Mixpanel + Kite

Kite starts with the reports and definitions your team has already built in Mixpanel. It searches your project for saved reports, dashboards, and metrics, and it reads the results of the ones that matter, such as a weekly active users chart or a signup report someone pinned months ago. It also reads the business context your team saved in Mixpanel, the short notes that explain what an event like "Activated" means at your company.

That context matters because event names are rarely self-explanatory. If your team counts an account as activated once it invites a second teammate, Kite uses that definition instead of guessing from the name. From there it runs fresh queries on the same events, such as signups by week broken down by UTM source, and writes up what changed and why it matters in Slack. You can ask for that readout once or have Kite deliver it every Monday.

Kite builds funnels from the events you already track to find the exact step where people give up. A typical funnel might run from a pricing page view, to a "Start free trial" click, to a completed signup, to a first project created. Kite reads the conversion rate at each step and splits it by properties such as device, plan, country, or campaign, so it can tell you that the drop is mostly mobile visitors from one paid campaign rather than everyone.

Flows show where people go instead. Kite reads the most common paths before and after a step, such as the pages people visit right before they cancel or where they wander when they leave a signup form half finished. When session replay is on in your project, Kite also reads the replays recorded for specific people, along with the events that happened during each one. That combination usually turns "signups dropped" into a specific cause Kite can propose fixing.

A campaign that brings in signups who never return costs more than it looks like, and Mixpanel's retention data is where that shows up. Kite runs retention queries grouped by acquisition source or campaign to see what share of each channel's signups come back in the weeks that follow. It also measures how long people take to get from first visit to activation for each campaign.

Here's how that plays out. Say a LinkedIn campaign and a Google search campaign each brought in 200 signups last month. If 40 percent of the search signups are still active four weeks later and only 10 percent of the LinkedIn signups are, Kite flags that gap. It then recommends where to shift effort, and it can turn that finding into a landing page, an email, or a change to the campaign itself through your other connected tools.

To make those reports readable, Kite can build a lookup table in Mixpanel that maps campaign IDs to campaign names. It can also create formula-based custom properties, such as a plan tier calculated from raw billing fields, so a breakdown reads in the terms your team uses.

Kite builds Mixpanel cohorts, which are saved groups of users defined by what they did or who they are. One example is everyone who viewed the pricing page at least twice in the last 14 days without starting a trial. Another is everyone who activated last quarter and hasn't logged in for 30 days. Kite can also update a cohort's definition as your criteria change, or remove one your team no longer uses.

A cohort gives the rest of the work a precise audience. Kite uses it to break down funnels and retention, so it can compare the people who stalled with the people who didn't. Your team can also use the cohort anywhere Mixpanel lets you use one, including cohort syncs you've already set up to your email or ad tools.

Once Kite knows which step to fix, it can test the fix as a Mixpanel experiment. Before it sets anything up, it searches your project for earlier experiments on the same feature or metric, so your team learns from the last test instead of repeating it. Then it writes the hypothesis, picks the primary metric, and creates the experiment as a draft. Before launch, it checks the setup against Mixpanel's pre-launch checks, including whether your expected traffic can reach the sample size the test needs.

Kite brings that setup to you in Slack, and launches the experiment once you approve. While the test runs, it reads the lift, confidence interval, and significance for each metric. It also checks whether the traffic split has drifted from what was configured, which would make the results unreliable. When the test ends, Kite recommends shipping the winner or ending without shipping, and it records the decision in Mixpanel.

Mixpanel decides who sees which version, and your product shows that version by checking the experiment's feature flag. Kite can create and update those flags too, including rolling a feature out to a growing percentage of users or switching it off.

Tracking plans get messy as teams grow, and a funnel built on the wrong event gives the wrong answer. Kite reads the data quality issues Mixpanel finds in your project, such as a property that suddenly arrives as the wrong type. It also finds groups of near-duplicate names, like "Add to Cart," "add_to_cart," and "addToCart," and proposes merging them into one. A merge can be undone from Mixpanel if your team changes its mind.

Kite also maintains Lexicon, Mixpanel's data dictionary. It writes descriptions and display names for events that have none, marks the events your team should rely on as verified, hides or deprecates ones that are no longer used, and flags properties that hold personal information. It organizes events with tags and can update up to 50 events or properties at a time. Kite can also save a shared metric, such as "Marketing qualified signup," so every later report and experiment uses one definition.

Kite builds Mixpanel dashboards that combine reports with short text explanations, so your team can open one board and understand the numbers. It can create a dashboard from scratch, add or remove reports on an existing one, or copy a dashboard as the starting point for a new launch. A launch dashboard might hold the signup funnel, retention for users who arrived that week, and a note on what Kite changed and when.

Kite can also bring Mixpanel numbers into a dashboard it publishes itself. That dashboard can combine live Mixpanel charts with research and results from your other tools at one stable link, and it refreshes the data when someone opens it.