# Track results over time

import { Steps } from '@astrojs/starlight/components';

Run a journey or collection again to see what changed since last time. Runs start only when someone in your organization starts them. There is no schedule.

## Before you begin

- A journey or collection that has run at least once.
- Any member of your organization can run journeys and collections and read their reports.

## Steps

### Run it again

<Steps>
1. For a journey, do one of these:
   - On the **Journeys** list, click **Run now** on its row. The row then shows that the run is queued, with a link to its reports.
   - On the journey's **Reports** tab, click **Run N times**, or **Try it once** for one run per agent.

2. For a collection, open it and click **Run collection**, or click the run button on its row in the **Collections** tab.
</Steps>

### Find earlier reports

- **A journey:** its **Reports** tab groups reports into series, one per agent and model. Each series is headed by the agent's name and notes "deltas are comparable within this series only". Each row shows the date, **Baseline** on the first report, **Running** while a run is going, how many runs finished, and **Open**.
- **A collection:** the **Reports** section of the collection page lists one report per run, with **Baseline** on the first. **Latest report** opens the newest.

## Compare reports

### In a journey report

Comparisons work one agent at a time. If the journey has more than one agent, select one under **Showing** first; results are never averaged across agents.

- **Step through a series.** When the agent's series has more than one report, arrows above the scorecard move to the previous or next report in it, labeled "same definition, same run configuration".
- **Read the movement column.** Each scorecard row shows how many more or fewer runs passed that check than in the previous report of the series, such as +2 or -1. The column stays empty when you view all agents at once.
- **See the trend.** The **Trend** section at the bottom lists the other reports in the same series.

A journey report has no charts. It uses the scorecard grid.

### In a collection report

Every later run is compared with the collection's first run:

- The headline says how many more or fewer buyer questions you win than on the first run's date.
- A question whose group changed shows what it was: **Was a win**, **Was split**, **Was a loss** or **Had no result**.
- **New** marks a journey added since the first run.
- **Question changed** marks a question whose wording changed since the first run. It is not compared, because its numbers would not mean the same thing.

## What makes two runs comparable

| Keep the same | Why |
| --- | --- |
| **Agents and models** | The app enforces this. Changing an agent's model starts a new series with its own baseline, so you never compare two models by accident. See [Choose agents and models](/journeys/agents-and-models/). |
| **Questions, persona and checks** | This one is up to you. Keep the wording the same between runs you want to compare. For a different question, create a new journey. |
| **Number of runs** | The movement column counts passing runs, not a percentage. **Try it once** and full runs land in the same series, so a one-run report next to a ten-run report shows a large jump that is not real. |

Some variation is normal. Answers can differ from one run to the next, and search results change, so a small movement between two reports can be noise rather than a change in your docs. Runs that errored are left out of the counts rather than counted as failures; every figure says how many runs finished. See [How results are measured](/journeys/how-results-are-measured/).

### The baseline

The baseline is the first report in each journey series, and the first run of a collection. It cannot be moved.

## Check it worked

- The new report appears at the top of the journey's **Reports** tab, or under **Reports** on the collection page.
- In a later collection report, questions that moved group show what they were, such as **Was a loss**.

## Next

- [Turn a gap into a fix](/journeys/turn-a-gap-into-a-fix/)
- [Collections and sharing](/journeys/collections/)
- [How results are measured](/journeys/how-results-are-measured/)