Cohort Retention by SmartVisuals

Bind an entity and an event date. That is the whole setup.

Cohort Retention by SmartVisuals derives the first-event date, the cohort, the period offset and the denominator for you, and it tells you which denominator it used. The five-part DAX recipe that cohort analysis usually needs - a first-event calculated column, a cohort key, a disconnected period table, an offset measure and a future-period guard - becomes unnecessary.

IT DOES NOT LIE

A cohort table is easy to get wrong in ways that look right. This one is built around not doing that, and every one of these is free in every tier:

• Unobserved cells are hatched, never rendered as zero. A cohort that has not lived twelve months yet does not report 0 percent at month twelve.
• Incomplete periods are outlined and excluded from averages by default, so a half-finished month cannot pass for a finished one.
• Cohorts with too few entities show their base count instead of a meaningless percentage.
• The denominator is named on screen, in a status strip that survives export to PDF and PowerPoint.

FOUR DENOMINATORS, THREE DEFINITIONS, THREE VIEWS

• Cohort base, observed base, previous period, surviving base. Classic, rolling and sequential retention, each named on the chart.
• Calendar or anniversary period alignment. Day, week, month, quarter or fiscal-year grain. Cohorts ordered by date, size or the metric at any period.
• The retention triangle, survival curves per cohort with an optional reference line, and a layer cake of active entities over calendar time.

BUILT FOR REPORTS, NOT DEMOS

Click a cohort, a period or a cell to filter the rest of the report - with a Cohort Date bound, one filter covers a cohort of any size and survives bookmarks. Format on the canvas (on-object formatting), conditional formatting for cohort labels, dynamic format strings, full keyboard navigation with a single tab stop, high contrast, and 20 languages. Works on pre-aggregated data when the model is too large to send entity-level rows.

PREMIUM: VALUE AND COHORT COMPARISON

The free tier tells you how many came back. Premium tells you what they are worth and what to do about it.

• Follow the money: value retention (net revenue retention), revenue per active entity, and the cumulative lifetime value curve per cohort, with the period each cohort repays your acquisition cost.
• Know how long customers stay: pooled Kaplan-Meier survival over every cohort, with the median survival time (T50) and its confidence interval.
• Ask what if: cut churn by any percentage in the periods you choose and see the projected retention next to the actual, with the gain per period.
• Compare what matters: segments side by side on one colour scale, with significance markers that separate a real gap from noise.
• Judge against a standard: a benchmark row from your best, median or first cohort - or a target you type in - plus threshold flags on every cell above or below the level you set.
• Read the cause, not just the number: the age-vs-calendar panel, the insights strip, and retention projection with a shifted-beta-geometric model validated against the published reference.

Every premium feature is fully unlocked while editing, so you can evaluate everything before buying. Unlicensed viewers always see the full free tier. Never blocked, never watermarked. No data leaves Power BI: the visual makes no external calls.

Sample data and documentation: smartvisuals.io
Support: hello@smartvisuals.io