How to Define Success Metrics: From CPI to ROAS to Payback

Your success metric changes as your app matures.

Early on, you might optimize for cheap installs. Later, you shift to return on ad spend. Eventually, payback period becomes the constraint.

Using the wrong metric at the wrong time kills growth. Here's how to define success metrics that match your stage.

The Metrics Hierarchy

Success metrics exist in layers. Each tells you something different:

Volume metrics: How much did we get?

• Installs
• Spend
• Impressions

Efficiency metrics: How efficiently did we acquire users?

• CPI (cost per install)
• CPM (cost per thousand impressions)
• CTR (click-through rate)

Quality metrics: How valuable are the users?

• Retention (D1, D7, D30)
• Conversion rate
• Session frequency

Value metrics: How much are they worth?

• LTV (lifetime value)
• ROAS (return on ad spend)
• Payback period

The deeper you go, the more meaningful the metric—but the longer it takes to measure.

CPI (Cost Per Install)

Definition: How much you pay to acquire one install.

Formula: Spend ÷ Installs = CPI

Example: $1,000 spend, 500 installs = $2.00 CPI

When CPI Matters

Early testing:

When you're validating channels and don't have LTV data yet, CPI is your primary metric.

Goal: Find channels that deliver installs at acceptable cost.

Channel comparison:

CPI lets you compare efficiency across channels with different pricing models (Facebook, TikTok, ASA).

Benchmarking:

CPI gives you an immediate signal about whether campaigns are working.

CPI Limitations

Installs aren't all equal:

A $1 install that never opens the app is worse than a $5 install that becomes a subscriber.

Doesn't account for revenue:

You can have a great CPI and still lose money if users don't monetize.

Optimizing for CPI alone attracts low-quality users:

Ad networks will find cheap installs if that's what you optimize for. They won't necessarily be valuable users.

When to Use CPI

• Stage: Early testing (first 1,000-5,000 installs)
• Goal: Validate that channels can deliver users affordably
• Caveat: Track retention alongside CPI from day one

CPI targets by category (2025):

• Games: $0.50-2.50
• E-commerce: $2-8
• Finance: $5-15
• Subscription apps: $3-10
• Utility: $1-4

These vary significantly by platform and targeting.

ROAS (Return on Ad Spend)

Definition: Revenue generated per dollar spent on ads.

Formula: Revenue ÷ Spend = ROAS

Example: $10,000 revenue, $5,000 spend = 2.0x ROAS

When ROAS Matters

Post-monetization:

Once users start generating revenue, ROAS becomes the key metric.

Scaling profitably:

ROAS tells you if you can afford to spend more while staying profitable.

Channel optimization:

Compare channels based on revenue efficiency, not just user acquisition cost.

ROAS Time Windows

ROAS changes based on measurement window:

D7 ROAS: Revenue in first 7 days ÷ Spend

• Useful for quick optimization cycles
• Incomplete picture for subscription or long-term monetization apps

D30 ROAS: Revenue in first 30 days ÷ Spend

• Better predictor of long-term performance
• Standard for most apps

D180 ROAS: Revenue in first 180 days ÷ Spend

• True lifetime value proxy
• Too long for optimization cycles

Choose measurement window based on your monetization timing.

ROAS Targets

Minimum viable: 1.0x (break even)

Healthy growth: 1.5-2.0x

Highly profitable: 3.0x+

Venture-backed aggressive growth: 0.5-1.0x (acceptable short-term loss for market share)

Your target depends on business model and growth stage.

When to Use ROAS

• Stage: Post-monetization (first revenue events occurring)
• Goal: Scale campaigns that generate profitable revenue
• Caveat: Need 30+ days of data for meaningful ROAS calculation

LTV (Lifetime Value)

Definition: Total revenue a user generates over their lifetime.

Formula: Average revenue per user × Average user lifespan

Example: $5/month subscription × 8 month average lifespan = $40 LTV

Calculating LTV

For subscriptions:

LTV = ARPU × (1 / Churn Rate)

Example:

• Monthly subscription: $10/month
• Monthly churn: 10%
• LTV = $10 × (1 / 0.10) = $100

For transaction-based apps:

LTV = Average Transaction Value × Purchase Frequency × Customer Lifespan

Example:

• Average order: $30
• Orders per year: 4
• Average customer lifespan: 2 years
• LTV = $30 × 4 × 2 = $240

For ad-monetized apps:

LTV = ARPDAU × Average DAU Lifespan

Example:

• ARPDAU (average revenue per daily active user): $0.05
• User stays active for 90 days average
• LTV = $0.05 × 90 = $4.50

LTV:CAC Ratio

CAC (Customer Acquisition Cost): What you spend to acquire a user.

LTV:CAC ratio: LTV ÷ CAC

Interpretation:

• 3:1 or higher: Healthy, sustainable growth
• 2:1 to 3:1: Acceptable, room for improvement
• 1:1 to 2:1: Barely profitable or unprofitable
• Below 1:1: Losing money on every user

Example:

• LTV: $60
• CAC: $20
• Ratio: 3:1 (healthy)

When to Use LTV

• Stage: Mature app with stable retention and monetization patterns
• Goal: Understand maximum viable CAC for sustainable growth
• Caveat: Requires 90+ days of data for accurate calculation

Payback Period

Definition: How long it takes to recover the cost of acquiring a user.

