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

# Analyze your data

> Ask questions about your revenue, orders, ad spend and customers, and get answers from your connected data

Ask Minerva Agent a question about your connected data in plain language. It answers with numbers, tables and charts, and shows the sources behind them.

## Before you start

* **Connect the sources your question needs.** Use the table below as a guide. It names each kind of data the way [Setup check](https://docs.minerva.io/platform/connect-data.md#setup-check) does.
* **Make sure your data is ready.** The Agent can analyze your data once it's connected and ready for analysis.
* **Make sure your data covers the whole calculation.** A metric needs all of its inputs. Return on ad spend, for example, needs both revenue and ad spend, for the same period and in the same currency.

| Question about | Data it needs |
| - | - |
| Customers and leads | **Customer records** |
| Revenue, orders and purchases | **Orders & revenue** |
| Leads and pipeline | **Sales pipeline** |
| Website and app behavior | **Web & app activity** |
| Email and lifecycle performance | **Email & messaging** |
| Ad spend and campaign performance | **Ad spend** |

A question that spans sources also needs those sources to share fields that can be matched. Connecting two tools doesn't by itself link their records.

## Ask a question

**1. Open the Agent**

Select [**Agent**](https://app.minerva.io/agent) in the sidebar, or select **Ask Minerva** on the page you're working on.

**2. Ask with the details that matter**

State the question, the date range, the comparison and the breakdown you want. Say how you define any metric that could be read more than one way.

**3. Check the answer**

Review the time period, the units and the filters the Agent used. Select **Sources** on a chart to see what's behind it, as **Data** or as **SQL**. Ask the Agent to explain anything surprising.

**4. Follow up**

Change the period, compare channels, look at a cohort or ask for a chart. Each follow-up can run new queries.

**5. Save it or run it again**

Ask for a [report](https://docs.minerva.io/platform/agent/reports-and-sharing.md) to keep the analysis as a document. Create a [schedule](https://docs.minerva.io/platform/agent/schedules.md) to run it again on a cadence.

## Example questions

Each of these works once you've connected the data it needs.

* "Compare revenue and orders for the last 28 days with the previous 28 days. Show the biggest changes by channel."
* "Which campaigns had the largest change in spend last week, and what happened to conversions?"
* "How does repeat purchase rate differ by first-purchase month?"
* "Which product categories grew fastest, and was it more buyers or larger orders?"

## Read the results

* **Gaps are called out.** When a source is missing or a query fails, the Agent says so rather than showing a zero.
* **Detail tables are capped.** A table that lists individual rows shows a limited number of them. Ask for a total when you want to know the size of a group.
* **Reports are saved analysis.** Opening a report doesn't run its queries again. Ask the Agent to refresh it for current numbers.
* **Platforms count differently.** Conversions reported by an ad platform aren't the same as paid orders, and platforms differ in attribution window, currency and time zone.
* **Facts and suggestions are separate.** The Agent keeps the figures it queried apart from its own suggestions, so you can test a suggestion before you act on it.

## Limits and billing

Analysis is read-only. Asking for analysis doesn't change a campaign, activate a segment or modify anything in a connected tool.

Running analysis and saving reports don't use your record export allowance. If you then ask the Agent to export or activate people, that is a separate step with its own confirmation, billing and limits.


This documentation is built and hosted on [Mintlify](https://mintlify.com), a developer documentation platform.