Six E-Commerce KPIs That Show If Growth Is Profitable

Higher e-commerce sales do not necessarily mean healthier growth. Advertising costs, discounts, fulfillment expenses and returns can increase faster than revenue, so a useful scorecard must connect demand and checkout performance to customer economics and contribution margin.

The core formulas remain useful, but measurement has become more precise: platforms distinguish among gross sales, purchase revenue, refunds, sessions and event-based purchases in ways that can produce different results from the same trading period. The six KPIs that belong on the main scorecard are purchase conversion rate, checkout completion rate, average order value, customer acquisition cost, cohort customer value and contribution margin.

Set the measurement boundary first

A KPI is comparable only when its definition stays stable. For each measure, document the numerator, denominator, reporting period, currency treatment, order statuses and exclusions. A change in bot filtering, attribution logic or refund handling can move a metric even when customer behavior has not changed.

Platform definitions demonstrate the problem. Shopify’s analytics field reference calculates average order value as gross sales minus discounts, divided by orders, and excludes post-order adjustments. By contrast, Google Analytics’ official metric schema defines purchase revenue after refunded transaction revenue, while gross item revenue excludes tax, shipping and refunds.

Neither convention is universally wrong; they answer different questions. The management scorecard should use definitions that reconcile with the company’s order and financial data, while platform-native figures can remain useful for diagnosing traffic and merchandising performance.

The six-KPI scorecard

1. Purchase conversion rate

Purchase conversion rate measures completed purchase sessions as a share of eligible store sessions. The denominator needs a consistent scope: known bots, staff visits, non-store landing pages and markets where purchasing is unavailable should not appear in one period and disappear in the next.

The blended rate is an alert, not an explanation. Break it down by device, market, acquisition channel and new or returning visitor status. A decline caused by a larger share of low-intent traffic requires a different response from a decline confined to returning customers on mobile devices.

Conversion should also be tied to completed, valid orders. Counting a button click or payment attempt as a purchase can hide payment failures and inflate the apparent performance of the store.

2. Checkout completion rate

Checkout completion rate is completed purchases divided by checkout starts. Its inverse is checkout abandonment. Cart abandonment begins earlier in the journey, usually with an add-to-cart action, so it uses a different denominator and should not be presented as the same KPI.

Baymard’s cart-abandonment compilation, last updated September 22, 2025, calculates a 70.22% average across 50 studies. That figure is useful as broad context, not as a universal target: the component studies cover different years, stores and measurement environments.

A store’s own funnel is the more actionable benchmark. Separating cart additions, checkout starts, payment attempts and confirmed purchases shows where customers leave. Delivery restrictions, payment failures and unexpected costs may all depress the final rate, but they occur at different stages and require different investigation.

3. Average order value

Average order value divides the chosen order-revenue measure by completed orders. The phrase “chosen order-revenue measure” matters: gross merchandise value, net sales and revenue collected at checkout can diverge after discounts, cancellations and returns.

AOV should therefore be read beside units per order, discount rate and contribution per order. A bundle can increase the checkout total while lowering the amount retained after product and fulfillment costs. Likewise, a free-shipping threshold can produce a larger basket but transfer additional delivery expense to the merchant.

The objective is not the highest possible AOV. It is an order value that improves contribution without relying on discounts or subsidies that cost more than the incremental gross profit they create.

4. Customer acquisition cost

Customer acquisition cost divides acquisition spending by newly acquired customers over a defined period. The calculation must state which costs are included. A media-only CAC is suitable for comparing advertising campaigns, while a fully loaded figure may also include affiliate commissions, agency fees and variable creative costs.

Timing is another boundary. Customers may purchase after the period in which the acquisition expense was recorded, especially for products with a longer consideration cycle. A consistent attribution window reduces the distortion created by comparing one month’s spend with a different set of customers.

Blended CAC is useful for financial planning, while channel-level CAC helps allocate budget. The two should not be treated as interchangeable because unpaid demand, repeat exposure and attribution rules can shift costs among channels without changing total acquisition spending.

5. Cohort customer value

Cohort customer value tracks the economic value generated by customers acquired in the same period. Comparing cohorts over the same fixed window avoids giving older customers more time to buy than newer ones.

Revenue-based lifetime value can exaggerate the amount available for acquisition. A decision-grade version uses gross profit or contribution after relevant product, fulfillment, payment, returns and service costs. It should also use the same currency, customer definition and time horizon as the CAC comparison.

Repeat-purchase rate provides useful context but cannot replace the monetary measure. A cohort may contain many returning customers whose small or heavily discounted orders contribute less than a cohort with fewer, more profitable repeat purchases.

6. Contribution margin

Contribution margin tests whether orders leave enough money to support fixed costs. Begin with net sales and subtract costs that vary with the order, which may include cost of goods, payment fees, picking and packing, merchant-funded shipping, expected returns and variable promotional expense.

The cost boundary depends on the operating model, so it must be explicit. If acquisition spending is excluded from order contribution, CAC should be deducted separately before assessing the economics of a new customer. Otherwise, the scorecard can make an apparently profitable order look more attractive than the full customer economics justify.

Track both contribution dollars and contribution margin percentage. Dollars show the amount available for payroll, rent, software and other fixed expenses; the percentage supports comparisons across periods, products and markets with different sales volumes.

Benchmarks require comparable conditions

External averages can provide context, but they are weak targets unless the comparison group resembles the store. Category, price point, geography, device mix, traffic source, seasonality and treatment of refunds can all alter the result.

An internal benchmark should compare equivalent periods and segments using unchanged definitions. A promotion should be compared with a similar promotion as well as the preceding period; mobile paid traffic should be compared with the same segment, not with the store-wide average. Median performance and normal variation are more informative than a single record day.

Benchmarks also need business context. A conversion rate below an industry average may still support profitable growth if acquisition costs and contribution are strong. Conversely, an impressive conversion rate can conceal excessive discounting, low-value customers or expensive returns.

Why the six metrics must reconcile

The scorecard works as a connected system. Conversion produces orders, checkout completion isolates late-stage losses, AOV measures revenue per order, CAC prices customer growth, cohort value shows what acquired customers return, and contribution margin determines how much of that activity supports the business.

The figures should reconcile with one another. Orders multiplied by AOV should broadly explain the corresponding sales measure; acquisition spending divided by new customers should reproduce CAC; and cohort contribution should eventually connect customer behavior with the margin recorded in order data. Material gaps usually indicate differences in timing, attribution, currency or order status.

Consider a conditional example: if conversion and AOV rise while contribution per order falls, larger baskets are not automatically evidence of better performance. Product mix, discounts, shipping subsidies and expected returns deserve attention before additional traffic is purchased. The purpose of the six-KPI framework is to distinguish sales growth from growth that creates durable economic value.

Six E-Commerce KPIs That Show Whether Sales Growth Is Profitable