Product health dashboard: adoption, retention, and revenue

A product health dashboard compares adoption, retention, and revenue across consistent customer cohorts. Build it from the usage and customer revenue exports you already have, and keep the eligible population visible so that a changing customer mix cannot masquerade as healthier usage.

A useful dashboard lets product, customer success, and finance inspect the same accounts through different measures. A single blended health score can hide the reason those measures disagree.

Define the account and the observation window

Start with one business question: do accounts that receive guided onboarding adopt the main workflow, return to use it, and preserve their recurring revenue? Compare accounts that started in the same period and have completed the same observation window. Do not compare a newly signed cohort with one that has had several months to return.

Define adoption as completing a meaningful workflow during the first month. Define product retention here as an account returning to that workflow in the second month, divided by all original eligible accounts. This is a product activity measure, distinct from still having a paid subscription.

Choosing the start action and return action is fundamental to retention analysis, as Amplitude's retention documentation explains. Use your supplied export's existing actions and dates; the dashboard calculation begins after those records have been collected.

Prepare one row per account

Take an account list with account ID, start date, onboarding group, and eligibility. Bring the usage export with account ID, action, and date, plus opening and closing subscription balances. Reduce qualifying actions to account flags before joining revenue. Otherwise repeated usage records can multiply subscription amounts.

Use 1 for an account meeting an activity condition and 0 for a confirmed absence. Leave missing or incomplete evidence unresolved instead of turning it into zero. Keep the original account list, including churned and inactive accounts, as the reporting base.

The following illustrative US software company compares accounts through the end of their second month. Every account was eligible for the workflow. Columns A through F are account, cohort, adopted, returned, opening MRR, and closing MRR.

The adoption dashboard goes deeper into active usage and feature uptake; keep this dashboard focused on the connection to customer and revenue outcomes.

Calculate the cohort measures in Excel

Put the cohort name, such as Guided, in H2. With the source table in rows 2 through 9, use I2 for eligible account count, J2 for adoption, K2 for product retention, and L2 for net revenue retention:

I2: =COUNTIF($B$2:$B$9,H2)
J2: =SUMIFS($C$2:$C$9,$B$2:$B$9,H2)/I2
K2: =SUMIFS($D$2:$D$9,$B$2:$B$9,H2)/I2
L2: =SUMIFS($F$2:$F$9,$B$2:$B$9,H2)/SUMIFS($E$2:$E$9,$B$2:$B$9,H2)

Microsoft's SUMIFS documentation describes summing rows that meet a condition. Here it sums either binary activity flags or subscription balances. Format J through L as percentages, and show an unavailable result rather than a zero rate when its denominator is zero.

The revenue rate divides closing MRR from the original accounts by their opening MRR. Expansion can offset cancellation, so it can exceed full retention; Stripe's cohort definitions distinguish subscriber retention from revenue retention. Percentages here are shown to one decimal place.

Check the totals before assigning a cause

The cohort account counts must add to the source population: 4 + 4 = 8. Activity flags must never exceed those counts. The commercial check is 10,000 + 6,500 = $16,500 of closing MRR, matching the account rows, against $20,000 at opening.

The overall revenue retention rate is 16,500 / 20,000 = 82.5%. Revenue fell by 16,500 - 20,000 = -$3,500. Use an MRR dashboard to separate expansion, contraction, and churn rather than guessing from the net change.

Guided adoption exceeds Standard adoption by 75.0% - 50.0% = 25.0 percentage points. That is an observation, not proof that guided onboarding caused the difference. Account complexity, customer size, and selection into onboarding could also differ. Product owns the workflow definition, customer success checks account circumstances, and finance validates recurring revenue.

For a wider workflow review, use a product analysis dashboard; turn agreed findings and follow-up priorities into a product management report.

Where it goes wrong

• Different cohort ages. A cohort still completing its return window cannot be compared with a fully observed cohort.
• Event rows become customer counts. Repeated actions must not create extra adopted accounts or duplicate revenue.
• Paid retention is labeled product retention. An account can keep paying while usage stops, or return before a cancellation becomes effective.
• Missing records become inactivity. Verify export coverage before assigning zero activity flags.
• One score conceals opposing signals. Keep adoption, return activity, and revenue visible separately so the owner can see what changed.

Covirage is not affiliated with Microsoft, Amplitude, or Stripe.

Compare your own product cohorts in Covirage

Bring your usage export, account list, and customer revenue file to Covirage, which maps the columns and uses its tools to compute and check cohort comparisons. The external AI model explains the results with every figure traced to its rows. Create a free Covirage account and bring a sample of your own data after setting up your business.

See self-service analytics for the supporting approach to business files.