Analytics

On 9 September 2026, Mixpanel added version history to Insights, Funnels, Flows and Retention reports. Every save now creates a retrievable version with a timestamp, the name of the person who saved it and a summary of what changed. You can restore an older version, duplicate it as a separate report, or copy a direct link to it. Edits made through Mixpanel Agent and the Mixpanel MCP server also appear in the history, credited to the user whose session started them.

That sounds like housekeeping, but it addresses a real problem. A product manager opens the retention chart on Monday and sees a number that disagrees with last week’s slide. Somebody changed a filter, a breakdown or a date range, and nobody can say who or when. Without a change record, teams argue about which number is right instead of what to do about it. The risk grows as AI agents start making bulk edits across boards, because one request can alter many reports at once.

Version history fixes the audit trail inside Mixpanel. It does not tell you how often reports change, which ones are unstable or whether a change moved a headline metric. That is the part you have to measure yourself.

Data Points to Track

• report_version_saved events, with report_id, report_type (insights, funnels, flows, retention) and board_id
• edit_source: manual, Mixpanel Agent or MCP server, so automated edits can be compared with human ones
• changed_field: filter, breakdown, date range, event definition or metric formula
• headline_value_before and headline_value_after for the report’s main number, captured at save time
• restore_action: restored, duplicated or link shared, with the version age in days
• reports_per_board_edit: how many reports one bulk edit touched
• Report owner and team, so changes can be grouped by who is responsible

Setup Steps

1. List your critical reports. Pick the 10 to 20 reports that feed weekly reviews or board decks, and record an owner for each.
2. Define what counts as a material change. For example, a headline value moving more than 5% after a save, or any change to an event definition.
3. Pull version history on a schedule. Use your warehouse export or the Mixpanel documentation for programmatic access, and write the results to a simple changes table.
4. Tag the edit source. Make sure Agent and MCP edits are labelled separately from manual saves in your table.
5. Alert on material changes. Send a short message to the report owner when a critical report changes in a way that moves its headline number.
6. Review monthly. Check the most-edited reports and the most-restored versions, then lock down or document the definitions behind them.

Actionable Insights

A report that is saved many times a week is usually one whose definition nobody agrees on. Treat it as a metric-definition problem and write down the agreed formula, rather than letting people keep tweaking filters.

If restores are frequent, someone is making edits that turn out to be mistakes. That points to a need for review steps, especially for Agent-driven bulk edits. Compare the share of restores that follow automated edits with those that follow manual ones, and you will see whether the approval step before an Agent change is doing its job.

Where a headline value shifts sharply after a save, check the cause before the number reaches a stakeholder. A changed date range or breakdown can look like a real trend. Linking each shift to a named version lets you answer “what changed” in minutes instead of a day of detective work.

Finally, use the history to find stable reports. Those that rarely change and are often viewed are good candidates for your official dashboard.

Related Resources

• Product Analytics for Web & Mobile
• cc:analytics API Docs
• How cc:analytics works

Expert help

Need help tracking this in your app?

Our team sets up analytics pipelines for mobile and web teams every day. Talk to us and get your first events flowing in under an hour.

Talk to an expert