Lifetimely: Cohorts and Cohort Breakdown

Cohorts and Cohort Breakdown

At a glance

Cohorts groups customers by when they placed their first order and follows their purchasing value at equal ages. Cohort Breakdown groups customers by a selected product, promotion, geography, marketing, integration, or survey value. Use the reports together to identify whether customer quality is changing over time and which attributes are associated with the difference.

Availability and data freshness

Availability and prerequisites: Both reports are routed in Shopify and Amazon store contexts. Amazon hides the Shopify-only Klaviyo, Fairing, and KnoCommerce breakdowns. Survey breakdowns require the corresponding connected integration and a selected question. Predictive cards require the predictive-data account feature; Segment by appears only for accounts with cohort segmentation enabled.

Freshness and history: The Cohorts table response is treated as fresh in the page for 10 minutes; Cohort Breakdown and LTV-average responses are treated as fresh for one minute. These browser cache periods are not source-import SLAs.

Use Cohorts

Open Lifetime value > Cohorts (/lifetime_value/cohorts). The page title is Lifetime Value Report.

With no saved report range, the grid starts six months before today and ends yesterday, using store-timezone day boundaries. A previously saved report range overrides this fallback. The separate LTV-average request has a 90-day fallback; changing Time period aligns the grid and average cards to the selected dates.

Controls available on this page

Time period: Choose custom dates or Last 6 months, Last 12 months, This year, or a Since 2020 through Since 2025 preset. Presets end yesterday.
Timeframe: Group first orders and subsequent activity by Day, Week, Month, Quarter, or Year. The default is Month. Changing this changes cohort boundaries and the meaning of each age column, not merely the label.
Metric: The default is Accumulated sales per customer. See Cohorts metrics and calculations for all choices.
Format: Switch between currency/value ($) and percentage (%). Value is the default. Percentage format is unavailable for Customers who purchased X times.
Decimals (.00): Show two decimal places for currency metrics. It is disabled for percentage format and non-currency metrics.
Filters: Complex report filters narrow the report within the store-wide population. Store prefilters, calculation settings, timezone, display currency, platform, and saved date range are applied automatically from store context; they are not separate controls on this page.

Grid, chart, prediction, and export features

Grid: Each acquisition cohort is a row. Darker purple cells represent higher values within the displayed metric's range; hover a cell for its cohort, age, value, and first-purchase-relative percentage. The Average footer is an API-provided age summary; the New customers, CAC, and R-% footer values are simple averages across displayed cohort rows, not totals.
Grid/line-chart switch: The chart plots the selected metric by cohort age. Predictive data, CAC payback, Blended CAC, and cohort segmentation are unavailable in chart view.
Predictive data: Adds forecast cells for accumulated sales per customer, cohort sales, gross or contribution margin per customer, and the three CAC ratios. It is unavailable for other metrics and in chart view. Predicted cells are visually striped.
CAC payback: Adds a green marker at the age where accumulated gross margin per customer exceeds CAC. This is acquisition-cost payback, not company net profitability; it does not subtract operating expenses.
Blended CAC: Appears only after a complex filter is applied. Filtered CAC and CAC-dependent metrics are normally suppressed because marketing spend cannot be attributed to the narrowed cohort. Turning this on instead uses each period's whole-store blended CAC: total marketing spend divided by all new customers, not only filtered customers. The same basis is also used for filtered contribution margin.
Segment by: When enabled for the account, expand cohort rows into Upsell: Purchased vs Not Purchased or Upsell: Viewed vs Not Viewed segments. This changes the request and is not a general product/customer segmentation control.
Historical and Predictive LTV Averages: Separate cards show 3-, 6-, 12-, and 24-month values and a comparison for accumulated sales, net sales, or gross margin per customer. Other metrics show N/A. Prediction cards are account-gated, are disabled by filters, and can show N/A when the selected period is too short.
Export: Downloads a CSV for the active dates, timeframe, metric, format, predictive setting, and optional cohort segmentation. The file also includes New customers, CAC, total sales, returning customers, repurchase rate, total sales per customer, and an averages row. Export is disabled for Customers who purchased X times.
Ask AI: If available, the page action receives the selected dates, timeframe, metric, format, predictive state, and report data. Grid-cell tooltips can also offer an inline AI action.

