Define cohort membership and return events separately in Mixpanel, then export a consistent retention report. Compare completed activity windows with paid outcomes using a worked business example and spreadsheet checks.
Mixpanel cohort analysis compares defined groups over time, but the result depends on who enters the group and which action counts as returning. Set those rules separately, export completed periods, and compare the activity counts with business outcomes on the same population. A stronger activity rate is a useful signal; it does not by itself measure renewal or profit.
Start with a business question: “Did people who joined after our onboarding change return to complete useful work, and did more of them become paying customers?” The starting event should identify eligible entrants. The return event should represent the value they came back to receive.
For this example, an illustrative US scheduling software business uses “Signup Completed” for entry and “Schedule Published” for returning activity. It compares two disjoint signup groups. These are illustrative event names; select the actual events approved by your reporting owner.
Mixpanel's cohort documentation describes groups defined from properties or behavior. You can open the cohort builder from a report through Create Custom > Cohort, or save users selected from a report visualization. Use membership conditions that describe the intended entrants, rather than requiring the very future outcome you intend to measure.
For example, selecting only people who eventually paid and then calculating paid conversion would build the answer into the group. Start with all eligible signups, including those who never paid. Keep acquisition and exclusion rules in the report notes. The distinction between starting groups and actions is developed further in behavioral cohort analysis.
Open + Create New, choose Retention, and select the starting event and returning event. Choose the measurement, filters, breakdowns, and interval. Use On for activity in the specified interval. On or After, the documented default, also counts returns in later intervals. Mixpanel's Retention guide explains these choices and marks incomplete buckets with an asterisk.
In this example, the activity outcome is a qualifying return in the completed first-week bucket under the selected On definition. Record the project's interval settings rather than silently substituting calendar weeks or another elapsed-time convention. Ask the analytics owner to confirm the exact window before comparing it with billing.
To download the report, use the upper-right … menu, select Export, and choose CSV. For a user profile file, use Users, select the desired profiles and columns, and click Export. Mixpanel's export instructions describe these steps and the export limits. Check that the downloaded row count covers the intended population.
Save the cutoff date, event definitions, membership conditions, interval, filters, and report measurement with the CSV. A screenshot preserves context, but the numeric export is what you need for a reproducible comparison.
Both groups below have completed the first-week activity bucket and a separate first-30-day business outcome window. The reporting owner supplies the paid-user and revenue columns from a reconciled billing summary. For this small example, each signup corresponds to one customer account, so users and paying accounts share the same reporting unit.
The paid outcome means becoming a paying customer within the first 30 days. Revenue is net billed revenue within that same window, after credits, in USD. These outcomes are separate from the first-week product return event.
Put the following inputs in columns A through E:
| Signup group | Eligible signups | First-week active users | Paid users within 30 days | First-30-day net revenue (USD) |
|---|---|---|---|---|
| Before change | 120 | 72 | 30 | 6,000 |
| After change | 80 | 56 | 24 | 5,280 |
| Total | 200 | 128 | 54 | 11,280 |
Add columns F through I for activity retention, paid conversion, revenue per entrant, and revenue per payer. Enter the formulas in row 2 and copy to row 3:
F2 =IF(B2=0,"",C2/B2)
G2 =IF(B2=0,"",D2/B2)
H2 =IF(B2=0,"",E2/B2)
I2 =IF(D2=0,"",E2/D2)
These formulas work in Excel and Google Sheets. Display percentages to one decimal place and dollar ratios to two decimal places; retain unrounded values for the checks.
| Signup group | First-week activity retention | 30-day paid conversion | Revenue per entrant (USD) | Revenue per payer (USD) |
|---|---|---|---|---|
| Before change | 72 / 120 = 60.0% | 30 / 120 = 25.0% | 6,000 / 120 = 50.00 | 6,000 / 30 = 200.00 |
| After change | 56 / 80 = 70.0% | 24 / 80 = 30.0% | 5,280 / 80 = 66.00 | 5,280 / 24 = 220.00 |
| Combined | 128 / 200 = 64.0% | 54 / 200 = 27.0% | 11,280 / 200 = 56.40 | 11,280 / 54 = 208.89 |
The combined activity rate must divide the summed counts, rather than averaging the displayed percentages:
=SUM(C2:C3)/SUM(B2:B3)
Check each row through this identity:
Revenue per entrant = Paid conversion × Revenue per payer
Before the change, 25.0% × $200.00 = $50.00. After the change, 30.0% × $220.00 = $66.00. Total net revenue must tie to the billing summary: $6,000 + $5,280 = $11,280. Count unmatched account identifiers separately and resolve them before accepting that tie.
The after group has 10.0 percentage points more activity retention and 5.0 percentage points more paid conversion. Revenue per entrant rises by $16.00. These figures justify investigating the change, but different acquisition channels or customer needs could also explain them. An observational before-and-after comparison cannot establish causation.
The product owner checks the return event. The revenue operations owner checks membership and account mapping. Finance owns the billed revenue definition. For a broader management view, place these results alongside customer base KPIs. Use the cohort retention table reading guide when reviewing patterns across additional periods.
Covirage is not affiliated with Mixpanel, Microsoft, or Google.
Bring a prepared cohort activity export and a billing summary with consistent account identifiers and outcome windows. Covirage maps the columns, its tools compute and check the figures, and the external AI model explains the comparisons with every figure traced to its rows. Create a free Covirage account and bring a sample of your own data.
See Covirage for self-service analytics for the supporting workflow.
A cohort defines the people being studied. Retention measures whether those people return to perform a specified action after a starting event. Keep both definitions in the report: selecting a group alone does not establish what returning means.
It may accompany higher revenue, but the measures answer different questions. Activity counts qualifying users; revenue sums business transactions. Compare a consistent population and observation window, then examine paid conversion and revenue per payer before interpreting the relationship.
Yes, at a customer age both groups have completed. A recent group may have a finished first-week result but no finished later-week result. Compare the shared completed window and keep later cells blank until enough time has passed.
Agree on the reporting unit before combining activity and billing. If several users belong to one paying account, aggregate both files to that account using an approved mapping. Otherwise retain a user-level activity report and a separate customer-level financial report.