Customer Cohort Profitability

A subscription company compares customers who joined in January with those who joined in April. The April group has lower total profit partly because it has existed for fewer months, so a fair analysis compares both groups at the same month of customer life.

For an owner, cohort profitability answers whether newer customer groups create durable contribution as they mature, and it is strongest when it shows original group size, comparable age and transparent cost assumptions. Define the cohort: acquisition month is common, but a contract-renewal or product-adoption cohort can answer a different question, and membership rules should not change midway through the comparison.

Track revenue over time, including subscriptions, upgrades and refunds on a documented basis, and separate invoices, cash and recognised revenue when timing differs. Assign direct and attributable costs such as product delivery, support, payment fees and acquisition spending, without pretending broad fixed overhead is automatically caused by one cohort.

Stripe's cohort-analysis guidance explains how grouping customers by a shared starting period can reveal retention patterns hidden by company-wide averages, and profitability adds cost and revenue data to that cohort view. Corporate Finance Institute describes customer profitability analysis as examining the costs of serving customers rather than ranking them by sales alone, and the same cost discipline applies when customers are grouped into cohorts.

Use a common age axis, comparing month three for each cohort rather than present-day totals, because older groups naturally have more opportunities to earn revenue and incur costs. Watch survivor bias: if many customers leave early, looking only at the remaining customers can overstate profitability, so keep the original cohort size and show revenue and cost from the whole group.

An illustrative cumulative cohort contribution per acquired customer is cohort revenue through month six minus attributable costs through month six, divided by the original number acquired, so if 100 customers generate $50,000 of revenue and $30,000 of costs, contribution is $200 per original customer. Include acquisition cost explicitly when asking whether a channel pays back, because a cohort can show positive servicing contribution and still not cover its advertising and sales cost within the observed period.

Segment with care, since paid-search customers and referrals may have different acquisition costs or needs, and small groups are noisy, so present counts and avoid overinterpreting one month. Handle price changes by showing whether differences come from customer behaviour, product mix or pricing policy, and check delayed costs such as returns, warranty claims and service credits, because a too-short window can make a cohort seem profitable before those costs appear.

Do not confuse projected lifetime value with observed contribution, since a future-value model requires assumptions about retention and margin, and forecast portions should be labelled separately from actuals. Use a cohort matrix or chart with each row an acquisition month and each column a month since acquisition, marking incomplete periods clearly rather than treating them as zero revenue or cost.

Reconcile to finance totals so cohort revenue and costs tie to the relevant ledger scope with unassigned items explained, remembering that duplicate customer IDs can split one relationship across groups while merged accounts can hide the original acquisition source. If one cohort performs poorly, investigate onboarding, channel quality, product fit and service costs before cutting a group, because correlation alone does not identify the cause.