When a Platform reports “25% Growth,” the next question should not simply be “Is that good?”
It should be: “What caused the Growth?”
A 25% increase in Transactions can describe very different business situations.
One Platform may be acquiring many New Users who use it once and disappear.
Another may have almost no User Growth, while existing users return more frequently.
A third may have stable Users and Frequency, with Revenue Growth coming entirely from higher Average Order Value.
The headline Growth Rate can therefore hide very different Growth Engines, each requiring different decisions across Marketing, Product, Retention and Monetization.
The goal is not to add more KPIs. It is to decompose the headline number into components that explain whether Growth comes from “more people,” “more frequent usage” or “more value per use.”

Do not ask only whether the Platform is growing. Ask what is driving the Growth.
If Transactions increase by 30%, we know that activity on the Platform has increased.
We do not yet know why. Growth could come from:
• More Active Users
• Existing users returning more frequently
• More Transactions per visit
• Higher Value per Transaction
• A combination of several factors
A useful starting point is:
Platform Activity = Active Users × Average Usage Frequency
To connect activity with business value: Platform Value = Active Users × Usage Frequency × Value per Use
Then Drill Down further into whether Active User Growth comes from New or Returning Users and which Segments are driving higher Frequency.
Definitions matter. Active Users, Sessions, Transactions and Critical Events measure different levels of behavior. Google Analytics explicitly distinguishes Total, Active, New and Returning Users, while Product Analytics frameworks distinguish Retention from Usage Frequency or Stickiness.

First, define what “Platform Growth” actually means

Platform Growth can refer to many different metrics:

• Registered Users
• Active Users
• Sessions
• Transactions
• Orders
• Content Consumption
• Gross Merchandise Value or GMV
• Revenue
• Subscriptions

These metrics are not interchangeable.
More Registered Users does not necessarily mean more Active Users.
More Active Users does not automatically mean better Retention.
More Sessions does not necessarily mean more Transactions.
More Transactions does not guarantee Revenue or Profit will grow at the same rate.
Before analyzing Growth, define the Outcome the business is trying to explain.
For a Marketplace, it may be Orders or GMV.
For a Subscription Platform, it may be Active Subscribers or Recurring Revenue.
For a Content Platform, it may be Active Users and Content Consumption.
For a Service App, it may be Completed Transactions or another Critical Event representing the core value delivered by the service.
Amplitude describes a Critical Event as an action closely aligned with the value users receive from a Product and recommends focusing Engagement Analysis on meaningful actions rather than arbitrary activity.

Start by separating “how many people” from “how often they use”

For a transaction-based Platform, a simple decomposition is: Transactions = Active Users × Average Transactions per Active User
Last month:
Active Users = 100,000
Transactions per User = 2.0
Transactions = 200,000

This month:
Active Users = 120,000
Transactions per User = 2.0
Transactions = 240,000
Transactions grew by 20%.
In this case, Growth came from more Active Users while Frequency remained unchanged.

Now consider:
Active Users = 100,000
Transactions per User = 2.4
Transactions = 240,000
Transactions still grew by 20%.
But the Growth Story is completely different.
The first case is User-base Growth.
The second is Frequency Growth.
The headline number is identical.
The Business Implication is not.

User Growth and Frequency Growth answer different questions

User Growth asks: “Are more people using the Platform?”
Frequency Growth asks: “Are existing users using it more often?”
Both can create Growth, but they point toward different underlying dynamics.
Users ↑ + Frequency ↓
The Platform may be acquiring more people while Engagement per User weakens.
Users → + Frequency ↑
The existing user base may be developing stronger Usage Habits or more frequent Use Cases.
Users ↑ + Frequency ↑
Growth is being supported by both User Base and Frequency.
Users ↓ + Frequency ↓ + Revenue ↑
Investigate Value per Use, Pricing, Mix or Monetization.

If Active Users grow, Drill Down into New vs Returning Users

Suppose Monthly Active Users increase from 100,000 to 130,000.
Do not immediately conclude that the Platform is becoming healthier.
The Growth could come from:
Case A New Users increase sharply while Returning Users remain flat.
Case B New Users remain flat while Returning Users increase.
Case C Both New and Returning Users increase.
Case D New Users increase sharply while Returning Users decline.
Case D is particularly important.
Headline Active Users may still look strong because Acquisition is replacing users who are not returning.
Google Analytics distinguishes New Users from Returning Users: New Users are associated with first_visit or first_open, while Returning Users have initiated at least one previous Session. Separating these groups helps prevent User Growth from being treated as one homogeneous metric.

New User Growth is not the same as Retention Growth

Consider: Month 1
New Users = 50,000
Returning Users = 70,000

Month 2
New Users = 80,000
Returning Users = 68,000
Total Activity may grow because of an Acquisition Campaign.
What the data supports is: Acquisition increased.
What it does not yet establish is:
Those New Users will Retain.
Follow the Cohorts into:
Week 1 Retention
Week 4 Retention
Month 3 Retention
or Repeat Transaction Rate based on the Platform’s Natural Usage Cycle.
Retention measures whether users return over time, while Stickiness or Usage Frequency examines how frequently they use the Product within a defined period. They are related, but they are not the same metric.

