> For clean Markdown of any page, append .md to the page URL.
> For a complete documentation index, see https://docs.6mm.com/llms.txt.
> For AI client integration (Claude Code, Cursor, etc.), connect to the MCP server at https://docs.6mm.com/_mcp/server.

# Statistics

> Statistics — Turn conversation, ticket and knowledge activity into questions your team can act on. Start with a defined reporting period, then investigate the records behind a change in the numbers.

Turn conversation, ticket and knowledge activity into questions your team can act on. Start with a defined reporting period, then investigate the records behind a change in the numbers.

1. **Set the period** — Keep dates and filters consistent.
2. **Compare indicators** — Look at volume, handling and feedback together.
3. **Inspect the records** — Find examples and choose a concrete improvement.

## Understand support activity

Open **Statistics** and choose the relevant time range and available filters. Review real-time overview, conversations, tickets, satisfaction, messages, agents, feedback and knowledge statistics.

Use several metrics together. Conversation volume shows demand, ticket status shows follow-up work, and ratings provide feedback on individual experiences. A high knowledge view count alone does not prove an issue was resolved.

## Compare like-for-like data

Keep the same time range, timezone, filters and module scope when comparing results. Check the field labels rather than assuming all numbers count unique people. If a value differs from another screen, compare its definition and filtering first.

MCP can query statistics and help summarize trends. It cannot modify statistical records through the read-only statistics tools. Ask it to include the queried period and source metrics in the summary so the result can be checked in the console.

## Choose a view for the question

| Question                            | Useful views                                                                       |
| ----------------------------------- | ---------------------------------------------------------------------------------- |
| What changed in support demand?     | Messages, human conversations, bot activity and country distribution.              |
| Is follow-up work accumulating?     | Tickets and feedback, with the same reporting period.                              |
| How do visitors evaluate service?   | Satisfaction, followed by a review of the underlying conversations.                |
| Which help content needs attention? | FAQ and knowledge statistics, compared with repeated questions from real sessions. |

### Example: prepare a weekly review

1. Select the agreed week and record the date range and active filters in the review.
2. Summarize demand, unfinished work and visitor feedback separately. Do not add unlike metrics together.
3. Open a small set of relevant conversations or tickets to understand the reasons behind the change.
4. Choose a concrete action: revise an unclear article, update a quick reply or review team handling. Compare the next period using the same definitions.

You can ask an MCP-connected AI to summarize the selected statistics and identify questions to investigate. Require it to name the period and distinguish observed figures from suggested explanations.

## Common questions

#### Does a popular article mean it resolves most questions?

Views and interactions describe activity, not guaranteed resolution. Read visitor questions and outcomes before making that conclusion.

#### Why do two reports show different totals?

Compare the time basis, filters and metric definition first. Session counts, messages and unique visitors are different quantities.

## Continue with

#### [Inspect a conversation](/customer-support/history)

Understand the context behind a service metric.

#### [Review service usage](/customer-support/billing)

Check intelligent-service consumption separately.