> ## Documentation Index
> Fetch the complete documentation index at: https://docs.simulant.tech/llms.txt
> Use this file to discover all available pages before exploring further.

# Build a cohort

> Define the audience you want to reach — from source data, demographic filters, and data-backed psychographic tags.

A **cohort** is the audience a test or focus group runs against. It's a slice of the grounded population, defined by the data it's built from and the filters you apply. Every cohort reports a sample size (`n`) so you always know how many personas stand behind a result.

<Frame caption="The Cohorts view — each audience with its age band, sample size, source count, and psychographic tags.">
  <img src="https://mintcdn.com/simulant/j7X70BC9NUZuF_VF/images/app/cohorts-list.png?fit=max&auto=format&n=j7X70BC9NUZuF_VF&q=85&s=b6e7208a9c917c38688599394eb99aa9" alt="Cohorts list showing voter audiences such as Suburban Swing Voters and Rural Independents" width="1428" height="779" data-path="images/app/cohorts-list.png" />
</Frame>

## Create a cohort

Go to **Cohorts → New cohort** and work down the builder. A live preview on the right shows the sample size and an example respondent as you refine.

<Frame caption="The cohort builder — toggle source datasets and apply filters while the preview shows the sample size and an example respondent.">
  <img src="https://mintcdn.com/simulant/j7X70BC9NUZuF_VF/images/app/cohort-builder.png?fit=max&auto=format&n=j7X70BC9NUZuF_VF&q=85&s=7acdabe5f8b310a5a258fa33e09fab2d" alt="Cohort builder for Suburban Swing Voters with source datasets and a sample respondent preview" width="1428" height="779" data-path="images/app/cohort-builder.png" />
</Frame>

<Steps>
  <Step title="Name your cohort">
    Give it a name you'll recognize later, such as `Suburban Swing · Undecided`.
  </Step>

  <Step title="Choose the source data">
    Under **Source data**, toggle which datasets this cohort is built from. Turning a source off removes its signal from the personas. See [Datasets](/data/datasets).
  </Step>

  <Step title="Apply demographic filters">
    Narrow the sample by **age range**, **gender**, **state or territory**, **income**, **household**, and **education**. Leave a filter as **Any** to keep it open.
  </Step>

  <Step title="Add psychographic tags">
    Select **data-backed tags** — attitudes and dispositions such as `Cost-of-living anxious` or `Climate-focused`. Tags are grounded in the source data, not guessed.
  </Step>

  <Step title="Create the cohort">
    Click **Create cohort**. It's now available to any test or focus group in the workspace.
  </Step>
</Steps>

<Tip>
  Keep cohorts focused. A tight, well-defined cohort produces sharper segment cuts than one broad audience — and you can always build several and compare them.
</Tip>

## Sample size

As you filter, Simulant shows the available sample size (`n`). When you run a test, you choose how many personas to draw from the cohort; the panel is capped at the cohort's total. Larger samples tighten the confidence of quantitative measures; smaller samples run faster.

## Editing and reusing cohorts

Open any cohort from **Cohorts** to adjust its sources, filters, or tags and **Save changes**. Cohorts are reusable across tests and focus groups, which keeps results comparable — running two creatives against the same cohort is a clean like-for-like read.

## Next

<CardGroup cols={2}>
  <Card title="Add a creative" icon="clapperboard" href="/guides/creatives">
    Upload the asset you want this cohort to react to.
  </Card>

  <Card title="Run a test" icon="flask" href="/guides/tests">
    Put your cohort to work.
  </Card>
</CardGroup>