> For the complete documentation index, see [llms.txt](https://evidano.gitbook.io/ailyze-docs/llms.txt). Markdown versions of documentation pages are available by appending `.md` to page URLs; this page is available as [Markdown](https://evidano.gitbook.io/ailyze-docs/analyze-data/compare-groups-segments.md).

# Compare Groups (Segments)

See how answers differ across groups like department, region, or plan, so you can point to exactly who differs and how.

This page shows you how to compare your results across groups — like department, region, or plan — so you can see exactly who differs and how.

At the end you will have side-by-side comparisons — charts and quotes — showing how each group answers differently.

## Before you start

* A dataset with a grouping column (spreadsheet), or several documents you can tag by group.

"People are unhappy" is weak. "*Enterprise* users on *mobile* are unhappy about *onboarding*" is actionable. Comparing [segments](/ailyze-docs/reference/glossary.md#segment) is how you get there.

## In spreadsheet analysis

### Compare across all columns

Pick "Analyze text, numerical, categorical, etc. survey data across all columns". On the **Configure Column Types** board, drag your grouping columns into the **Segments** bucket (for example, **Department, Region, Plan, Tenure, Rating band**). Evidano then compares every numerical, categorical, and text result **across those segments**.

<figure><img src="https://1178311500-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fi4GBXPMUDxszIpUVOARU%2Fuploads%2FXnxxa7qsbcL2CKWj0HbY%2Fimage.png?alt=media&amp;token=4d4c6834-aea9-4b44-a0ce-a8cc07d37760" alt="Dragging grouping columns into the Segments bucket"><figcaption><p>Only columns dropped into the Segments bucket become groups you can compare.</p></figcaption></figure>

### Compare selected columns

Pick **Analyze text responses in selected columns**. Here the grouping is detected for you (**AI will auto-identify this for you**), so you don't choose segments manually. If you need explicit control, use the across-all-columns option above.

### Reading the comparison

* On any chart, use the **Segment** dropdown to switch groups, and **Put segments on y-axis** to see them side by side.
* The **Overall** dashboard shows each column broken down by segment.
* Download chart data as **CSV/PNG** for your own deck.

<figure><img src="https://1178311500-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fi4GBXPMUDxszIpUVOARU%2Fuploads%2F6jKHGz1CXdgHyx7WI63t%2FCompare%20Groups%201.png?alt=media&amp;token=baa7a7b9-9aab-425c-95c5-28033dc3af57" alt="A chart broken down by segment"><figcaption><p>One bar per segment makes the gap between groups visible without reading the table.</p></figcaption></figure>

## In document analysis

In document analysis, segments are set up in the draft, and each file (or each focus-group participant) is tagged with a value. Every builder control is listed in [Document Analysis Settings](/ailyze-docs/reference/settings/document-analysis.md#segments). The figures use the guide's sample project, compared by District (North Valley, South Ridge) and Respondent type (staff, parents).

1. **Check the segments.** When your brief or supporting files mention groups, the AI drafts them as segments. When each row is a document or participant, they are columns of **You will receive a table of answers and metadata** (while **Compare groups (e.g., by gender, year)** is ticked), labeled "Segment · N values" with their values underneath; otherwise hover a file name to see its values. The AI tags the files with them as soon as the draft opens ("Tagging files with AI…"). **Add** › **Add group to compare** adds one. You should see each segment list its values, for example North Valley and South Ridge under District.
2. **Turn on the comparison.** Under **You will receive a report**, at the top of the draft, tick **Compare groups (e.g., by gender, year)**; it is already ticked when the draft has segments. Its pencil opens the tagging dialog; it turns amber, with a note, while some files are not tagged yet.
3. **Tag each file.** In the dialog, use **Auto-populate with AI** to let the AI read the values from the files, **Add manually** to add a segment and type values into the cells, **Import XLSX/CSV** to load a sheet of file names and values, or **Use previous** to reuse an earlier project's segments. Click **Done**; it keeps values nobody is tagged with yet and offers to remove them (**Remove it** / **Remove them**). You should see the amber note disappear once every file has a value.
4. **Read the comparison.** Results include **Groups to compare** and a **Comparison across segments** section inside each [theme](/ailyze-docs/reference/glossary.md#theme) and question. It names any segment it could not compare, for example one with only one value tagged. [How to read your results](/ailyze-docs/analyze-data/how-to-read-your-results.md) explains both.
5. **Fix a tag after the run.** On the results page, click **Edit groups to compare** (Pro) in the **Codebook** header's ⋯ menu or in any comparison section, change the tags, and save. The comparison updates afterwards.

You need several files to compare. A focus group split by participant still counts as one file: "Comparing groups needs at least two files; untick it or add files".

<figure><img src="https://1178311500-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fi4GBXPMUDxszIpUVOARU%2Fuploads%2Fgit-blob-06e241e46e612ea5ed9316cab8648130bca00b7b%2Fsettings-document-analysis-27-segments.png?alt=media" alt="Which participant or document belongs to which group dialog with four tagging buttons and a Respondent group cell per row"><figcaption><p>Each focus-group participant gets their own row, so two people in the same transcript can carry different values.</p></figcaption></figure>

<figure><img src="https://1178311500-files.gitbook.io/~/files/v0/b/gitbook-x-prod.appspot.com/o/spaces%2Fi4GBXPMUDxszIpUVOARU%2Fuploads%2Fgit-blob-f9d4c70b0bd7f7f2788aea066ddb17166a9d659f%2Fsettings-document-analysis-42-edit-segments.png?alt=media" alt="Edit groups to compare dialog on a completed project with each row&#x27;s group values and Save and update comparison"><figcaption><p>Fix a mis-tagged file here instead of creating the project again.</p></figcaption></figure>

## What makes a good segmentation

* **Pick variables you can act on** — region, plan, role, version, time period.
* **Keep groups big enough** to be meaningful (and, for employee data, to protect anonymity).
* **Limit the number of segments** — 2–5 clear groups beat a dozen tiny ones.
* **Watch for confounds** — a difference by region might really be a difference by product version.

## Common comparisons by team

| Team                | Compare by                         | Question answered                        |
| ------------------- | ---------------------------------- | ---------------------------------------- |
| Customer experience | channel, plan, satisfaction band   | Where is dissatisfaction concentrated?   |
| Product             | app version, platform, cohort      | Did the new release fix or break things? |
| HR                  | department, tenure, location       | Which groups are disengaging, and why?   |
| Market research     | demographic, region, customer type | How do needs differ by group?            |

## A realistic example

> **Scenario:** A pulse survey across 5 departments. Segment by **department**, and the Overall dashboard reveals that *workload* dominates in Support while *career growth* dominates in Engineering — each with quotes — turning a flat average into a targeted action plan.

## Next steps

* [Analyze employee feedback](/ailyze-docs/common-research-workflows/analyze-employee-feedback.md) — a common segmented workflow.
* [Discover themes and patterns](/ailyze-docs/analyze-data/discover-themes-and-patterns.md) — find the themes you'll compare.
* [How to read your results](/ailyze-docs/analyze-data/how-to-read-your-results.md) — read the comparison charts.