Skip to main content

​
Overview

Investigating anomalies or trends in your data is extremely manual. Teams spend hours each week running the same repetitive process (picking properties, running breakdowns, eyeballing segments) to figure out what drove a change in a metric. Mixpanel’s AI-Powered Root Cause Analysis (RCA) feature does that work for you. Launch RCA from any Insights report or fired alert and an AI agent automatically diagnoses what changed: it validates the anomaly, runs the relevant breakdowns, ranks the dimensions that contributed most, and writes up an interpretation with a confidence level and suggested next steps — all delivered as a Board you and your team can keep working from.

​
How it works

When you run an RCA, Mixpanel immediately launches a new Board and exposes the agent’s reasoning as it works:
  1. Validates the anomaly. Confirms there’s a real change to explain before going further. RCA compares your flagged point against a baseline window that ends right before it, sized automatically from your chart’s time grain (hourly uses roughly the prior 24 hours, daily ~14 days, weekly ~9 weeks, monthly ~6 months); if your chart shows a longer range than that, RCA uses your range instead. It then measures how many standard deviations your point sits from that window’s average (the z-score), with 2 or more being the bar for a real anomaly rather than routine noise. A seasonality check compares like-for-like periods (same weekday for daily data, same hour-of-day for hourly) so normal recurring patterns don’t get flagged.
  2. Selects properties to analyze. An LLM agent picks which properties to break down on, combining properties Mixpanel suggests for the event with any you’ve prioritized in Custom RCA Settings. It spreads picks across plausible cause categories (acquisition, audience, platform, data quality, user identity), filters out high-cardinality identifiers and near-duplicates, and factors in your project’s business context.
  3. Runs breakdowns. Executes the chosen breakdowns to see which segments actually moved.
  4. Writes an interpretation. Fills the Board with contribution factor analysis (each segment’s share of the total change), ranked dimensions, an AI-written explanation of the likely cause, a confidence level, and suggested next steps.
Cards stream into the Board as each step completes, with a live progress banner at the top while the agent is still working.
What the numbers mean. The percentage change (sometimes called “lift”) compares the data point you flagged against a baseline, which is the average of the periods just before it: (flagged value - baseline) / baseline, so +30% means it came in 30% above the recent norm and a negative number means below. Each segment’s contribution percentage is its share of the total change: (how much the segment moved) / (how much the whole metric moved), so if a metric dropped by 500 overall and one country accounts for 400 of that, it shows an 80% contribution. Confidence works differently: it’s not a calculated number but a label (“High,” “Medium,” or “Low”) the AI assigns to each finding based on how unusual the change is versus normal fluctuation (the z-score) and how cleanly a single segment explains it, so a big, clear change driven by one segment earns higher confidence than a small change spread across many segments or one that looks seasonal.

​
Where to launch RCA

1

From an Insights report

Open any Insights chart, hover over a data point that looks unexpected, and choose Root Cause Analysis from the popover. RCA launches scoped to that event, segment, and time window.
2

From a fired alert

When an alert fires, the notification (Slack, email, or webhook) includes a link that runs Root Cause Analysis and generates a Board for that alert. See Alerts for how RCA fits into alerting.

​
The RCA Board

Because RCA generates a standard Mixpanel Board, it inherits all the capabilities of our typical Boards:
  • Anyone in your project can open it. RCA Boards are project-shared by default, so teammates can pick up the analysis where you left off.
  • You can edit it like any Board. Add cards, rearrange, annotate, or duplicate it if you want a private working copy.
  • Cards include contribution factor analysis. Each ranked dimension lands as its own card with the breakdown, contribution percentage, and supporting context. You can edit this as needed before sharing.
  • The interpretation card summarizes findings. The AI’s written explanation, confidence level, and suggested next steps live at the top of the Board so anyone scanning the results gets the punchline first.

​
Custom RCA Settings

Oftentimes, customers know which properties matter for their business and want to direct the root cause analysis to focus there first. In the AI Settings tab of your Project Settings, you can specify which event properties RCA should prioritize during analysis. For example, a marketing team that knows anomalies usually trace back to utm_campaign_id can point RCA at that dimension so it’s always considered first. Properties you prioritize are combined with Mixpanel’s automatic suggestions, so you never lose the agent’s broader view — you just guide it toward the dimensions you care most about.
These settings only influence which properties RCA prioritizes. They don’t control RCA’s analysis sequencing or methodology — regardless of the instructions you provide, the agent still validates the anomaly, runs its breakdowns, ranks contributing dimensions, and writes up its interpretation the same way.

​
Feedback

Every generated Board has a thumbs-up / thumbs-down feedback widget. We use this signal to evaluate RCA quality and improve future runs. If RCA missed the mark, please rate it — it’s the most direct way to influence what we work on next.

​
Limits

  • Daily RCA runs per project are capped by plan: Enterprise 300/day, Growth 70/day, Free 10/day.
  • RCA today reasons over events and properties. Adjacent context like annotations, session replays, experiments, and metric trees aren’t included in the V0 agent’s reasoning.