Skip to main content

What to Expect When Optimizing For LTV

  • Data and Analytics Committee
  • Jul 3, 2020
  • 5 minute read

To learn more about using predictive modeling and data to optimize for lifetime value, contact the data scientists at Alliant!

 The Data and Analytics Committee would like to share this article from Alliant with SUBTA members. No doubt these are challenging times for maximizing return from your existing customers, but getting to understand your new customers and how differently they will engage with your brand is also critical to ensure profitability for the long term.  

Lifetime value is an important but sometimes mysterious metric that many marketers strive to measure and meet. We’ve compiled this short “bootcamp” to help you get started on your own journey towards optimizing for LTV.

Managing marketing operations to focus on acquiring customers with optimal LTV takes patience, persistence, and dedication. It impacts strategy in multiple areas, including creative, media and especially, data and analytics. Along the way, Alliant’s experience has resulted in a few basic tenets you can use to focus your strategy and ultimately build a high-LTV customer base.

LTV Basic #1

Understand your costs, your profit, and your goals.

Lifetime value is the sum of many behaviors observed throughout a customer’s lifecycle. The first step is to define what LTV means to you — and what metrics you will use to measure it.

Begin by reviewing how much customers cost to acquire. Some channels are more expensive than others, and some customers take multiple efforts. Here, and with all customer lifecycle stages, time is money, so assign a meaningful timeline to your definition of LTV.

After acquisition, look at indicators like a number of purchases, shipments, product mix, or total dollar amount spent – all are worthwhile measures of value. Quantify the costs of fulfillment, product, returns, and other common customer behaviors— each is tied to a myriad of costs, and to truly define what lifetime value is to your business, you must be aware of all the little things that add or subtract from value.

All of this information will let you calculate and understand your historical LTV based on actual data from the criteria above.

LTV Basic #2

Make sure you have the right data to manage LTV.

Once you have identified the relevant costs, they need to be correlated to time-stamped consumer behaviors captured in your CRM. These performance-driving behaviors include: orders, payments, selection of deferred payment options, returns, canceled orders, and declined credit cards. If you’re running a subscription business, be sure to flag other relevant behaviors, such as pauses and re-subscribes, in addition to cancels.

Clean data is the foundation of effective LTV management. Capturing these metrics is imperative to helping you understand the lifecycle performance of each customer. Monitoring performance can help you identify and respond to changes in customer behavior over time. You will easily be able to identify the “intro-only” customers who cancel after a discounted introduction offer compared to those who are repeat or reactivated customers.

Capturing data at this level will become even more valuable when you begin working a data scientist. He or she will be able to help you identify one or more dependent variables that define lifetime value. For example, your historical LTV may have identified one simple behavior as the key indicator for lifetime value, for instance, number of shipments. Therefore, you’ll need to collect and aggregate shipment data in order to know when a customer has hit that mark.

LTV Basic #3

Build a solid behavioral model.

The final step to constructing your LTV toolkit is to create a behavioral model that will drive your marketing and customer management decisions in the future. Models let you project the future value of new prospects or customers over time. Armed with this insight you can make better decisions about when to invest marketing in a given segment of consumers, or to pass them by for those that will be more profitable.

It is beyond the scope of this post to dive deep into the available model types, but given the right data you should have many options. Options include simple response models or a more complex multi-behavioral model that can optimize for consumers likely to exhibit several behaviors, like three shipments and payments. This type of model can also take into account other behaviors such as pause or returns. Your data scientist can help you consider your modeling strategy, the channels you should be modeling, and how to apply a model at the promotion stage or on receiving an order. 

LTV Basic #4

Play the long game and expect other metrics to take a short-term hit.

Once you build a model solution, you should test a small audience and be prepared to invest the appropriate timeline you established in your historical LTV to prove the predictive accuracy of the model. If you identified that LTV takes eight months with four shipments, grab some popcorn and watch the test segment with rapt attention. As you observe this group, you may notice that other metrics, perhaps your former key drivers of success, may suffer. For instance, response rates may be lower since a model tuned for LTV may exclude the money losing “intro-only” consumers.

If short-term metrics begin to decline, a natural reaction is to make decisions that interfere with the test. Stick to your guns and let the campaign run its course, as abandoning it could undermine your analytic investment. Remember, the model was designed to avoid responsive, but low-spending buyers. Remain patient and lean-in on the upfront work that was done to manage expectations of internal stakeholders.

Optimizing for lifetime value is truly a milestone achievement for a marketer, requiring preparation, a workshop of experts, and time. Set expectations and goals within your organization early, invest in establishing what LTV means to you, gather the right data, commit to testing a predictive model, and develop your strategy around the facts.

To learn more about using predictive modeling and data to optimize for lifetime value, contact the data scientists at Alliant!

Learn more and apply for SUBTA’s Committees

SUBTA color logo