Skip to navigation
Using Fyno Connect

Create Cohorts

To create a one-time or recurring campaign, you need to create a cohort as first step.

What is a Cohort?

A Cohort is a collection or a segment of users created on a set of filters or a database query. This allows you to segment users based on any criteria and send targeted messages to them.

A few examples of filters include filtering by specific status, location, or sign-up timestamp. There are countless other possibilities that you can explore by using different combinations of filters or writing your own query.

How to create a cohort?

Let’s have a look at how to create a cohort.

1From the left navigation panel, click on Cohorts. On the page that opens, click on “Create Cohort”.
2In the pop-up, provide a relevant name for the cohort, select a database to create the cohort and click on “Submit”.
3On submitting, a new page will open consisting of a dropdown menu with the list of all the schemas present in the selected database.
4Select a schema and view its data in a tabular form below. Apply filters, or run a query on it to get the desired segment of users.
5Click on “Save Cohort” to save the selected schema and the applied filters or query to complete the cohort creation process.
6Alternatively, you can click on the dropdown next to “Save Cohort” and select “Save and Create Campaign” to not only save the cohort but also create a campaign from it.

Filters and Query Builder

Filters and Query allow you to get the segment of users for certain conditions or factors.

Filters Builder

A powerful visual builder using which you can apply a set of filters to the schema data to get the desired output. This builder consists of the following major components:

Rule

It is the most important component as it consists of the single condition or the filter that needs to be applied.

A rule is created by combining 3 things:

Rule = Key + Operator + Value

The “Key” is the column name or the object key from your schema. It is present as a dropdown menu with a list of column names or object keys from your schema.

The “Operator” is the operation or check that is to be performed between the key and the value. It is present as a dropdown menu with a pre-defined list of operators.

The “Value” is the data value that you input. The data type of this field will depend on the “Key” selected and will change as you change the selected “Key”.

Example of a rule:

Age > 30, Company = Fyno, Email ends with @fyno.io, and many more.

Distinct Values

Use the highlighted icon to instantly see all unique values in your selected column. This automatic detection saves time, especially when working with large datasets, by showing you all possible filter values in a convenient dropdown menu.

Group

A group is a collection of rules that can be used to organize filters and create complex filters by combining multiple groups with different combinators.

Combinators

Determine how 2 rules or groups will combine together in order to form a complete condition or a filter. It can have 2 values:

  1. AND: The “AND” combinator evaluates to true if both of its rules or groups are true. If at least one of the it is false, the entire combined condition evaluates to false. The “AND” operator can be thought of as representing the intersection of conditions.
  2. OR: The “OR” operator evaluates to true if at least one of its rules or groups is true. If both are false, the entire combined condition evaluates to false. The “OR” operator can be thought of as representing the union of conditions.
After completely adding all the rules and groups, hit the “Apply” button to filters and view the updated data below.
Click the ‘Get Count’ button to view the total number of users in the cohort.

Query Builder

A simple code editor to write or paste your db queries and run it directly on the schema data to get the desired output. This editor gives you more freedom as you can write as complex a query as you can and also helps you reuse what you have already written someplace else.

It’s essential to use the appropriate query when working with different databases. For instance, MongoDB requires a JSON-formatted Aggregation Pipeline query which looks like this:

[
  {
    "$match": {}
  }
]

Generate a query with Fyno AI

Don’t know how to write a database query? Describe the set of users you want in plain English, and Fyno AI converts it into a query that runs on your schema data.

For example, if your user table has created_at and language columns, you could type:

Get the User_id, Name and Email of users who signed up in 2026 and whose language is English.

Fyno AI generates the query in the right format for your database:

Here’s the query for this prompt in each supported database:

[
  {
    "$match": {
      "created_at": {
        "$gte": { "$date": "2026-01-01T00:00:00Z" },
        "$lt": { "$date": "2027-01-01T00:00:00Z" }
      },
      "language": "en"
    }
  },
  {
    "$project": {
      "_id": 0,
      "user_id": 1,
      "name": 1,
      "email": 1
    }
  }
]

The query returns the User_id, Name and Email of each matching user. Notice how Fyno AI interprets the prompt: “signed up in 2026” becomes a date range on created_at, and “English” becomes the language code en as stored in the data. The table and column names in your query come from the schemas you select, so yours will differ.

1In the Use Fyno AI to generate query field, describe the segment of users you want, including all conditions.
2From the Select relevant schema/s dropdown, choose the schema or schemas that contain the data your prompt refers to. You can select more than one.
3Optionally, select Let Fyno AI use the first 3 rows of relevant schemas for accuracy. This gives Fyno AI a sample of your actual data, which helps it understand field formats and values (such as en for English) and produce a more accurate query.
4Click Generate Query. The generated query appears in the Query Editor.
5Review the query, edit it if needed, and click RUN to view the results.
6Optionally, rate the query using the thumbs up or thumbs down icon next to What do you think about this? Your feedback helps improve Fyno AI.
If the results aren’t what you expected, refine your prompt and generate the query again, or edit the query directly in the Query Editor.
Fyno AI only reads the first 3 rows of the selected schemas when you enable the sample-data option. Leave it unchecked if those rows contain sensitive information you don’t want to share.

How to download cohort data?

To download cohort data, go to cohorts page and click ellipsis present at the end of each cohort. You will see the below options.

Click Download CSV option to download the cohort data.