Mobile app acquisition is the channel ecommerce founders most often misprice. They either assume CAC has to be desktop-low or they accept it as inevitably high and stop trying to optimize. Both moves are wrong. The right framing starts with mobile LTV, which is not the same as your overall LTV because the install graph behaves differently.
Why mobile LTV is different
A buyer who installs your app behaves measurably differently from a buyer who only visits via web. Higher repeat purchase rate (30–50% higher in our data), higher order frequency, higher push-driven incremental revenue. The same person, depending on which channel they engage on, generates different lifetime value. The install is the discriminator.
This means you cannot just plug your overall LTV into a CAC calculation and reason about mobile acquisition. You need an LTV specific to the install cohort, modeled forward, that captures the channel-specific behavior.
Calculating install LTV
Three inputs: install-to-first-order rate, average order value among installed users, and order frequency over a 12-month window. Multiply them together for total revenue, multiply by gross margin for contribution margin, and you have a working install LTV.
A worked example. 60% install-to-first-order, $75 AOV among app buyers, 3.5 orders in 12 months. Revenue per install: 0.60 × $75 × 3.5 = $157.50. At 55% gross margin, contribution: $86.63. That is your 12-month install LTV before any push or referral uplift.
Push uplift adds another 15–25% to the install LTV in our data. Referral adds 5–15% depending on how the program is designed. The all-in install LTV including these effects sits around $110–$130 in the example above. That number is what your install CAC ceiling should reference.
Install CAC ceiling
A common rule of thumb is to keep CAC at one-third of 12-month LTV. For the example brand, that means an install CAC ceiling of $35–$45 on a $110–$130 install LTV. Above that, acquisition stops being a positive-margin activity in any reasonable timeframe.
In practice, paid install acquisition on Apple Search Ads, Meta, or TikTok currently delivers CPIs in the $3–$12 range depending on category and creative quality. Well below the $35 ceiling. That is why paid mobile acquisition usually pencils out for ecommerce — the LTV is high enough relative to the CPI to leave meaningful margin.
Accounting for owned channels
Most installs in the first six months of an app channel come from owned media — mobile-web banners, email, packaging, referrals. These installs do not have a paid CAC, but they are not free either. The cost is opportunity cost: every mobile-web visitor who installs the app is not converting to a mobile-web purchase in that session.
For modeling purposes, the practical rule is that owned-channel installs have a CAC roughly equal to the foregone mobile-web revenue, minus the install LTV they enable. In almost every case this nets out positive — the LTV of an installed buyer exceeds the value of a single mobile-web session — but it is worth modeling so you do not overestimate the "free" nature of owned acquisition.
Payback period
Payback period — how long it takes for the cumulative contribution margin of an install cohort to exceed its acquisition cost — is the metric most useful for cash management. A 12-month payback is fine if you have plenty of working capital. A 4–6 month payback is what you want if cash is tight.
Most healthy ecommerce app channels see payback in the 3–8 month range, depending on category and acquisition mix. Channels heavy in paid social tend to pay back slower because CPIs are higher; channels heavy in owned acquisition pay back faster but are volume-constrained.
“The cheapest install you can buy is the one you already have. Owned acquisition is the only channel where payback is measured in days, not months.”— Appolar growth deck
Common modeling pitfalls
Pitfall 1: assuming web LTV equals install LTV. They diverge meaningfully. Use channel-specific LTV.
Pitfall 2: ignoring incremental push revenue. Push-driven revenue does not show up in standard LTV calculations because it is attributed to the campaign, not the install. Add it back when modeling install LTV.
Pitfall 3: not accounting for the install half-life. Not all installs stay active forever. Uninstalls happen. A small percentage of the cohort each month drops off. The 12-month LTV should be net of expected uninstall rates, which for healthy apps run 2–4% per month after the first 90 days.
Pitfall 4: counting one-time discounts as costs of acquisition. A first-order discount of 15% is part of the install acquisition cost, but it is also a part of the install LTV (it makes the first order happen). Brands that double-count this conservatively underestimate the channel.
Done correctly, the mobile LTV math is the most optimistic part of an ecommerce founder's spreadsheet. The install is an asset; once it is on the buyer's phone, it generates revenue with no ongoing acquisition spend. The brands that internalize this find that the right amount to spend on install acquisition is meaningfully more than they were spending, not less.
