#relationships·Jul 17, 2026·6 min read
CPI vs LTV: Which Metric Tells You If Your User Acquisition Is Profitable?
The core decision question is: Are you paying too much to acquire users who never pay back? CPI (Cost Per Install) measures the cost to get a new user, while LTV (Lifetime Value) measures the total revenue that user generates over time. Together they answer whether your ad spend is an investment or a loss.
Cost vs. Revenue: The Fundamental Split
CPI is an acquisition cost metric — it tells you how much you spent per install. LTV is a monetization metric — it estimates the total revenue a user will generate.
- CPI = Ad spend / Number of installs
- LTV = Average revenue per user over their entire lifetime
The critical relationship: LTV must exceed CPI for a campaign to be profitable. If CPI > LTV, you lose money on every user.
Why the confusion?
Both metrics are used to evaluate user quality, but they measure opposite ends of the funnel. A low CPI can hide poor retention; a high LTV can mask high acquisition costs.
Which to use when
Pick CPI when:
- You're optimizing for volume and need to control cost per install
- Running UA campaigns where the immediate goal is installs
- Comparing ad networks on cost efficiency
Pick LTV when:
- You're evaluating long-term profitability of a user cohort
- Deciding whether to scale or kill a campaign
- Comparing user quality across sources (e.g., organic vs. paid)
Use both together when:
- Calculating LTV / CPI ratio — a ratio > 1 means profitable UA
- Setting bid caps based on predicted LTV
- Reporting to stakeholders who care about ROI, not just installs
How they diverge
What they measure
- CPI: Cost to acquire one install (input)
- LTV: Revenue generated by one user over lifetime (output)
Time horizon
- CPI: Snapshot — measured at the moment of install
- LTV: Cumulative — requires days, weeks, or months of data
Profitability signal
- CPI: Alone cannot tell you if you're profitable
- LTV: Alone cannot tell you if acquisition is efficient
Where they overlap
Both are user-level averages
Both CPI and LTV are calculated by dividing total spend or total revenue by the number of users. They hide individual variance.
Both depend on attribution
Accurate install attribution (for CPI) and revenue/retention attribution (for LTV) are required. Misattribution skews both.
Both are used in UA optimization
Ad networks and DSPs use both CPI and LTV signals to optimize delivery and bidding.
Real scenarios
The cheap install trap
A gaming studio runs a campaign with CPI = $0.30 — far below their $1.00 target. They scale spend.
- What happened: After 30 days, retention drops to 5% and in-app purchases are negligible.
- What they checked: LTV after 30 days = $0.15.
Takeaway: Low CPI does not guarantee profitability. LTV must be measured to avoid scaling a losing campaign.
The high-quality source
A subscription app sees CPI = $5.00 from a premium ad network — double their average.
- What happened: Users from that network have 40% retention at 90 days and high subscription conversion.
- What they checked: LTV after 90 days = $12.00.
Takeaway: A high CPI can be profitable if LTV is proportionally higher. Never kill a source based on CPI alone.
How they work together
When you need to control cost per install — e.g., setting a maximum bid for a UA campaign, comparing ad network efficiency, or hitting a target CPA.
When you need to evaluate long-term user value — e.g., deciding whether to scale a campaign, segmenting users by revenue potential, or forecasting ROI.
When you need to calculate profitability — e.g., computing LTV / CPI ratio to determine if your UA spend is generating positive returns.
Side-by-side snapshot
| Lens | CPI | LTV |
|---|---|---|
| Definition | Cost per install | Lifetime value per user |
| Formula | Ad spend / Installs | Total revenue / Number of users |
| Time horizon | At install moment | Over entire user lifetime |
| Profitability signal | None alone | None alone |
| Used for | UA cost control | Revenue forecasting & ROI |
Common pitfalls
Optimizing CPI without LTV
Why it's wrong: You can drive CPI down by targeting low-quality users who never monetize.
- What to do instead: Always pair CPI with at least a short-term LTV estimate (e.g., D7 revenue) to ensure cost efficiency doesn't destroy value.
Using LTV without a time horizon
Why it's wrong: LTV is a forecast. If you use a 365-day LTV for a campaign you'll evaluate in 7 days, the numbers won't match.
- What to do instead: Use cohort-based LTV with a clear time window (e.g., D30 LTV) and align it with your optimization cycle.
For learning only. Not advice on bids or spend.