Shopify customer lifetime value analysis measures the value customers generate across repeated purchases. Cohort analytics groups customers by a shared starting event—usually first purchase month—and tracks revenue, contribution, repeat purchase and retention as the cohort matures.

The most reliable model starts with observed order and refund data, separates new from returning customers, calculates contribution rather than revenue alone, and shows forecast assumptions independently. Teams should compare cohorts by acquisition source, first product, market and offer only when sample size and data quality support the decision.

Project Supply builds ecommerce data and analytics systems for acquisition, retention and profitability decisions: Project Supply AI and Data Analytics

Why average LTV misleads

A single historical average mixes customers acquired in different periods, markets, channels and commercial conditions. Recent customers have had less time to repeat. High-value wholesale or employee orders can distort the mean. Refunds, discounts and fulfilment costs can make revenue-rich customers less profitable.

Cohorts solve the maturity problem by comparing customers at equivalent ages such as 30, 60, 90 or 180 days after first purchase. Segmentation then reveals which acquisition and product strategies create durable value.

Define the metric before calculating it

Revenue LTV

Revenue LTV is cumulative net revenue per acquired customer over an observed or forecast horizon. Define whether net revenue excludes discounts, refunds, taxes, duties and shipping.

Gross-profit LTV

Gross-profit LTV subtracts product cost from net revenue. It is more commercially useful than revenue when product margins vary, but still excludes other variable fulfilment and payment costs.

Contribution LTV

Contribution LTV subtracts the variable costs required to fulfil and serve orders according to a finance-approved definition. It provides a stronger basis for acquisition and retention investment.

Predictive LTV

Predictive LTV estimates future value using historical behaviour and assumptions. It must show model version, horizon, training population, uncertainty and validation. It should never be presented as observed performance.

The cohort model

Assign each eligible customer to a cohort based on first completed purchase date. Calculate customer age from that event. Aggregate orders and adjustments into standard maturity windows. Keep a stable customer key so email changes, guest checkout and merged profiles do not silently create duplicates.

Report both total cohort value and per-original-customer value. The denominator should normally remain the number of acquired customers in the cohort, allowing retention decline to appear in cumulative value.

Required Shopify data

Customer

Customer ID, first-order date, market, consent state and approved segmentation fields. Avoid using unnecessary personal data in analytical outputs.

Order

Order ID, customer ID, creation and processing timestamps, currency, channel, gross sales, discounts, refunds, tax, shipping, net sales and order status.

Order line

Product and variant IDs, quantity, item revenue, discount allocation, refund allocation and cost reference. Line data enables first-product and product-family cohorts.

Acquisition

First-touch and governed acquisition fields may include source, medium, campaign, click identifier, landing page and acquisition cost. Preserve the attribution model and lookback window.

Cost

Product cost, payment fee, pick-and-pack, packaging, variable shipping subsidy, return or RTO cost and other approved variable expenses. Store valid-from dates because costs change.

Data preparation

Normalise timezones, currency and identifiers. Define eligible orders, test orders, exchanges, cancellations and wholesale records. Link refunds to the original order and line where possible. Reconcile monthly totals against Shopify and finance reports before calculating cohorts.

Create explicit rules for guest checkout, duplicate profiles and customer deletion. Identity resolution should be proportionate and privacy reviewed; a higher match rate is not automatically better if it relies on inappropriate data use.

Core cohort outputs

Cumulative revenue per customer

For each cohort and maturity window, divide cumulative net revenue by original cohort customers. This shows how value builds over time.

Cumulative contribution per customer

Apply the approved contribution formula to every order and divide cumulative contribution by the original customer count. Use this for payback and acquisition decisions.

Repeat purchase rate

Calculate the share of the original cohort reaching a second, third or later completed purchase by each maturity point. State whether exchanges or subscription renewals count.

Time to second order

Measure the distribution, not only the average. Median and percentiles help schedule retention interventions without allowing a few very late purchases to distort timing.

Orders per customer

Track cumulative completed orders per original customer. Pair it with order value and margin so frequency improvement is not celebrated when contribution declines.

Active customer retention

For transactional commerce, define activity windows such as purchased within the last period. This differs from subscription retention and should not be labelled identically without explanation.

Segmentation dimensions

Useful cohort cuts include first acquisition source, campaign, landing page, first product, category, discount status, subscription status, geography, device, first-order value and fulfilment method. Begin with dimensions connected to a decision.

Avoid slicing until every cell appears interesting. Set minimum cohort sizes, combine sparse categories and display confidence. Differences in early small cohorts may be random.

Acquisition-source cohorts

Compare customers acquired from paid search, paid social, affiliates, creators, organic search, referrals, marketplaces and direct activity using a governed acquisition rule. Show spend and new-customer count alongside LTV.

Channel comparisons must control for maturity, product mix, offers and attribution differences. A campaign may acquire high-value customers because it promoted a high-margin product, not because the channel inherently produces better customers.