Read the Cohorts grid

The selected range identifies customers whose first qualifying order belongs to each acquisition period after store-wide prefilters and complex filters are applied. One row is one acquisition period; New customers is the original cohort size used by per-customer lifetime metrics.

The columns distinguish the first order from later activity:

First order contains first-purchase value.
0 contains activity in the remainder of the same day, week, month, quarter, or year as the first purchase. For an accumulated metric, the cell is the running value through that age; for a discrete metric, it is additional same-period activity.
1, 2, and later columns move through successive cohort-age periods.

Newer cohorts have blank right-hand cells because they have not reached those ages. Compare the same numbered column across rows rather than each row's latest populated value.

The fixed columns provide additional context:

Field	Definition
First order at	Store-timezone day, week, month, quarter, or year containing customers' first qualifying order.
New customers	Customers acquired into the row's original cohort. This is the denominator for per-customer accumulated metrics.
CAC	Front-loaded CAC: the cohort period's total marketing spend divided by new customers. It is suppressed under complex filters unless Blended CAC is enabled.
R-%	Percentage of cohort customers who made a second purchase.

Worked example

A monthly cohort has 100 new customers. Their first orders produce $8,000 net sales, and later orders bring accumulated net sales to $12,000 by month 3. Accumulated net sales per customer at month 3 is $12,000 / 100 = $120, even if only 30 customers returned. Dividing by 30 would measure returning-customer value instead. Compare this cohort's $120 with other cohorts in their month-3 column.

Cohorts metrics and calculations

Unless noted as discrete, “accumulated” means the running value from first purchase through the displayed cohort age. Currency values use the store's selected display currency and shop calculation profile.

Metric	Unit and calculation
Accumulated sales per customer	Currency. Tax-inclusive order totals, which already reflect discounts, minus refunds; accumulated through that age / original cohort customers.
Accumulated net sales per customer	Currency. `(accumulated sales − taxes)` through that age / original cohort customers. Discounts and refunds are already reflected in accumulated sales and are not deducted again.
Accumulated gross margin per customer	Currency. `(accumulated net sales − product cost − shipping cost − handling cost − transaction cost − applicable duties and Amazon fees)` through that age / original cohort customers.
Accumulated contribution margin per customer	Currency. `accumulated gross margin per customer − front-loaded CAC`. The same cohort CAC is deducted once at every displayed age; discounts and refunds are already reflected upstream and are not deducted again. It is hidden for filtered cohorts unless Blended CAC is enabled, which substitutes whole-store CAC for the cohort's filtered CAC.
Average order value	Currency, not accumulated. Sales after discount codes and before refunds / orders in the first-order or age bucket.
Customers	Count, not accumulated. Unique customers from the original cohort who purchase in the displayed age bucket; the First order cell represents the acquired customers.
Cohort transactions	Count, not accumulated. Orders from the cohort in the displayed first-order or age bucket.
Cohort sales	Currency, not accumulated. Sales from the cohort in the displayed first-order or age bucket.
Cohort net sales	Currency, not accumulated. Net sales after taxes, discount codes, and refunds in the displayed first-order or age bucket.
Accumulated orders per customer	Ratio. Orders through that age / original cohort customers.
Customers who purchased X times	Count distribution rather than an age grid. Columns are customers with 1 or more, 2 or more, 3 or more, through 8 or more orders. These thresholds overlap: a customer with three orders appears in the first three columns.
Accumulated sales per customer / CAC	Ratio. Accumulated sales per customer / CAC. A value above 1 means sales have exceeded acquisition cost, not that margin has recovered it.
Accumulated net sales per customer / CAC	Ratio. Accumulated net sales per customer / CAC.
Accumulated gross margin per customer / CAC	Ratio. Accumulated gross margin per customer / CAC. A value above 1 marks gross-margin CAC payback in this report.

Missing cost or marketing setup can overstate margin or suppress CAC-dependent values. Confirm cost source precedence, refund treatment, fallbacks, and ad-account exclusions in the shop calculation profile.

Use Cohort Breakdown

Open Lifetime value > Cohort Breakdown (/lifetime_value/cohort_breakdown). With no saved range, the report starts 90 days before today and ends yesterday; a saved report range overrides it. Defaults are Month, Product, Accumulated sales per customer, page 1, and 10 rows per page.

