What Insights answers
A well-configured Insight helps you answer questions such as:- How many users performed an event?
- How many times did an event occur?
- How many sessions contained an event?
- How frequently do users perform an event?
- What is the average, median, minimum, maximum, or total value of an event property?
- How does a property value behave at the user level?
- How does a metric change over time?
- How does a metric differ across user segments?
- How does the metric compare between two periods?
- How do multiple event metrics relate to each other?
Core concepts
The Insight output as a configurable metric result with filters, breakdowns, a time window, and a visualization.
How Insights works
Insights follows a simple analysis flow: Select Events → Choose Measurement → Configure Events → Apply Filters → Add Breakdown → Select Time Window → Choose Visualization → Analyze Results You can start with a simple event metric and progressively refine the analysis using the available configuration options.Build an Insight
1. Select an event
Start by selecting the event you want to analyze. For example:Application SubmittedPayment CompletedProduct ViewedCourse StartedPurchase Completed
2. Choose a measurement
After selecting an event, choose what you want to measure. Insights supports the following measurement types:
The first three measurements form the core event measurements, while the other measurement types support more advanced analysis.
Measurement types
Unique Users
Unique Users counts the distinct users who performed the selected event.Example
Event:Application Submitted
Measurement: Unique Users
Result: 8,500
This means 8,500 distinct users performed the event within the selected analysis configuration.
Total Events
Total Events counts all occurrences of the selected event.Example
If 1,000 users performedVideo Played and the event occurred 4,500 times:
Total Events = 4,500
Total Sessions
Total Sessions counts sessions containing the selected event.Example
Event:Product Viewed
Measurement: Total Sessions
The result represents the number of sessions in which the selected event occurred.
Frequency per User
Frequency per User measures how frequently users perform an event. It supports:- Average
- Median
- Distribution
Example
Suppose users performVideo Played different numbers of times during the selected period.
Frequency per User can help answer:
How often do users typically play videos?You can view the average or median frequency, or analyze the distribution of users based on their event frequency.
Aggregate Property
Use Aggregate Property when you want to measure a property associated with an event rather than simply counting the event. Supported aggregation options include:- AVG
- COUNT
- DISTINCTCOUNT
- MAX
- MIN
- SUM
- MEDIAN
- MODE
Example
Suppose thePurchase Completed event has a Revenue property.
You can configure:
Event: Purchase Completed
Property: Revenue
Aggregation: SUM
This answers:
What is the total revenue represented by the selected purchase events?
Aggregate Property per User
Aggregate Property per User lets you first calculate a property value for each user and then aggregate those user-level results. The calculation has two stages: Property → Per-user aggregation → Across-user aggregationPer-user aggregations
You can use:- Sum
- Average
- Distinct Count
- Maximum
- Minimum
Across-user aggregations
You can then use:- Average
- Sum
- Median
- Maximum
- Minimum
- Distribution
Example
Suppose you want to calculate the average revenue generated per user. You could configure: Property: Revenue Per-user aggregation: Sum Across-user aggregation: Average Conceptually:Average(Sum of Revenue per User)
Event-Level Custom Formula
Use an Event-Level Custom Formula when you need to combine event-level measurements. The formula configuration uses Events and Formula tabs. Supported arithmetic scenarios include:A + BA - BA × BA ÷ B
Example
Suppose you have: Metric A:Orders Completed
Metric B: Orders Cancelled
A formula can combine these measurements to create another event-level calculation.
Configure events
Filter an event
You can apply a filter directly to a selected event. For example: Event:Purchase Completed
Filter: Payment Method = UPI
Only matching event data is included in the calculation.
Remove an event
You can remove an event from the metric configuration. The Insight is then recalculated using the updated configuration.OR operator
Insights supports the OR operator for supported event conditions. For example:Signup Completed OR Login Completed
This allows supported event conditions to be combined within the analysis.
The AND operator is not currently included in the supplied product scope.
Filter the analysis
Filters restrict the data included in the analysis. Insights supports:- All Properties
- User Properties
- Cohorts
Property filters
Use property filters to focus the analysis on a specific population or property value. For example: Country = India or Platform = AndroidExample
To analyze Android users who completed a payment: Event:Payment Completed
Measurement: Unique Users
Filter: Platform = Android
Event property filters
Event property filters apply to the selected event. For example: Event:Purchase Completed
Event Property: Plan
Value: Premium
This allows the metric to focus on matching event occurrences.
