Product analytics metrics measure how people reach value, use features, return, and generate revenue from a digital product. Useful measures name an event, a population, and a time window so the result can guide a decision.
If your dashboards disagree, adding another chart won’t settle the question. First check what each number counts: an active user, a paying customer, and a subscription are different units.
15 product analytics metrics to track
The 15 metrics below cover three connected questions: are people getting value, do they stay, and does the business benefit?
- Activation and engagement: activation rate, time to first value, active users, feature adoption rate, and stickiness.
- Conversion and continued use: conversion rate, retention rate, churn rate, and user growth rate.
- Business outcomes: monthly recurring revenue (MRR), annual recurring revenue (ARR), average revenue per user (ARPU), average revenue per paying user (ARPPU), customer lifetime value (CLV), and customer acquisition cost (CAC).
Choose the measures needed for your team’s next decision.
Activation and engagement metrics
1. Activation rate
Activation rate measures the share of eligible new users who reach a defined first-value event within a set period.
Formula: activated users / eligible new users x 100.
Choose an action that shows value, such as completing a useful workflow, rather than simply signing in. Mixpanel’s adoption guide makes this distinction with its own product: opening a report alone does not show that someone has gained value by exploring data.
For a team product, you may need activated accounts rather than individual users. Set that unit and the activation window before comparing onboarding versions.
2. Time to first value
Time to first value measures how long a user takes to complete the first-value event after a defined starting point.
Calculation: first-value timestamp - starting timestamp; summarize the elapsed times with a median.
Pick a start such as signup or workspace creation, and keep it consistent. Then inspect steps that delay users, including setup or data imports.
Report activation rate beside the median among activators. Faster times in that group can hide users who never reached value.
3. Active users
Active users are unique users who perform a qualifying action during a stated period. Daily, weekly, and monthly active users are commonly shortened to DAU, WAU, and MAU.
Calculation: count distinct users performing the chosen event in the day, week, or month.
Use a behavior that matters to the product, and document whether background events or simple logins qualify. Otherwise, a tracking change can look like audience growth.
Compare equal-length periods with the same identity rules. An increase in MAU tells you that more people were active, but doesn’t establish that an earlier signup cohort stayed.
4. Feature adoption rate
A launch can produce many first clicks without showing that the feature has become part of a useful workflow.
Feature adoption rate measures how many eligible users use a particular feature. One useful definition is the share of eligible active users who trigger its meaningful usage event within the reporting period.
Formula: eligible active users using the feature / eligible active users x 100.
Exclude users whose plan or role cannot access the feature. That keeps the denominator relevant when evaluating discovery or onboarding.
Separate trying the feature once from repeated use. A launch can produce many first clicks without showing that the feature has become part of a useful workflow.
5. Stickiness
Stickiness describes how frequently an active audience uses a product. DAU/MAU compares a day’s active users with monthly active users; some reports instead use average DAU over the month.
Formula: DAU / MAU x 100, with the exact daily and monthly definitions stated.
For a weekly workflow, WAU/MAU may fit better. Choose the cadence users need, rather than treating daily use as a goal for every product.
Don’t label a ratio healthy without context. A monthly billing tool and a daily collaboration app serve different habits, and stickiness does not replace a retention cohort.
| Question | Metric | Definition to settle first |
|---|---|---|
| Are new users reaching value? | Activation rate | Value event and activation window |
| How long does the path take? | Time to first value | Starting event and non-activators |
| Are people using the product? | Active users and stickiness | Qualifying event and usage cadence |
| Is a feature being used? | Feature adoption | Access eligibility and repeat use |
Conversion and continued-use metrics
6. Conversion rate
Conversion rate measures the share of eligible users who complete a defined outcome within a chosen window.
Formula: unique users completing the outcome / unique eligible starters x 100.
For a trial-to-paid funnel, start with eligible trial users rather than all website visitors. Count people once, and specify whether completing the outcome after the window counts.
Break the funnel into steps to locate the problem. A low overall rate doesn’t tell you whether users failed at account creation, setup, or payment.
7. Retention rate
Retention rate measures how many members of a starting cohort return to perform a chosen event. A cohort is a group sharing a starting condition, such as signup week.
Formula: eligible cohort members performing the return event at the selected interval / eligible starting-cohort members x 100.
Amplitude’s retention documentation makes the starting and return events explicit. Specify whether you mean return on day seven or on or after day seven; those produce different answers.
Only compare cohorts old enough to reach the interval. New signups don’t belong in an older cohort’s denominator; use customer retention metrics for deeper customer-level measures.