First-product cohorts

Group customers by the first product or category purchased. This can reveal gateway products that create repeat behaviour, one-time products with weak retention and bundles that accelerate cross-category adoption.

Evaluate first-order contribution, repeat contribution and stock or fulfilment constraints. Do not optimise acquisition toward an apparent gateway product that cannot support profitable scale.

Discount and promotion cohorts

Separate full-price, introductory discount, bundle and free-shipping cohorts. Compare conversion context, first-order contribution, repeat rate, later discount dependency and return behaviour.

Promotion decisions should consider incrementality. Customers who would have purchased anyway can make a discount cohort look valuable while eroding contribution.

Subscription and non-subscription cohorts

For subscriptions, measure activation, successful renewals, skips, pauses, churn, reactivation and contribution. Distinguish involuntary payment failure from voluntary cancellation. Compare with non-subscription customers using equivalent maturity windows.

Cohort maturity

Label recent cohorts as immature. Do not compare a 30-day-old cohort’s final-looking LTV with a 12-month cohort. Create fixed views such as 30-, 60-, 90-, 180- and 365-day observed contribution.

When reporting predictive values for younger cohorts, show observed value and forecast value separately. Back-test previous forecasts against actual maturation and publish error by segment.

CAC and payback

Calculate customer acquisition cost using acquisition spend divided by eligible new customers under the chosen attribution rule. Compare CAC with observed cumulative contribution LTV at matching cohort maturity.

Payback occurs when cumulative contribution attributable to the customer covers acquisition cost under the approved definition. Report the share of cohorts reaching payback and the timing distribution rather than one blended claim.

Need a Shopify cohort model reconciled with marketing spend and finance costs? Contact Project Supply: Talk to Project Supply

Dashboard design

Executive summary

Show new customers, CAC, 30/90/180-day observed contribution LTV, payback progress, repeat rate and forecast confidence. Include prior-period and plan comparisons.

Cohort heatmap

Rows represent acquisition cohorts and columns represent customer age. Cells show cumulative contribution per original customer, repeat rate or another single selected metric. Do not mix metric definitions in one heatmap.

Segment comparison

Allow comparison by channel, first product, market and promotion with sample size, maturity and acquisition spend visible.

Data-quality panel

Show unmatched spend, missing cost, unidentified customers, refund maturity, late-arriving orders, currency conversion and pipeline freshness.

Predictive modelling options

A simple model can use observed retention and order contribution curves by mature cohorts. More advanced models may estimate purchase probability and expected contribution by customer features. Complexity should be justified by improved, validated decisions.

Split training and validation periods, avoid leakage from future behaviour, monitor calibration and drift, and compare with simple baselines. Provide ranges rather than false precision. Do not use sensitive attributes without a legitimate, reviewed purpose.

Decision use cases

Use cohort evidence to set acquisition bids, choose first-order offers, prioritise retention journeys, forecast cash needs, plan inventory, evaluate subscription strategy, identify high-value products and decide which markets deserve investment.

Every dashboard view should connect to an owner and action. If a low-value cohort has no defined response, the analysis is descriptive rather than operational.

90-day implementation roadmap

Days 1–15: definitions

Agree customer, first order, revenue, contribution, acquisition, repeat, retention and maturity definitions. Inventory data and reconcile sample orders.

Days 16–30: foundation model

Build customer, order, line, refund, spend and cost tables. Produce unsegmented monthly cohorts and validate totals with finance and ecommerce teams.

Days 31–60: segmentation

Add channel, first-product, market and offer cuts. Create minimum sample rules and contribution-based payback views.

Days 61–90: operationalisation

Embed cohort reviews in budget and retention planning, back-test forecasts, document decisions and assign data-quality ownership.

Common mistakes

Avoid using revenue as profit, mixing cohort ages, ignoring refunds, changing customer identity rules silently, dividing by active rather than original customers, applying one product margin to all orders, adding predicted value to observed value, and comparing attributed channels without model context.

Another mistake is treating LTV:CAC as timeless. Both numerator and denominator change with cohort maturity, spend level, product mix and cost structure. Always display horizon and definition.

Governance

Finance owns contribution definitions and cost approval. Marketing owns campaign taxonomy and acquisition decisions. Ecommerce owns order operations. Data teams own pipelines and model versions. Privacy and security owners govern identifiers, access and retention.

Publish a metric dictionary with formulas, owners, source tables, exclusions, refresh frequency and effective dates. Log changes so historical reports remain explainable.

Commercial recommendation

Begin with observed contribution cohorts and reliable acquisition spend. Use predictive LTV only after mature cohorts exist and simple baselines are validated. Optimise for marginal contribution and cash payback, not the highest headline revenue LTV.

A useful Shopify cohort system makes uncertainty visible. It helps leaders distinguish a genuinely improving customer base from an attribution change, temporary promotion or immature cohort.

For Shopify CLV pipelines, cohort dashboards and retention analytics implementation, contact Project Supply: Talk to Project Supply