Tableau. Repeat-purchase cohorts. See when paid customers come back for a second order.

What this means for your store

A cohort groups customers by their first purchase period - often month - and tracks what share order again in period +1, +2, and so on. Paid media can look profitable on first-order ROAS while destroying LTV if cohorts never repeat. Tableau builds the triangular retention table with LOD for first order date and DATEDIFF for months since acquisition.

Scenario on a real storefront

Chewy's retention team cohorts pet owners by first order month and first-touch channel. The paid vs organic split tells merchandising whether prospecting buys durable customers:

// First order date per customer
{ FIXED [Customer ID] : MIN([Order Date]) }

// Cohort month
DATETRUNC('month', [First Order Date])

// Months since first order
DATEDIFF('month', [First Order Date], [Order Date])

// Repeat flag (order after first)
[Order Date] > [First Order Date]

// Viz: Cohort month on rows, Months Since on columns
// Metric: COUNTD(Customer ID) / LOOKUP for % of cohort
// Filter: exclude wholesale B2B customers

What to do next

• Cohort charts are sensitive to returns and cancellations - align with net revenue policy.
• Small cohorts (< 100 customers) swing wildly; suppress or blend quarters.
• Compare paid vs organic cohorts separately - blended CAC distorts whether a campaign rents or retains customers.

Bottom line

Repeat-purchase cohorts show whether ecommerce growth is rented or durable. Use LOD for first-order timing, cohort by month and channel, and read them alongside MER and payback windows.