8. Churn rate
Customer churn rate measures customer loss over a period. For a fixed beginning-of-period cohort, count which of those customers leave during the period.
Formula: customers lost from the starting cohort / customers in that starting cohort x 100.
Document what counts as leaving: cancellation, the last subscription ending, or another business rule. Product inactivity is a separate behavior definition, and revenue churn measures money rather than people.
Check dashboard definitions before reconciling totals. Stripe’s subscriber-churn calculation uses its own population rules, including new subscribers in the denominator; the fixed-cohort formula above won’t automatically reproduce it.
9. User growth rate
User growth rate measures the percentage change in a consistently defined user count between periods.
Formula: (ending count - starting count) / starting count x 100.
For example, moving from 500 to 550 monthly active users gives 10% growth. Use the same qualifying behavior and comparable periods, so the increase reflects a like-for-like count.
A zero starting count makes percentage growth undefined. Report the absolute increase instead, and examine acquisition, retention, and returning users to understand what changed.
| Question | Metric | Definition to settle first |
|---|---|---|
| Did starters complete the outcome? | Conversion | Eligible starters and completion window |
| Did the original users return? | Retention | Starting cohort and return interval |
| Which existing customers left? | Churn | Starting population and loss rule |
| Did the active audience expand? | User growth | Same qualifying behavior in both periods |
Revenue and acquisition metrics
Behavior measures need a business context. These financial measures belong alongside activation and retention, with definitions matched to the business model; the broader SaaS analytics guide covers how those reporting jobs fit together.
10. Monthly recurring revenue (MRR)
MRR is the monthly-normalized value of recurring subscriptions. It separates recurring revenue scale from the timing of payments.
Calculation: sum the monthly-equivalent recurring amounts across included subscriptions.
An annual $1,200 subscription contributes $100 MRR, as Stripe’s MRR guidance explains. Adding the full annual payment to that month’s MRR would overstate the run rate.
Exclude one-time charges, and document the treatment of discounts, overdue subscriptions, and other billing states. Compare recurring growth separately from cash received.
11. Annual recurring revenue (ARR)
ARR expresses the recurring-revenue run rate on an annual basis. Under a consistent MRR-based definition, annualize the current monthly value.
Formula: ARR = MRR x 12.
For example, $5,000 MRR corresponds to $60,000 ARR. New subscriptions, expansion, contraction, and churn change the underlying recurring base.
Keep the snapshot date and inclusion rules aligned when comparing periods. ARR is not the cash already collected this year or a promise of the next year’s sales; Stripe’s ARR guide distinguishes it from total revenue.
12. Average revenue per user (ARPU)
ARPU measures revenue relative to a defined user population. The label alone doesn’t tell you which users or revenue a dashboard includes.
Formula: relevant revenue for the period / defined users for that period.
For a usage-based product, that might be monthly revenue divided by monthly active users. A subscription dashboard may instead use MRR divided by paid subscribers, as Stripe does.
Write the definition beside the chart before comparing plans or segments. A change in customer mix can move the average even when no individual customer’s spending changes.
13. Average revenue per paying user (ARPPU)
ARPPU measures revenue per paying user, excluding free users from the denominator.
Formula: relevant period revenue / users who paid during that period.
State whether the numerator is total purchase revenue or a defined recurring-revenue amount. For the same revenue, dividing by all users answers a different question from dividing by paying users.
Read ARPPU beside the share of users who pay. A rising average among a shrinking paying group can conceal weak conversion, so investigate the population change before celebrating the result.
14. Customer lifetime value (CLV)
CLV estimates a customer’s value over the relationship. Revenue-based and margin-based estimates answer different questions, so label the model.
Repeat-purchase revenue model: average order value x purchases per customer per period x expected customer lifespan in those periods.
Simplified subscription revenue model: monthly revenue per customer / monthly customer churn rate, using churn as a decimal.
The subscription shortcut assumes stable conditions; zero or very low churn makes it unreliable. Stripe’s CLV guide describes multiple estimation methods rather than one formula for every business.
If you include gross margin, state that adjustment explicitly. A lifetime revenue estimate is not realized profit, and early cohorts may not yet provide a dependable lifetime estimate.
15. Customer acquisition cost (CAC)
CAC measures acquisition spending per newly acquired customer.
Formula: acquisition-focused sales and marketing costs / new customers acquired in the matched period.
Include the costs your definition requires, such as advertising, acquisition staff, agencies, and relevant tools. Stripe’s SaaS CAC guide explains why the cost scope and time frame matter.
