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

# Segments and quotas

> Recruit the right mix of participants and compare how segments respond differently.

A screener tells you who qualifies. Segments and quotas control the *mix* — say, half iPhone users and half Android users, or a minimum of five participants per age bracket. You set targets per segment, and recruitment fills them.

## Why it matters

Without quotas, recruitment fills with whoever passes the screener first — and that's rarely a balanced sample. Quotas guarantee the composition you designed for. And because segments carry through to analysis, your results show where groups agree and where they genuinely split.

## Setting up segments

Segments are built from your screener questions. From the Recruitment tab, click **New quota**:

1. On the **Segments** tab, click **New segment** and name it (e.g., "iPhone users").
2. Drag answer options from your screener questions into the segment. Participants whose answers match belong to that segment.
3. On the **Quotas** tab, give each segment a minimum and/or maximum participant count.

Quotas display as "exactly 10" when min and max match, or "min 5, max 10" when they differ.

<Info>
  Segments come from single- and multi-select screener questions. If the attribute you want to segment on isn't in your screener yet, add the question first.
</Info>

## How quotas fill

During recruitment, each segment shows a progress bar against its target. When a segment hits its maximum, new applicants who match it are screened out automatically — they're told the study has received enough responses. Everyone else keeps flowing through. Quotas apply per recruitment round, so a new round starts fresh against that round's targets.

## Segments in your results

* **Segments section** — your Summary shows how each segment was defined and who fell into it, with per-segment findings where the groups genuinely diverged.
* **Theme tags** — themes are labeled with the segments that drove them.
* **Quote attribution** — with segments, quotes are attributed by segment and age ("Lapsed user, 38") instead of a participant number, so you can read a finding and see which group it came from.

Segments with too few participants are excluded from comparisons (the report says which), so a single loud voice doesn't get presented as a group pattern.

<Tip>
  Two segments is the sweet spot for most studies. More segments means fewer participants per segment, and comparisons need a handful of people on each side to mean anything.
</Tip>