Top

5 Steps to Better Customer Retention Analytics 

Bob Hayes, PhD | Jan 31, 2017 486 views No Comments

Share on LinkedIn

This blog describes a five-step process to improve your customer retention analytics efforts. This process leverages existing data to draw useful insights to show you how you can improve retention rates, increase purchasing behavior and maximize the lifetime value of your customers.

Customer retention is a key driver of business success. One approach to keep your customers from leaving is to optimize customer retention. Customer retention is defined as the activities a business participates in to create repeat customers and boost the profitability of future purchases.


Engage with customers in real-time across every channel, no matter the medium. Use visitor tracking and email analytics to know what your customers are seeing.

There are four reasons why you should you be concerned about customer retention. First, existing customers generate more revenue than new customers. Gartner estimates that, for businesses, around 80% of future revenue comes from just 20% of existing customers. Second, existing customers require less resources to keep compared to acquiring new customers. Third, existing customers cost less to service compared to new customers. Finally, existing customers act as your marketing team and bring in new customers.

The Goals of Retention Marketing

There are several goals or outcomes we want to optimize as marketers. At a minimum, the goal of retention marketing is to keep/retain your current customers. While keep customers around for a long time is important, other goals of retention marketing are focused on growing and expanding those relationships as measured through up/cross-selling, conversion rates and number of recommendations.

Our goals as marketers, then, require a close look at a variety of outcome measures. While this paper will present an example for one metric, it is important to state that the 5-step process can be applied to any of your marketing metrics.

Analyzing Retention Marketing Data

In our digitized world, businesses track a lot of data about their customers, including their attitudes, product usage and loyalty behaviors (e.g., staying, recommending, buying more). When it comes to retention marketing efforts, businesses are trying to leverage all these different types of data to draw insights about their customers. As such, retention marketing is necessarily a quantitative endeavor, requiring businesses to consolidate and analyze those data. While there are specific techniques businesses can apply to their data, they are all built on a more general approach to extracting insights from data. Below are five simple steps businesses follow when they use analytics to improve their business processes, including their marketing retention efforts.

  1. Formulate a question 1. or problem statement.
  2. Generate hypotheses/hunches/educated guesses that are testable.
  3. Gather/Generate data to understand the phenomenon in question. Data can be generated through experimentation; when we can’t conduct true experiments, data are obtained through observations.
  4. Analyze data to test the hypotheses/hunches / Draw conclusions.
  5. Communicate results to interested parties or take action (e.g., change processes) based on the conclusions. Additionally, the outcomes can help you refine your hypotheses for further testing.

The steps above might look familiar as they represent the scientific method. The scientific method is a proven method of finding insights using data. The application of the scientific method helps people be honest with themselves and minimizes the chances of them arriving at the wrong conclusion. By following where the data lead them, business leaders use their data to help them understand how their business really works. Through trial and error, marketers use the scientific method to uncover the reasons why customers stay or leave and the underlying processes that drive those relationships.

Let’s go through each of these steps.

1. Formulate the Questions

Before you analyze your data, you need to know what you want out of it. What do you want to know? What problem are you trying to solve? If you know that your churn is high, you might consider formulating questions that help you better understand why customers leave. Keep your business objectives in mind when formulating your questions. By posing specific, business-related questions, the answers you receive will help you improve how you run your business. For retention marketers, the questions they formulate are necessarily related to identifying the variables that are related to such things like churn rates and conversion rates.

It’s imperative that problem statements are either generated or reviewed by an executive who possesses good business and marketing acumen. Expertise in business and marketing will help ensure the questions that are being asked are relevant to business needs. Establishing the right problem statements will help guide decisions throughout each of the subsequent steps and increase the chances of successfully using data to improve retention marketing efforts.

While marketers might formulate their own questions or problem statements that are unique to their business situation, there are marketing questions that are common across all marketing professionals. I will focus on one of those questions:

  • What can I do to increase conversions?

2. Generate Hypotheses/Hunches

After establishing the problem statement, the next step is to try to generate hypotheses or hunches or educated guesses. A hypothesis essentially represents what is expected to happen. Hypotheses can be stated in either predictive terms, like “If I do A, then B will occur” or as a statement, like “A is related to B.”

Based on the problem statement, here is a testable hypothesis:

  • H1: Emails with a red call-to-action (CTA) button (compared to a blue button) result in higher conversion rates.

Finding the answers to this hypothesis will help the marketer build marketing campaigns that will improve conversion rates.

3. Gather/Generate Data

In this step, data are collected to determine if hypothesis is correct. Marketers can employ one of two methods to collect their data: experimental or observational. In the experimental approach, marketers control the levels of one variable to study its impact on the outcome of interest (e.g., conversion, renewal). In the observational approach, marketers do not control the levels of any variables; they simply use the data that are available to them to test their hypotheses.