Formula: Time until cumulative revenue ≥ acquisition cost

Example: Spend $10 to acquire user. They generate $2/week. Payback = 5 weeks.

Calculating Payback

Track cumulative revenue over time:

Week	Weekly Revenue	Cumulative Revenue	Cumulative - CAC
0	$0	$0	-$10
1	$2	$2	-$8
2	$2	$4	-$6
3	$3	$7	-$3
4	$4	$11	+$1

Payback period: 4 weeks (first week with positive cumulative)

Payback Period Targets

Bootstrapped / cash-constrained: 1-3 months

You need capital back quickly to reinvest in growth.

Venture-backed / growth mode: 6-12 months

You can afford longer payback if it enables faster growth.

Profitable at scale: 3-6 months

Balanced approach for sustainable, capital-efficient growth.

Why Payback Period Matters

Cash flow management:

Shorter payback = faster capital recycling = faster growth.

Risk reduction:

Recover investment quickly before user behavior changes.

Investor expectations:

Different investors have different payback requirements based on growth strategy.

When to Use Payback Period

• Stage: Scaling phase with consistent revenue patterns
• Goal: Optimize capital efficiency and cash flow
• Caveat: Requires accurate LTV curves and revenue timing data

Choosing the Right Metric by Stage

Stage 1: Validation (0-1,000 installs)

Primary metric: CPI

Secondary metrics: D1 retention, tutorial completion rate

Why: You're learning if channels work and what users cost. Don't have revenue data yet.

Goal: CPI < category benchmark, D1 retention > 35%

Stage 2: Monetization Activation (1K-10K installs)

Primary metric: D7 ROAS

Secondary metrics: CPI, conversion rate, D7 retention

Why: Revenue starts flowing. Optimize toward users who monetize quickly.

Goal: D7 ROAS > 0.5x, improving toward 1.0x

Stage 3: Scaling (10K-100K installs)

Primary metric: D30 ROAS

Secondary metrics: Payback period, LTV:CAC ratio

Why: You have enough data for meaningful LTV calculation. Focus on sustainable unit economics.

Goal: D30 ROAS > 1.5x, payback < 6 months

Stage 4: Optimization (100K+ installs)

Primary metric: Payback period

Secondary metrics: LTV:CAC ratio, channel-specific ROAS

Why: You're optimizing capital efficiency and scaling what works.

Goal: Payback < 3 months, LTV:CAC > 3:1

Metric Dashboards

Build dashboards that show metrics in context:

Campaign Performance Dashboard:

• Spend by channel
• Installs by channel
• CPI by channel
• D7 ROAS by channel
• Payback curve by channel

Cohort Dashboard:

• Retention curves by cohort
• Revenue curves by cohort
• LTV projection by cohort
• Payback timing by cohort

Channel Comparison Dashboard:

• CPI: Facebook vs TikTok vs Google
• D7 ROAS: By channel
• Retention: By channel
• LTV: By channel

This lets you make decisions based on multiple metrics simultaneously.

Common Mistakes

Optimizing for one metric exclusively:

Great CPI with terrible retention is worse than higher CPI with strong retention.

Not adjusting metrics by stage:

Using payback period when you have 500 installs is premature. Using CPI when you have 50K installs ignores revenue.

Comparing metrics across different time windows:

D7 ROAS vs D30 ROAS aren't comparable. Use consistent windows.

Ignoring retention entirely:

Revenue metrics lag. Retention gives you early warning signals.

Setting targets without context:

A "good" CPI depends on LTV. A "good" payback depends on capital constraints.

FAQs

What metrics should I track for user acquisition?

Track CPI (cost per install), retention rates (D1, D7, D30), conversion rate, LTV (lifetime value), ROAS (return on ad spend), and payback period. Which matters most depends on your growth stage and business model.

When should I optimize for CPI vs ROAS?

Optimize for CPI in early testing when validating channels and learning what works. Shift to ROAS once you have LTV data and want to scale profitably. ROAS ensures you're acquiring users at sustainable economics.

What's a good payback period?

Most venture-backed apps target 6-12 month payback periods. Bootstrapped apps often need 1-3 months. The acceptable payback depends on your capital situation, growth targets, and investor expectations.

Can I have good ROAS but bad unit economics?

Yes. If your ROAS looks good at D30 but users churn before paying back CAC, you're losing money long-term. Always track payback period alongside ROAS.

Should I stop campaigns with ROAS below 1.0x?

Depends on your strategy. Growth-stage companies often run at 0.5-1.0x ROAS to acquire market share. Mature companies usually require 1.5x+ for sustainability. Know your strategy before setting thresholds.

---

Success metrics evolve as your app grows. Start with CPI to validate channels, shift to ROAS as revenue flows, and eventually optimize for payback period as you scale. The metric you choose determines the growth you achieve.