Returning Users increase, but what are they returning to do?

An increase in Returning Users is useful information.
It is still not enough.
Users may return to:
Open the App
Browse Products
Search
Read Content
Check Status
or complete a Transaction.
If the Platform delivers its primary value through Purchase, counting every App Open as meaningful Engagement may overstate business-relevant behavior.
Define a Critical Event.
Examples:
E-commerce → Completed Purchase
Food Delivery → Completed Order
Travel Platform → Booking
Payment App → Successful Transaction
Learning Platform → Lesson Completion or another defined Learning Activity
B2B SaaS → Core Workflow Completion
A Critical Event does not always have to be a Revenue Event, but it should be closely related to the core value the Product is intended to deliver.

Measure Frequency according to the Natural Usage Cycle—not every Platform should optimize DAU

DAU, WAU and MAU are widely used Product Metrics.
But Daily Active Users should not automatically become the primary KPI for every Platform.
A Messaging App may naturally be used daily.
Food Delivery may be used several times per week by some Segments.
Some E-commerce categories may naturally be monthly.
Travel Booking may occur only a few times per year.
Some B2B Platforms follow weekly or monthly workflows.
Amplitude notes that Stickiness should be interpreted according to the Product Usage Interval because different Products have different natural usage frequencies. DAU/MAU can therefore be misleading for Products that are not designed for daily use.
Before asking: “Is our DAU/MAU good?”
Ask: “How often should users naturally receive value from this Product?”

Higher Frequency can indicate Engagement, or Friction

More Sessions per User is not automatically good news.
Suppose a Banking App reports a 25% increase in Sessions per User.
One explanation is: Users are more engaged.
Alternative explanations include:

• Failed Transactions require repeated attempts
• Status information is unclear, so users keep checking
• Payment requires too many Steps
• Notifications generate visits without meaningful Value
• Users repeatedly return to correct Errors

Frequency should therefore be interpreted alongside Critical Events and Outcomes.
For example:
Sessions per User ↑
Successful Transactions per User →
Error Rate ↑
In this case, Higher Frequency may reflect Friction rather than Healthy Engagement.
Google Analytics distinguishes Sessions, Engaged Sessions, Engagement Rate and Key Events precisely because a Session is not automatically equivalent to meaningful Engagement or a Business-relevant Event.

Do not rely only on Average Frequency

Suppose Average Transactions per User increase from 2.0 to 2.6.
Engagement appears stronger.
But the distribution may reveal:
Light Users: 1.2 → 1.1
Medium Users: 2.8 → 2.7
Heavy Users: 8.0 → 12.5
The average increased because a small group of Heavy Users became much more active.
The overall User Base did not necessarily become more engaged.
Consider a Frequency Distribution such as:

• 1 transaction per month
• 2–3 transactions
• 4–7 transactions
• 8+ transactions

or Segment users according to business-relevant usage patterns.
Stickiness Analysis follows a related logic by examining how many days or periods users perform a defined Event rather than relying only on a single aggregate average.

Extend Users × Frequency into Value per Use

If the objective is to explain Revenue Growth, Users and Frequency are not enough.
Extend the decomposition: Revenue = Active Users × Transactions per Active User × Revenue per Transaction
Suppose:
Active Users +10%
Transactions per User +5%
Revenue per Transaction -8%
Revenue may still grow while Monetization per Transaction weakens.
Now consider:
Active Users →
Frequency →
Revenue per Transaction +20%
Revenue can grow even though User Engagement has not improved at all.
Therefore: “Platform Revenue increased”
does not automatically mean: “The Platform has more users and stronger Engagement.”
Revenue Growth and User Growth are different phenomena.
Decompose the Components before drawing the conclusion.

For a Marketplace, separate both Demand Side and Supply Side Growth

Some Platforms serve more than one User Group.
A Marketplace may have:
Buyers
Sellers
Ride-hailing may have:
Riders
Drivers
A Delivery Platform may have:
Customers
Merchants
Riders
‍
Total User Growth can hide an imbalance between the different sides.
For example:
Customers +30%
Merchants +5%
Orders +20%
Growth looks positive.
But if Demand grows faster than Supply capacity, the Platform may experience:

• Longer Waiting Time
• Lower Availability
• More Cancellations
• Lower Service Quality

Multi-sided Platforms should therefore examine Growth on each side together with Interaction Metrics between them.

Cohort Analysis shows whether new Growth behaves differently from old Growth

Total Active Users combine people who joined at different times.
Cohort Analysis allows the business to ask:
What percentage of users acquired in January returned in later months?
Do users acquired after a new Campaign Retain better or worse than earlier Cohorts?
Does Channel A generate users with different Month 3 Frequency from Channel B?
Example:
January Cohort
Month 1 Retention = 42%
Month 3 = 28%
June Cohort
Month 1 = 55%
Month 3 = 39%
Even with the same Total Active Users, the June Cohort shows stronger return behavior in this example.
However, improved Retention does not establish the Cause.
Possible explanations include Product Changes, Customer Mix, Acquisition Channel, Seasonality or other factors.
Cohort Analysis identifies Patterns.
A Pattern is not Causal Proof.