Using our current example, the marketer devised a study to see how well her predictions held up. The marketer used the experimental approach by creating two marketing emails that only differed with respect to the color of the call-to-action (CTA) button (red vs blue) and tracked conversion rates for each one.

4. Analyze Data / Draw Conclusions

After marketers collect or gather their data, they analyze it to test their hypotheses. Retention marketers analyze the data using different types of analytics, including descriptive analytics and predictive analytics. Descriptive analytics are used to indicate the current state of the world. Predictive analytics are used to make predictions about future events or new metrics (other data). Predictive analytics essentially looks at the relationships among different metrics.

Descriptive analytics are commonly used to help marketers summarize key metrics, like current retention rates as well as rank product features on frequency of use. While useful, descriptive analytics do not help find answers to our hypothesis. Marketers rely on predictive analytics to provide those answers. For our examples, the marketer would use predictive analytics to calculate the degree of impact that a red CTA button has on conversion rates (e.g., using t-tests) and how strongly/weakly each of the product usage metrics is related to customer retention/renewal (e.g., using regression analysis).

When Data Become Overwhelming: Turn to Machine Learning

When marketers have integrated all of the disparate data silos, they are typically overwhelmed with the number of variables they have to manage and analyze. Even if they have access to a seasoned data scientist, the data scientist would still be hard pressed to manually and quickly sift through the sheer volume of data to find the optimal predictive model. Instead, to create the best predictive models of retention, marketers rely on the power of machine learning to quickly and accurately uncover the underlying reasons why customers are staying or leaving.

Based on math, statistics and probability, machine learning finds connections among variables that help optimize important organizational outcomes, in this case, retention. Machine learning simply uses historical data to build predictive models of customer retention. These models are then applied to new customer data to make predictions about the future. Iterative in nature, machine learning algorithms continually learn, and the more data they ingest, the better they get. Coupled with the processing capability of today, these algorithms, compared to humans, can provide customer insights quickly.

5. Take action / Communicate results

In this step, marketers can use the study findings in two ways. First, they can communicate the results to others in the organization. Second, they can take action on their own by making changes in the marketing processes to improve outcomes (retention, conversion, renewal). Prescriptive analytics is employed in this phase of the retention marketing analytics. Prescriptive analytics involves the combination of both quantitative results and human judgment.

Our marketer found that the red CTA buttons resulted in higher conversion rates (compared to blue); as a result of this finding, she took action by changing all of the CTA buttons to red. Additional analyses can be conducted to determine the lift in retention that occurred due to the color change.

It is important to note that the results of the analyses don’t have to end at this stage. Marketers often use the results of their analyses to set new hypotheses. They might discover new pieces of information that lead to additional questions. Basically, the results of one study act as the background research to inform new/future retention marketing projects. The results of your analysis lead to better questions which lead to better answers. Retention marketing analytics efforts are not a one-time endeavor. It’s best to think of retention marketing analytics as a way to continually improve retention efforts, either through implementing new marketing procedures or fine-tuning existing ones.

Summary

To be successful at retention marketing means that marketers need to leverage their company’s data. Marketers who can apply appropriate analytics to these data will get the insights they need to drive business growth. Figure 1 summarizes the steps that guide marketers’ analytics approach. The key to successful retention marketing analytics rests on asking the right questions, getting access to the right data and analyzing the data. As a result, it’s important that retention marketing analytics involves experts who are knowledgeable in marketing (to ask the right questions), engineering (to get at the data) and statistics (to analyze the data).

GettingValuefromCustomerData.pngFigure 1. The Empirical Enterprise: 5 Steps to Get Value from Data

While companies can rely on their employees to conduct retention marketing analytics efforts, they often don’t have employees with sophisticated analytics (read: data science) skills. Consequently, companies are relying more on third party vendors who provide a customer data platform that automates data integration and analyses, leaving the marketers more time to think about the questions to ask and how to implement what they learn. At Appuri, we developed a data and analytics platform that simplifies the way businesses are getting insight from their data (see Figure 2).

DataToInsightsAction.pngFigure 2. The Empirical Enterprise uses data to make better decisions.

We are able to easily integrate disparate data silos and build machine learning algorithms that automatically surface insights, leaving marketing professionals the time they need to take action on those insights to improve their business processes and take immediate action to save accounts if necessary. By automating many of the analytics steps to get from data to insights/action, we help business professionals focus on activities that will decrease churn, grow existing relationships and improve advocacy.

Print Friendly

Republished with author's permission from original post.


CustomerThink Recommends:


Categories: BlogCustomer AnalyticsCustomer LoyaltyEditor's Pick

486 views

No responses yet, why not leave yours?

Add Your Comment (All comments are reviewed by moderator, no spam permitted!)