Cohort filters
You can use an existing cohort to restrict the analysis to a defined group of users. For example: Cohort = Premium Users The metric then analyzes the selected event for users belonging to that cohort.Add a breakdown
Breakdowns divide the result into groups so that you can compare segments. Supported breakdowns include:- All Properties
- User Properties
Example
Suppose the overall result is: Payment Completed = 10,000 users Add aPlatform breakdown:
The breakdown makes it possible to compare the metric across supported segments.
Time window
Select a time window to determine which period of data is included in the analysis. Predefined time windows include:- Today
- Yesterday
- Last 7 days
- Last 30 days
- Last 3 months
- Last 12 months
- Custom Date Range
Compare time periods
Insights allows you to compare two analysis periods. For example: Current period: Last 7 Days Comparison period: Previous 7 Days This allows you to compare how the selected metric changes between the two periods.Visualizations
Choose the visualization that best represents the question you are answering.
These visualization types are supported by the supplied specification.
Time granularity
For supported time-based visualizations, you can use time granularity to group the metric over time. Time granularity is supported for:- Line
- Stacked Line
- Column
- Stacked Column
Analyze your results
Once your Insight is configured, review the resulting metric and refine the analysis as needed. You can change:- Event
- Measurement
- Event configuration
- Filters
- Breakdown
- Time window
- Comparison period
- Visualization
- Time granularity where supported
Example: Analyze application completion
Suppose you want to understand application completion among users.Configuration
EventApplication Completed
Measurement
Unique Users
Filter
Platform = Android
Breakdown
Region
Time Window
Last 30 Days
Visualization
Column
What this tells you
The Insight shows the number of unique Android users who completed an application during the selected period, segmented by region.Example: Analyze purchase revenue
Suppose you want to understand the total revenue generated from completed purchases.Configuration
EventPurchase Completed
Property
Revenue
Measurement
Aggregate Property
Aggregation
SUM
Time Window
Last 30 Days
This calculates the total value of the selected Revenue property across matching purchase events.
Example: Analyze revenue per user
Suppose you want to understand the average revenue generated by each user.Configuration
EventPurchase Completed
Property
Revenue
Per-user aggregation
SUM
Across-user aggregation
AVERAGE
This calculates:
Average of each user’s total revenueThis two-stage calculation is supported through Aggregate Property per User.
Example: Analyze event frequency
Suppose you want to understand how frequently users perform a core product action.Configuration
EventVideo Played
Measurement
Frequency per User
Analysis
Average
Time Window
Last 30 Days
This helps answer:
How frequently does a user play videos during the selected period?You can also use Median or Distribution to understand the frequency pattern.
Example: Compare two event metrics
Suppose you want to compare two event-level measurements. You can define the required event measurements and use an Event-Level Custom Formula to combine them. For example:A + B
or
A ÷ B
The formula is validated before execution.
Result actions
Insights provides additional actions for working with users represented by a result.View Users
View the users contributing to a supported metric result.Create Cohort
Create a cohort from the users represented by the result.Download Users
Download the users represented by the result. These actions allow you to continue analysis beyond the metric itself.Empty states
Insights displays an empty state when the selected configuration does not return data. For example:No data foundThe purpose of the empty state is to clearly communicate that the current configuration did not produce a result rather than leaving the visualization blank.
Loading state
When Klaritics is calculating the selected metric, a loading state indicates that the analysis is in progress.Error handling
Insights provides clear feedback for invalid or unsupported configurations. Examples include:- Invalid date range
- Invalid formula
- Unsupported configuration
- No matching data
What you can do
Build
Create an event-based metric using supported measurements.Measure
Measure users, events, sessions, frequency, event properties, user-level property values, or formula-based metrics.Filter
Narrow your analysis using supported event, user, property, and cohort filters.Segment
Break down your results using supported properties and cohorts.Compare
Compare metric performance across two analysis periods.Visualize
Switch between supported chart types to understand the result from different perspectives.Analyze
Refine your metric by changing events, measurements, filters, breakdowns, time windows, and visualizations.Act
View users, create cohorts, or download users from supported metric results.Common analysis patterns
In summary
Insights provides a flexible way to explore product behavior from event data. You can:- Select events.
- Choose how they should be measured.
- Filter event and user data.
- Segment results using breakdowns.
- Measure event frequency.
- Aggregate event properties.
- Calculate property metrics at the user level.
- Build event-level formulas.
- Analyze predefined or custom time periods.
- Compare two periods.
- Visualize results using multiple chart types.
- View, create cohorts from, or download users from supported results.