Acquisition and Channel Performance Metrics
Acquisition metrics help growth teams understand which channels bring valuable users and at what cost. Start with basic source-level metrics: Installs, New Users, Cost Per Install (CPI), Cost Per Acquisition (CPA), and Cost Per Mille (CPM) for ad impressions. Calculate CPI = Ad Spend / Number of Installs; CPA = Ad Spend / Number of Conversions (e.g., paying users). But raw install counts are insufficient: measure retained users at key milestones (D1, D7, D28) coming from each channel to estimate quality. Combine acquisition data with downstream behavior to compute effective CAC (Customer Acquisition Cost) = Total Spend on Channel / Number of Users who Reach Target Value (e.g., payers or LTV threshold).
Attribution nuances matter: be explicit about attribution windows (click-through vs. view-through), last-touch vs. multi-touch models, and deterministic vs. probabilistic attribution when privacy constraints apply (SKAdNetwork, ATT). For paid UA, track ROI metrics: Return on Ad Spend (ROAS) = Revenue Attributed to Campaign / Ad Spend. For subscription or LTV-heavy games/apps, measure payback period (days to recover CAC) and LTV:CAC ratio — a common healthy benchmark is >3:1 but depends on margins and churn.
Segment by creative, audience, placement, geography, and device. Track cohort-adjusted metrics to avoid channel bias: e.g., a channel may deliver many installs but poor D7 retention, inflating CPI efficiency but harming economics. Use UTM or MMP tags for consistent tracking, and validate server-side events to prevent double-counting. Set automated alerts for sudden CPI spikes, install drops, or anomalous conversion rates to quickly investigate channel health. Prioritize channels that deliver sustainably positive unit economics after accounting for retention and monetization.
Engagement and Retention: DAU, MAU, and Stickiness
Engagement and retention are the heartbeat of mobile growth — they determine sustainable monetization and organic virality. Core metrics include Daily Active Users (DAU), Monthly Active Users (MAU), and the DAU/MAU ratio (stickiness). DAU = unique users who open the app in a day; MAU = unique users in a 30-day window. Stickiness = DAU / MAU indicates how often monthly users are active daily; values closer to 1 suggest high habitual usage. However, these aggregated metrics hide cohort dynamics: track retention curves (D1, D3, D7, D14, D28) by acquisition cohort to see whether product improvements or campaigns change lifetime engagement.
Beyond simple active-user counts, instrument events that represent meaningful engagement: sessions per user, average session length, depth (levels completed, content consumed), and key feature adoption rates. Funnel metrics — e.g., onboarding completion, first purchase, level 5 completion — give actionable levers. Optimize onboarding to reduce early drop-off: measure time-to-first-value (how long until a user experiences the app’s primary benefit) and aim to minimize it. Implement behavioral triggers (push, in-app messaging, email) targeted by lifecycle stage — for example, re-engagement campaigns for D3–D7 churn risk users, and reward-driven outreach for lapsed high-LTV cohorts.
Beware of vanity metrics: raw installs and MAUs can grow while retention deteriorates. Instead define north-star metrics tied to user value (e.g., weekly active purchasers or sessions per paying user). Use segmentation (demographics, acquisition source, device, geography) to reveal differential retention patterns. Also factor platform constraints: privacy changes may reduce granular user-level attribution, so shift to aggregated and cohort-based analyses while validating signals with statistically significant sample sizes. Instrument server-side events for reliability, maintain consistent event naming, and include versioning so past cohorts remain interpretable.
Monetization Metrics: ARPU, LTV, and Conversion Funnels
Monetization metrics determine whether growth is profitable. The primary finance-oriented metrics growth teams track are ARPU (Average Revenue Per User), ARPPU (Average Revenue Per Paying User), LTV (Lifetime Value), and conversion rates across the purchase funnel. ARPU = Total Revenue / Total Users over a period; ARPPU isolates only paying users and is useful for pricing and promotion analyses. LTV estimates the present or projected revenue a user will generate over their lifetime; calculate LTV using cohort-based revenue curves or predictive models that account for retention decay and monetization trends. Common simplified formula: LTV ≈ ARPU_per_period * (1 / churn_rate) with appropriate period alignment, but more accurate models use discounted cash flows and cohort aggregation.
Track the purchase funnel: Viewed Offer → Initiated Purchase → Completed Purchase → First Receipt Redemption. Measure conversion rates and drop-offs at each step and instrument microtransactions and subscriptions separately. For subscription models, monitor MRR (Monthly Recurring Revenue), ARR, churn rate, and upgrade/downgrade flows. For IAP (in-app purchases), capture metrics like average order value (AOV), purchase frequency, and time-to-first-purchase. Also monitor deferred revenue and refunds to avoid overstating LTV.
Unit economics synthesis ties monetization to acquisition: compute LTV:CAC ratio and payback period (days until cumulative gross margin from a cohort covers CAC). Incorporate gross margin (after platform fees, payment fees) to avoid misleading LTV. Use cohort-level LTV to spot changes in product-market fit or pricing effects. Additionally, monitor promotional impact: discounts may lift conversion rate but depress long-term LTV if substituting full-price purchases. A/B test pricing and bundles while measuring long-term effects (not just immediate conversion uplift). Finally, ensure revenue data is reconciled with finance (server-side receipts, app-store reports) and that analytics handles currency normalization, refunds, and platform revenue shares.
Experimentation and Cohort Analysis for Growth Optimization
Experimentation and cohort analysis are the engines for iterative growth improvements. Implement a robust experimentation framework: random assignment, sufficient sample sizes, pre-defined primary and secondary metrics, and statistical significance thresholds. Primary metrics should reflect business outcomes (e.g., retention or revenue), not easy-to-change vanity metrics. Track experiments with a hypothesis, expected effect size, and runtime estimates; stop rules and sequential testing methods must be defined to avoid p-hacking.
Cohort analysis complements experiments by revealing lifetime effects and heterogeneity. Create cohorts by acquisition date, experiment variant, campaign, or user behavior (e.g., completed onboarding). Compare retention curves, monetization, and engagement across cohorts to understand long-term impacts. Use survival analysis to model retention decay and to estimate expected lifetime value more accurately than simple averages. When testing features that impact pricing or monetization, run long-duration holdouts to capture slower-moving behaviors like repeat purchases and subscription renewals.
Practice safe experimentation: incremental changes can have complex system-wide effects (e.g., UX tweak increases short-term conversion but reduces retention). Use guardrail metrics (unrelated metrics you monitor for negative impact) and run canary rollouts before full launches. For cross-platform products, consider interference between web, Android, and iOS experiments. With privacy constraints, adapt by using aggregated metrics and uplift modeling when individual-level identifiers are limited. Finally, operationalize learnings: maintain an experiment registry, summarize results with clear implications, and translate successful experiments into product or acquisition playbooks so wins scale across campaigns and markets.
