Putting Lifetime Value to Work

Hi, I’m Elia, Business Intelligence Manager at Buzzvil, a mobile lockscreen ad network. We use lifetime value, or LTV, in many areas, including user acquisition channel management. In this post, I’ll share the trial and error behind our LTV calculations and some practical ways to use them.

Is marketing an expense—or an investment?

The mobile advertising industry has a few universally important metrics. New registered users (NRU), daily active users (DAU), and retention rate are essential for anyone in mobile advertising, especially marketers. Managing new users arriving through different channels and observing their retention determines the business’s DAU. A marketer should have a reasonable understanding of the costs involved in acquiring and retaining those users. Doing that well makes someone a capable marketer, but not necessarily an exceptional one. From the business’s perspective, a more important question remains.

One interesting paper used text mining to identify the most frequent keywords in several decades of articles from the Journal of Marketing Research. The fourth most common keyword was “measurement.” That reflects how much attention researchers have paid to a marketer’s enduring dilemma: the CEO asking, “So, does it make money?” A Facebook post might earn 100,000 shares, or a Super Bowl ad might fill Twitter with the company’s name, but the underlying question remains the same.

The key is to recognize that, from the company’s perspective, marketing is an investment rather than simply an expense. Every investment is evaluated for suitability and performance through ROI, and marketing should be too.

Calculating LTV: the first step toward ROI analysis

LTV is the metric that lets us analyze the results of marketing—new users and retention—in terms of ROI. Its importance has long been recognized, and examples of companies managing it are easy to find. A coffee shop’s “buy ten, get one free” program is a classic example. Why offer something that initially appears to reduce profit? The immediate answer is retention; the deeper answer is LTV.

Stamping a loyalty card raises the likelihood of the customer’s next purchase. That probability of returning and buying again is a form of retention. But returning to the store or service is not enough by itself. The more fundamental business question is, “Does this customer’s return help our business?” There may be times when revenue growth takes priority over profit, but profit remains a business foundation. If selling a coffee costs more than it earns, repeat visits only increase the loss. Retention alone then loses its meaning. That’s why I consider LTV one of the ultimate metrics a marketer should watch.

Learning through trial and error

Buzzvil began trying to calculate LTV early in its life. We took our first steps using Sungmoon Cho’s blog post, “Understanding Customer Lifetime Value”, which defines LTV as follows:

Unfortunately, our initial attempts using that definition did not produce particularly meaningful results. There were two main reasons:

1. We lacked sufficient data.
2. It was too early in the business lifecycle to calculate LTV reliably.

The formula requires reasonably accurate measures of average revenue per customer, retention, and acquisition cost. The complexity of mobile advertising made average revenue per user difficult to measure. Revenue can arise per impression, per click, or when a condition such as installation or app execution is met. Simply dividing total revenue by user count discarded too much information. Yet we lacked sufficiently detailed behavioral data to measure each component individually.

Once we recognized the data gap, we carefully reviewed what to collect, when and where to collect it, how to store it, and how it should be made available for analysis. We then put that plan into practice.

Metrics such as retention also need time to stabilize. Early in a business’s lifecycle, revenue, retention, and acquisition costs may all be unstable, making it difficult to estimate suitable values for the formula. The necessary time varies by business. In Buzzvil’s case, it took at least a year.

How Buzzvil calculates LTV

Over time, our metrics stabilized and data accumulated, making LTV calculation possible. We could calculate LTV by country or as an overall average, but we kept working toward the smallest practical unit and eventually reached individual-user LTV. Let’s look at the calculation and then an example of how we use it in marketing.

The formula can be expressed simply as:

That single line contains a great deal. The time variable t is closely related to retention. For a user i with low retention, the range of t may span only a day or two. With high retention, it may extend beyond one or two years. Assuming revenue exceeds cost at each t, improving retention is essential to increasing LTV.

It also matters that both revenue and cost vary with i and t. As noted above, measurement is a central concern in marketing. Without proper measurement, tracing the revenue and cost generated by user i at time t is difficult.

Consider a restaurant. Measuring a customer’s LTV requires a record of that person’s visits, which most restaurants don’t inherently have. Loyalty programs help create that history. If we know when a customer visited, what they ordered, what they paid, and the margin on that food at the time, we can calculate LTV.

Online advertising offers vast amounts of data, allowing even small user actions to be tracked and making this approach possible. Depending on the service, t can be measured in finer units such as hours or in larger intervals.

Using LTV to manage acquisition channels

Once LTV is calculated, we need to decide how to apply it. The metric contains a wealth of information and can inform many decisions. We might identify when LTV drops most sharply over time and use marketing events to improve returns, or examine each user’s revenue minus cost and improve profitability through segmentation.

At Buzzvil, user acquisition was one area where LTV substantially improved our economics. Influencer marketing, common in markets such as gaming apps, provides a useful example. Different channels work with different influencers, who promote apps and services in different ways. The profitability of users acquired through those channels can therefore vary considerably.

Imagine one influencer saying, “Just install it—you can delete it immediately,” while another says, “This is a great app; install it and give it a try.” Their acquired users will likely have different retention rates. If acquisition through the first channel is significantly cheaper, it could still be more profitable despite lower retention. If costs are similar, retention may become a decisive factor.

But retention isn’t the only factor: user activity also affects revenue considerably. Judging a channel solely on retention is risky. LTV incorporates both retention and activity, producing a metric that can be compared with acquisition cost to make profitability much easier to assess.

Closing thoughts

Working with mobile advertising data reveals how much profitability can vary with user behavior. Low-activity and high-activity users differ across many measures, from impressions to click-through rates. Calculating LTV helped Buzzvil build a DAU base with strong underlying engagement. We are now working to connect those users with more suitable ads through improvements such as more sophisticated targeting.

I hope this post helps you become a more effective marketer and helps your company improve its profitability. :)