Example: Transactions grew 30%, where did the Growth come from?

Suppose a Platform reports: Transactions +30%
Start the Drill Down:
Transactions ↑ 30%
→ Active Users ↑ 25%
→ Transactions per User ↑ 4%
Then Drill Down Active Users:
→ New Users ↑ 55%
→ Returning Users ↑ 3%
Then New Users:
→ Paid Social ↑ 80%
→ Organic ↑ 8%
Then Cohort Analysis:
→ Paid Social New Users show lower Month 2 Retention than Organic Users
The evidence supports the conclusion that most of the observed Growth is associated with New User Growth, particularly Paid Social, rather than substantially higher Frequency among existing users.
It does not yet prove that Paid Social caused all of the Platform Growth.
Campaign overlap, Seasonality, Attribution Error and External Factors may also contribute.
The next question should therefore not simply be: “Should we increase Media Budget?”
It should be: “Do the New Users we are acquiring at greater scale Retain and generate enough Customer Value to justify the acquisition cost?”

A Platform can become stronger without User Growth

Consider:
Monthly Active Users → 500,000
Transactions per User 2.1 → 2.8
Retention ↑
Critical Event Completion ↑
Revenue per User ↑
The number of Users has not increased.
But existing users are using the Product more frequently and generating more Value.
For some Platforms, this may represent higher-quality Growth than acquiring large numbers of users who never return.
However, “higher quality” still requires an Economics check.
Examine:
CAC
Revenue
Margin
Incentive Cost
Cost to Serve
Higher Engagement does not guarantee Profitability.

Be careful when Promotion temporarily increases Frequency

Suppose Coupons increase Transactions per User from 2.0 to 3.2.
Do not immediately conclude that the Platform has developed stronger User Habits.
Check:

• Does Frequency remain higher after the Promotion ends?
• Are the additional Transactions truly Incremental or merely shifted in Timing?
• Are users buying more because of Product Value or Incentive?
• What happens to Margin after the Promotion?
• Does the promoted Cohort Retain differently from a Comparison Group?

Higher Frequency during the Campaign can be a FACT in the Behavioral Data.
“Promotion created Loyalty” is a Causal Claim requiring stronger evidence.

A Platform Growth Dashboard should have Layers, not just headline totals

Instead of placing dozens of KPIs on one screen, organize the Dashboard into Diagnostic Layers.
Layer 1: Business Outcome

• Revenue
• Transactions
• GMV or Critical Business Outcome

Layer 2: User Base

• Active Users
• New Users
• Returning Users

Layer 3: Engagement

• Frequency per User
• Active Days
• Critical Events per User

Layer 4: Retention

• Cohort Retention
• Repeat Rate
• Reactivation

Layer 5: Monetization

• Revenue per User
• Revenue per Transaction
• Average Order Value

Layer 6: Economics

• CAC
• Incentive Cost
• Margin
• Cost to Serve

This turns the Dashboard from a Reporting Tool into a Diagnostic Tool.
Instead of only showing whether Growth occurred, it helps explain where to investigate next.

Eight questions to ask before interpreting Platform Growth

1. Are we talking about Growth in Users, Sessions, Transactions, GMV or Revenue?
2. What activity qualifies someone as an Active User?
3. Is Active User Growth coming from New or Returning Users?
4. Is Usage Frequency per User increasing or decreasing?
5. Is Frequency increasing across the User Base or only among Heavy Users?
6. Are users returning to perform the Critical Event or merely opening the Platform?
7. Do new Cohorts Retain as well as earlier Cohorts?
8. Does the Growth still produce acceptable Revenue, Margin and Unit Economics?

These questions help turn Platform Metrics into inputs for business decisions rather than numbers for reporting alone.

The takeaway: Decompose Platform Growth into “Users × Frequency × Value”

When a Platform grows, do not explain the headline number too quickly.
Start with: Platform Value = Active Users × Usage Frequency × Value per Use
Then Drill Down: Active Users → New Users + Returning Users
Usage Frequency → Active Days / Sessions / Critical Events / Transactions per User
Value per Use → Price / Average Order Value / Revenue per Transaction / Mix
Then examine:
Retention
Cohort
Segment
Channel
Promotion
Economics
Each Metric answers a different question.
Active Users tell you how many people use the Platform.
Frequency tells you how often they use it.
Retention tells you whether they return over time.
Critical Events tell you whether they perform behaviors related to the Product’s core value.
Revenue tells you how much financial value is generated.
No single Metric answers all of these questions.
So when Platform Growth reaches +30%, the more useful question is not simply: “Is 30% good?”
It is: “Did the 30% come from more users, more frequent usage or higher Value per Use—and does that Growth Engine remain healthy when we examine Retention and Economics?”