Skip to main content
Stickiness measures how frequently users engage with a specific event over a selected time interval. Most analyses answer how many users performed an event. Stickiness answers a different question:
How consistently and repeatedly are users performing the event?
Use Stickiness to understand habit formation, engagement quality, retention behavior, the split between casual and power users, and how consistently a feature is adopted.

​
What Stickiness answers

Stickiness helps you answer:
  • How many days per week or per month do users perform this event?
  • What share of users are habitual rather than occasional?
  • Is engagement depth improving or declining over time?
  • How does repeat usage compare across features?
  • How strong is retention for this behavior?
Stickiness is one of the strongest indicators of product-market fit, habit formation, engagement quality, and retention strength. Teams commonly apply it to messaging apps measuring daily conversations, SaaS platforms measuring recurring workflows, collaboration tools measuring repeated interactions, and social platforms measuring return frequency.

​
Core concepts

Multiple events on the same day count as only one active day. A user who performs the event ten times on Tuesday has one active day, not ten.

​
Configuration options

​
Intervals

Window: 7 calendar daysExample: Monday → SundayMaximum bucket: 7 days

​
Build a Stickiness analysis

1

Select an event

Choose the event to analyze. Stickiness supports a single event only. You can apply property filters to the selected event.
2

Choose the interval

Select Weekly or Monthly. This determines the size of the measurement window and the maximum number of active days.
3

Set Computed as

Select Non-Cumulative or Cumulative. Non-Cumulative is the default.
4

Apply filters

Optionally narrow the analysis using event properties, user properties, or cohorts.
5

Select a date range

Choose a preset or custom range. The range always snaps to complete interval boundaries.
6

Review the results

Switch between the line chart, bar chart, and data table, then export if needed.

​
Metric section

The metric section contains the event selector along with the Computed as dropdown.

​
Event selection

You can select a single event only, and you can apply properties to the selected event.

​
Computed as

Two options are available: Non-Cumulative and Cumulative. Non-Cumulative is selected by default. The two modes answer different questions:

​
Filters

Click the Filter option to open the properties popup. The following property types are available:
  • All properties
  • Event properties
  • User properties
  • Cohorts

​
How Stickiness is calculated

Both modes work from the same underlying data: the number of active days each user recorded in the interval. For the examples below, assume this dataset:

​
Non-Cumulative

Non-Cumulative mode answers:
How many users performed the event on exactly N days?
Each user belongs to exactly one bucket.
Applied to the dataset above, the weekly distribution is: Total users is the sum of every bucket:
Each bucket is then expressed as a percentage:
For the 3-day bucket:
Meaning: 20% of users used the feature in exactly 3 days.

​
Monthly example

The same method applies across a calendar month. With this distribution: Total users: 8,581 For 14-day users:
Meaning: 10.42% of users used the feature on exactly 14 days.

​
Cumulative

Cumulative mode answers:
How many users performed the event on at least N days?
Users appear in multiple buckets.
Using the same dataset: The cumulative counts become: Percentages use the same shape of formula:
For users with at least 3 active days:
Meaning: 60% of users used the feature at least 3 days.

​
Visualization

You can switch the visualization between a Line chart and a Bar chart. Both use the same axes:

​
Data table

The data table displays these columns by default:
  • Interval date
  • Users count
  • Day-wise distribution
All users’ data is shown by default. Expanding the table reveals the data for each interval. The table includes a search bar that works on the interval date column, and results can be exported as CSV.

​
Date range selection

The date range selector adapts to the interval you selected, and always normalizes your selection to complete interval boundaries.
Presets
  • Last 4 Weeks
  • Last 8 Weeks
  • Last 12 Weeks
  • This Month
  • This Quarter
  • This Year
Custom options: Between, Last, Since, ThisSelection behaviorDate selection must always align to complete weekly boundaries, and you cannot select dates that fall in the middle of a week. If you select a custom range, Klaritics adjusts it to the corresponding week start date and week end date.ExampleIf the workspace week starts on Monday and you select May 14, 2026, the system converts it to:

​
Validation rules

The date picker visually indicates whether selection is happening at week level or month level. Partial week and partial month selections are not permitted, and all charts, tables, and calculations use normalized interval boundaries only.

​
In summary

  • Stickiness counts distinct active days, not event volume.
  • A day counts once no matter how many times the event fires.
  • Non-Cumulative splits users into exactly-N-day buckets; each user appears once.
  • Cumulative counts users with at least N active days; users appear in multiple buckets.
  • Weekly caps at 7 active days; Monthly caps at 30 or 31.
  • Date ranges always snap to whole weeks or whole months.

​
Next steps