into one usable data set.
Once you’ve gathered enough data points and uncovered trends in customer churn, there are a few best practices for putting this information to work:
1. Prepare your data to be analyzed. How your data is collected, stored and structured matters. Gather the types of data necessary for churn, and zero in on those points.
2. Let the data speak. Data often counters instincts or experience. Similar to the phrase “that’s why they play the game,” be prepared for surprises.
3. Plan to connect insights to action. It’s not enough to know why you’re losing or gaining customers; you have to do something about it. Start by integrating the output of the predictive algorithm with your CRM or customer contact system. Then, consider how you connect with customers — emails, call centers, web pages, etc. — and set alert actions. This type of preventative marketing will entice customers to stay in the fold.
4. Experiment with promotions, and be ready to deliver. No one promotion will work for your entire customer base. Create a promotions experimentation plan, and sample a large enough size to know which retention programs work for each customer segment.
5. Constantly update your database. Reversing customer churn should be an ongoing process that accounts for market fluctuations. Continuously add data regarding your promotions to inform even more accurate churn and models.
6. Institute change from within. Internal communication and culture can influence customer perceptions and, therefore, churn. Make sure each department knows why your customers buy and continue to buy.
To get everyone accounting for churn, you must associate the database of customers you’ve tagged as likely to churn with the overall customer recognition database. This way, when a customer calls and identifies himself via an account number, the customer service representative knows the individual is likely to depart and should be treated differently. This includes revising call-center scripts, training employees to recognize customers likely to churn, and preparing a stable of promotions and offerings tailored to these customers.
Customer churn is a costly yet preventable occurrence. Equipped with the right data points, you can uncover what drives subscriber choices and optimize your business accordingly. By accurately assessing the relative benefits of each change and eliminating ineffective promotions and programs, you can prioritize lucrative changes and secure a more sustainable revenue stream.
John Kelly leads the predictive analytics practice at Berkeley Research Group, which works with marketing, sales and operations leadership across a range of industries to leverage the power of econometrics and data science.
