Customer lifetime value (LTV) is one of the most important concepts in marketing today—it’s also the single-most pressing challenge. Customer LTV is all the money the average customer will spend with your business during their lifetime. Knowing that figure is clearly important, because it not only measures performance, it also helps guide many of your marketing decisions.

However, with changing customer touchpoints and ways of measuring interactions, LTV isn’t such a simple concept. And, as much as marketers talk about the importance of multi-touch attribution, customer data, and cross-channel measurement, and behavioral forecasting, the harsh reality is many of us are floundering to wrap our heads around measuring customer value.

Even though these concepts are integral to the future of marketing, they’re tough to implement into processes. To use an analogy, making measurement and optimization a part of your Marketing day-to-day can be the digital equivalent of a root canal: it can be a source of anxiety and pain. It doesn’t have to be that way though.

In many ways, technology is changing how we are measuring value, across the business’ departments, but for marketers, LTV is an evolving metric right now. Consumer-facing startup and enterprise marketers are all building their own LTV pictures, and we’re all learning from each other. Here are five key lessons from two e-commerce marketing teams, Art.com and Diamond Candles.

Step #1: Get started with an integrated customer view

One of the biggest challenges at Art.com involves the sheer volume of marketing activity. The company has created large scale digital marketing systems — particularly around search engine marketing (SEM).

Even more challenging is the fact that Art.com is an international company with a presence in 25 countries, websites in 17 different languages, and transactions in more than 12 currencies.

One of the most important decisions the company has made is to centralize information into a robust customer database that can be used to segment and analyze customer behavior to get a clearer view of LTV. Every piece of information also helps Art.com personalize campaigns which can bring more value to the equation.

“My recommendation would be for companies to take the first step to getting a 360-degree view of their customers and integrating all of the data across their marketing platforms into a single location,” says David Tjen, director of business analytics and strategy at Art.com. “This consolidation will help the marketing team have the proper inputs into calculating a lifetime value number.”

Step #2: Identify the business challenges you’re looking to solve

It’s important to have a clear understanding of why you’re collecting each and every data point. With the proliferation of information, marketers are often drowning in KPIs with a less-than-focused view of important business targets. Measurement is only part of the LTV equation.

“The biggest challenge for us is not the ability to measure lifetime value,” says Diamond Candles founder Justin Winter. “It’s how the changes to different customer touch points influence this metric. At Diamond Candles, the question that we’re asking is how to extend the value of each customer.”

At any given time, Winter relies on customer data to evaluate what his marketing team should do next.

“We’re wondering how to make big initiatives bigger,” says Winter. “Do we want to get people to buy more quickly, or invest more in nurturing our customer relationships?”

Goals are essential to the process of building action-oriented, accurate, and value-generating LTV processes. Some data will be worth more to your marketing efforts than others, and it’s important to know why.

Step #3: Create forecasts that make sense for your business

LTV forecasts are important because they allow marketers to take action, but those forecasts will be dependent on several considerations: purchases over time time, customer segments, and acquisition costs. If it makes sense, marketers may want to create business line or product-specific assumptions. No matter your approach, it’s important to keep assessments simple but flexible enough to accommodate a range of scenarios.

Art.com, for instance, uses an LTV model that forecasts new customer performance over a 3-year horizon. The company also utilizes its customer cluster models to better identify opportunities for personalization and targeting.

“We calculate the number of new customers acquired during a period and project a one-year, two-year and three-year repeat rate, average repeat orders per customer, average order size and then we estimate the marketing costs for retention and calculate a lifetime value for customers acquired through different channels,” says Tjen.

This analysis positions Art.com to make the most out of every customer segment. The company can more effectively personalize campaigns and determine whether to focus marketing priorities. This level of granularity helps Art.com elongate the potential value for every single customer type.

“For us, knowing that we can identify a cluster of customers that behave like our most loyal customers in their purchase patterns indicates to us that we should spend in order to retain these customers,” says Tjen. “Likewise, we can limit or eliminate spending on customers with very low potential lifetime value and optimize our marketing spend on the channels and customers with the highest returns for the business.”

Step #4: Identify specific optimizations

Forecasts should play a key role in a marketing team’s decision-making process. There are two strategic areas your data is likely to influence: spend and marketing activity.

With respect to spend, your marketing forecasts will outline what your team can afford in terms of customer acquisition.

“There are many businesses that focus on the initial transaction, only,” says Winter. “But there may be tough-to-win customer segments who require more investment upfront. In some situations, it may make sense to take a potential loss.”

LTV gives marketers much more flexibility and agility in the marketing spends. The end result includes higher quality customer acquisition and retention programs that optimize revenue over the long-term.

In terms of marketing activity, LTV forecasts should illuminate a brand’s highest value marketing opportunities.

“One thing we have identified through our search [engine] marketing is that some of the search terms we bid on can attract customers with very different lifetime values,” says Tjen. “For example, someone searching for a specific art print, from say, ‘Norman Rockwell,’ may convert more easily at Art.com, but tends to not be as valuable to the company as someone who converts on a search term like ‘decorate kitchen.’”

With this analysis, Art.com can respond to specific user behaviors.

Step #5: Build storytelling into your forecasting process

When it comes to enhancing LTV, your optimizations need to make sense. That’s why it’s important to build storytelling into your process. Marketers need to create a set of hypotheses and stories around key trends. This vantage point will ensure optimizations always remain human.

Take the Art.com example from the last section, as an example. Equally important to the trend is Tjen’s rationale for why different search terms yield varying LTVs.

“One theory is that a person who is looking for a particular print is being opportunistic and is only concerned in obtaining that single product,” says Tjen. “This person may be less likely to be a repeat customer than someone who has a functional problem they are trying to solve—like decorating a room, for instance.”

This approach to storytelling will position marketers to uncover opportunities they may not have considered previously.

“For instance, at Diamond Candles, we decided to test direct mail campaigns,” says Winter. “We realized that mail was an untapped marketing channel for high-value customers.”

Final thoughts

LTV optimizations are always a work in progress—no company will have a static process. As we collect more data, we can validate and build upon our models. What’s important is that we refine our methodologies, keep learning, and keep sharing our milestones with the marketing community. Precise LTV measurements make us better marketers and more importantly, make our ventures more profitable.