Choose and compare breakdowns

Time period, Timeframe, Filters, and automatically applied store settings affect the population in the same way described for Cohorts. Cohort Breakdown has the same presets but does not offer percentage format, decimal display, prediction, CAC payback, a chart, or LTV-average cards.
Breakdown by: Product, Product type, Product sku, Product variant, Discount code, Order tag, Customer tag, Country, State, Vendor, first- or last-touch Source, Medium, or Marketing channel, Klaviyo list or segment, Fairing Survey Question, or KnoCommerce Survey Question. Amazon hides Klaviyo and survey options. When one survey provider is connected, unconnected competing survey providers are hidden.
Survey question: Fairing and KnoCommerce require a question selection. Rows then represent answers to that question and include respondents only. Export remains disabled until a question is selected.
Metrics: Accumulated sales per customer, Accumulated gross margin per customer, Average order value, Customers, Cohort transactions, Cohort sales, or Accumulated orders per customer.
Search: In normal mode, search the current dimension's row values; survey mode searches answers. Results are paginated 10 at a time.
Compare: Replaces search and pagination with a multi-select comparison of exact dimension values. It is unavailable for Product variant, Fairing, and KnoCommerce survey breakdowns. Selections are remembered separately while switching breakdown dimensions during the page session.
Export: Starts a background CSV for the active dates, timeframe, breakdown, metric, search, page settings, and filters. Retrieve the finished file from Settings > Downloads.
Ask AI: If available, receives the selected dates, timeframe, breakdown, metric, comparison state, and loaded report data.

Read the breakdown table

One row is one returned dimension value, with New customers and R-% followed by First Order, age 0, 1, and later cells. Cell shading is relative to the displayed values. Rows can overlap—for example, one customer or order can have multiple products or tags—so do not sum breakdown rows to produce the all-customer cohort.

The current UI does not expose whether each dimension uses a first-order value, any qualifying order, current customer/product attributes, or event-time attributes. First- versus last-touch labels establish attribution position but not the provider's full attribution method.

For survey breakdowns, nonrespondents are omitted rather than placed in an unanswered row. If 40 respondents answered “Podcast” and accumulated sales through month 6 are $8,000, displayed sales per customer is $8,000 / 40 = $200. Compare answer groups only after checking respondent counts; do not treat them as representative of every customer without validating response bias.

Check and reconcile a result

Confirm store, Time period, Timeframe, metric, currency, timezone, and the count on Filters.
Check prefilters. Because they run before first-order classification, excluding an early order can move a customer into a later acquisition cohort.
For a value metric, hover the cell and identify whether you are reading First Order, age 0, or a later age. Recalculate accumulated metrics using the original New customers denominator, not returning customers.
For CAC or margin, check marketing connections, product/shipping/handling/transaction costs, refunds, and source precedence in the shop calculation profile. Confirm whether filtered values are suppressed or use Blended CAC.
Export Cohorts to inspect the row's customer, CAC, sales, returning-customer, and age columns. For Cohort Breakdown, retrieve the generated file from Settings > Downloads and verify whether its scope is suitable before reconciling.
Compare only equal-age cells. If checking a breakdown, remember that overlapping or respondent-only rows need not sum to Cohorts.

Resolve differences and limitations

Cohorts and Cohort Breakdown do not match: First confirm matching dates, timeframe, metric, filters, prefilters, currency, and cost settings. Breakdown rows can overlap or omit blank/nonresponding customers, and their dimension anchors are not fully documented, so their sum is not an all-customer control total.
A newer row looks worse: Its later-age cells are not complete. Compare at the newest age populated for every cohort in the comparison.
CAC or contribution margin is blank after filtering: Remove the complex filters or enable Blended CAC and accept that it uses whole-store CAC, not segment-attributed spend.
Margin looks too high: Missing product, shipping, handling, transaction, or marketing costs can bias margin upward. Verify the applicable cost source and fallback.
Prediction or averages show N/A: Use a supported metric and longer date range and confirm predictive access. The minimum data requirements are not published.
Historical values changed: Prefilter, refund, cost, order, and customer-identity changes can alter cohort membership or economics.

Next actions

Use LTV Drivers to form hypotheses about first-order attributes, then validate them at equal ages in Cohort Breakdown.
Use Subscriptions when the question is provider-based subscriber or recurring-revenue retention rather than commerce-order cohorts.
Review prefiltering and shop calculation profiles before treating a cohort difference as an acquisition or retention change.