The LTV metric is only partially useful to modern mobile marketers.
For many mobile advertisers, the LTV / CAC ratio sits at the very heart of commercial operations — LTV / CAC is the fundamental measurement of product viability, encapsulating both marketing efficiency and product monetization.
But the LTV metric is an anachronism on mobile. For one, no one thinks about monetization on the basis of some abstract “lifetime”: in this second act of the mobile economy, users are sticky, products are enduring, and monetization streams are complex. The “LTV paradigm” really represents the first act of the mobile economy, when products could launch and gain traction with relatively modest budgets and users casually, frictionlessly flitted between them as fungible commodities.
This isn’t what the world looks like anymore: there are fewer winners in the second act of the mobile economy, but they’re gigantic, and they enjoy some level of customer lock-in — either through network moats, brand loyalty, product differentiation based on real technological innovation, or a combination of all three. On this basis, there is no “user lifetime,” at least not in a way that can practically be applied to marketing operations. If a user can remain with an app for three or five or seven years, what’s the point of calculating some hypothetical “value” of that lifetime? An estimated LTV over the course of years would be unreliable, fragile, susceptible to massive swings as the product changed, and most importantly, too expensive to buy traffic against.
Rather, marketers in this second act should approach growth from the opposite perspective: what is the optimal way to structure cash flows to support the growth of the business? In “How does LTV / CAC fit into a growth strategy?,” I wrote:
Really, cash flow management requires comprehension of the way in which cohort daily revenues compound to make daily P&L positive and how much cash the company has in the bank to fund marketing and all other operations. These are truly strategic points for a company to dissect, and that process can’t start without knowledge of what relevant LTV / CAC ratios are.
What this means is that a payback window — that is, the period over which the money outlaid on marketing is paid back by a cohort — is really the critical component of marketing strategy, not LTV. If an advertiser is cash constrained, a Day 365 LTV is irrelevant. Advertisers are better served focusing on month-to-month cash generation than the hypothetical LTVs of users acquired today.
Yet advertisers torture themselves over LTV models that are highly fragile and almost impossible to validate (how do you back-test an LTV model when the product is updated every two weeks?). LTV modeling is actually fairly straightforward (although some marketing analysts overcomplicate it), but it’s futile to try to project an LTV curve out years into the future, especially for an immature product.
That approach simply doesn’t work. I showed in this presentation that even trying to calculate Day 4 ARPU at a 95% confidence level with 500 new users per day is fraught; projecting LTV out to Day 30 or 90 or 180 requires much, much more data than most advertisers realize.
I believe the better approach to thinking about LTV in the abstract, terminal sense is to buy traffic at a 100% ROAS target against an early-stage benchmark until enough data is accumulated to establish new, later-stage benchmarks. An example is outlined below: the marketing team sets a ROAS goal of 100% at the Day 15 benchmark and accumulates enough data to understand how the ARPU curve evolves to that point. Using that curve, the team estimates the bid level for 100% recoup at the Day 30 benchmark and buys against that (reducing the Day 15% ROAS to 80%). The team then accumulates data until they can estimate the bid for 100% ROAS at Day 60, and so on.
This is methodical, and it’s slow: if the underlying Day 60 ROAS bid for a campaign is high, an advertiser is leaving money on the table by bidding less than that just to accumulate data. But slow isn’t necessarily a bad thing when millions of dollars in ad spend are at risk. And coupling this approach with a mechanic like the Blended Same-Money Return metric allows advertisers to build robust, defensible growth strategies that don’t rely on ever-more cash being pushed into an unpredictable system against unrealistically long payback windows. A slow, winning strategy beats a losing strategy over a long enough timeframe.
What’s more, this strategy is consistent with the new, algorithmic campaign management reality that advertisers live in now. Google and Facebook et al will manage to ROAS targets for advertisers; it is wholly unnecessary to spend time and energy trying to estimate LTV across a long timeline when the largest networks are doing the heavy lifting.