Small business owners know that profits come from customers. They also know how much work it takes to acquire and keep a customer. The old adage may be “the customer is always right,” but the customer doesn’t always boost your bottom line significantly. For small businesses in particular, knowing which customers are valuable enough that they should be retained is vital. This requires knowing not only what the customer was worth in the past, but also what his or her value will be in the future. One valuable tool for this is customer lifetime value (CLV), which is the present value of future cash flow from a single customer or a group of customers.

Small business owners often don’t think about CLV, however. Some rely on a gut feeling about whether a customer will make a purchase. When asked how they know which potential customers will buy from them, a common reply is something like, “Well, actually, I don’t know. And last year quite a few prospects didn’t spend anything with us, so it was a bad year.”

CLV is becoming a vital predictive indicator. Many larger companies use it to generate valuable insights, and small businesses should do the same.


Many factors need to be considered when calculating CLV, such as the changes in customer spending, the success of marketing campaigns, the cost of goods and services sold, and the time horizon of customer loyalty—and that’s only the beginning. Some considerations present different or unique challenges for small businesses.

Cash-flow estimates, for example, are crucial to a successful calculation but are dependent on the accuracy of the data. And data collection is an issue for many small businesses. Every customer-­related transaction must be assigned to a specific customer, but a common mistake small businesses make is having multiple customer accounts for one customer. Even misspelling a customer’s name can invalidate the data. Small business managers are often so frantic in the execution of daily operations that many fail to devote sufficient resources to implementing the systems and processes required to collect accurate, relevant data. When the time comes to analyze the data, sorting through the collection of inconsistent inputs seems overwhelming.

Furthermore, many details of customer-related transactions may take place outside the business’s accounting system. For instance, a customer may purchase an item that’s properly documented through the software program and assigned to the specific customer. If that item is returned or ex­changed, however, then the customer service representative handling the return may just swap the product for a new one without identifying the customer. Essentially, the transaction with the customer actually costs twice what the data reveals. From that point, the CLV is skewed and may cause the small business owner to make an unprofitable decision to work to keep that customer. Lackluster customer profitability may result from internal company weaknesses that can be identified through analysis of CLV.


There are three levels of CLV, each with different uses:

  • Individual customer-level CLV is best used for marketing campaigns.
  • The CLV of specific segmented groups is used to analyze factors that drive demand for the product or services.
  • Customer-base-level CLV, which spans the entire population of customers, is used primarily for overall business strategy, planning, budgeting, and allocation of resources.

CLV models rely on comprehensive revenue, product-cost, and ­marketing-cost data at the individual transaction level. Using the available profitability data, a modified discounted cash-flow model or linear regression is adequate for most small businesses. Linear regression is used to compute the correlation between a continuous dependent variable (y) and one or more continuous independent variables (x1, x2…).

To build a CLV model (where y = CLV), the independent variables should include the top predictors of CLV, such as the customer’s historical annual revenue to a specific date (x1), annual income (x2), and so on.

Even with a simple CLV model, it’s important to consider specific attributes of the data. Seasonal purchase behavior can be unpredictable, so you need at least two years of data inputs for seasonal-based CLV. And more recent and more frequent purchases should have greater predictive qualities. For example, analysts usually exclude customers who haven’t purchased anything within a year and use the weekly average purchase amount per customer.

After listing variables that may affect CLV, use a decision tree or linear regression model to assess correlation among a set of variables to see which factors impact CLV and to determine their effects. For example, ask yourself, “Does a customer’s buying frequency affect the total revenue or profits that he or she generates? Is there a correlation between customer income level or other size components and the total revenue and profit from the customer?” Brainstorm hypotheses to determine potential correlations between CLV and relevant variables. Score each variable based on the significance of its impact on the hypothesis.

Test the accuracy of the rules as predictors by applying each rule to a set of customer data with known outcomes. For instance, if your hypothesis is that customers who purchase during the busiest promotional seasons generate the largest revenue per transaction, then the associated rule could be that customers who make purchases during November and December spend more than twice the per-transaction average.

To test this rule, select the customer data for all the customers with transaction dates between November 1 and December 31. Compare the data of the promotional period transaction amounts to the average transaction amount to decide if the answer is yes. Is the answer yes in every case or most cases? Revise the rule to improve its accuracy. Test the data from other purchases to rule out the possibility that another group also consistently exceeds the per-transaction average.

Once you’ve identified the correlation between variables in customer buying behavior, such as order frequency and transaction size, update the CLV forecast with each change in a variable. For instance, if profitability has been associated with order size, track and flag changes in individual customer order volumes and amounts.

While CLV calculation is based on an uncertain set of assumptions rather than an exact science, the process of estimating the value of a customer or set of customers reveals extremely valuable information that can help increase the profitability of your small business. Here are some of the benefits a small business might gain from using CLV:

  1. Identifying customers who detract from the company’s overall mission and goals;
  2. Setting more realistic sales goals;
  3. Detecting unproductive spending;
  4. Making changes to a price, product, promotion, or other aspect of a marketing strategy to en­sure it reaches the business’s valuable customers; and
  5. Refining business processes to improve weaknesses in areas such as data collection.

The gains from business insights extend far beyond improving efficiency and enlarging market share. Ultimately, CLV helps a small business achieve its goal of building lasting relationships with its best customers.

About the Authors