For a long sales cycle, this month’s spend may create customers later. Use a period or cohort approach that reflects that lag, and compare CAC with a clearly defined margin or lifetime-value measure rather than an incompatible revenue figure.
| Question | Metric | Data needed |
|---|---|---|
| How large is the recurring base? | MRR and ARR | Subscription amounts and billing intervals |
| What revenue does each population generate? | ARPU and ARPPU | Matched revenue and user populations |
| What might a relationship be worth? | CLV | Revenue or margin and lifetime assumptions |
| What does acquisition cost? | CAC | Acquisition costs and newly acquired customers |
How should you choose product analytics metrics?
Choose metrics around a decision, then define the behavior, population, time window, and data source needed to answer it. A small set that explains a product problem is more useful than a dashboard collecting every available number without a clear owner or next action.
For an onboarding problem, start with activation, time to first value, and one retention measure. Pair a primary goal with a guardrail, such as errors or support requests, so a faster workflow doesn’t hide another problem.
Write these five details into a metric definition:
- Decision and owner: name who will act on a change and what they can investigate.
- Event and unit: specify the action and whether you count people, accounts, or subscriptions.
- Population: record eligibility, plan access, and internal/test-user exclusions.
- Time: set the interval, timezone, cohort entry, and completion window.
- Source and checks: identify the events or billing records, then inspect sample records for missing or duplicate data.
A hypothetical cohort example
Suppose 100 eligible users sign up in one week. You define activation as completing a useful export within seven days; 40 users do so, giving a 40% activation rate.
Once all 100 users have reached day seven, 25 perform the chosen return event in the defined day-seven window. Their day-seven retention is 25%, using the original signup cohort, not the 40 activators, as the denominator.
If the 40 activators have a median time to value of two days, report it alongside the 40% activation rate. The time figure says nothing about how quickly the other 60 users might eventually activate.
These are invented arithmetic examples, not customer results or a claim that activation caused retention.
When does a metric become a vanity metric?
A metric becomes a vanity measure when it looks impressive without helping explain performance or guide a decision. Context, comparison, and a useful denominator can make a number actionable; the metric’s name alone does not determine whether the number is useful or misleading.
Nielsen Norman Group illustrates why cumulative totals can rise while giving little insight into experience changes. Add a time window, rate, or relevant segment before deciding what the trend means.
Bounce rate and time measures aren’t automatically useless. A longer task time might reflect either engagement or confusion; check completion and user feedback before interpreting it.
Also use the current definition: GA4 bounce rate is the share of non-engaged sessions. Engagement can come from a session longer than ten seconds, a key event, or at least two page/screen views, so a single-page visit is not necessarily a bounce.
Match the data source to the question
In-app activation, feature adoption, and behavioral retention need reliable event tracking and an appropriate analytics tool. Billing and transaction records support a different set of revenue, customer, and subscription questions.
Putler’s commerce product dashboard is relevant when you need product sales, purchase history, or commerce reporting. Don’t assume those reports automatically instrument an application’s first-value or feature-use events.
For broader reporting choices, start with the sales analytics guide and keep the source definitions beside your dashboard. Choose the tool after agreeing on the question and data required.
Start with one decision and a checked definition
Choose one product question this week, write its metric definition, and trace a few real records through the calculation. Add a retention or revenue measure that checks whether the immediate improvement supports the longer-term goal.
Review a change only after checking tracking quality, comparable cohorts, and the time window. A number moving in the desired direction is a reason to investigate, not proof that the last release caused it.
FAQ
What do you measure in product analytics?
Product analytics measures user behavior and outcomes, including activation, feature usage, conversion, retention, and related revenue. The exact measures depend on the product and decision. Define meaningful events, a consistent user or account population, and a time window; pricing or interface design alone is not a measurement of how people use the product.
Which product metrics should SaaS teams track first?
Start with activation, time to first value, one cohort-retention measure, and recurring revenue if the business uses subscriptions. Add a conversion or feature-adoption measure for the problem being investigated. Treat this as a starting set, not a universal scorecard: the value event and expected usage cadence must fit the product.
What is the KPI of a product analyst?
A product analyst does not have one universal KPI. The relevant key performance indicator depends on the business question, such as increasing activation or reducing a specific funnel drop-off. An analyst helps define the measure, validate data, interpret changes, and support a decision; producing more dashboards alone does not establish product success.
What is the difference between product metrics and analytics?
Product metrics are defined quantitative measures, such as a seven-day activation rate or monthly recurring revenue. Product analytics is the work of collecting and examining data to understand those measures, compare segments, investigate changes, and inform decisions. A metric reports a value; analysis checks what happened, possible explanations, and what evidence is still missing.