E-commerce analytics: the metrics that matter at each growth stage

An online store that tracks everything from day one usually learns nothing, because nobody has time to read forty charts. The better approach is to match what you measure to the decisions you face at your current size. This guide splits the path into stages by monthly revenue and lists the few metrics that matter at each, with formulas and small worked examples. The revenue bands are rough; use the stage that describes your problems, not only your turnover.

Stage 1: launch to about $50K a month

The question at this stage: do people who visit actually buy, and where do they come from?

Conversion rate

Conversion rate = orders / sessions

With 20,000 sessions and 400 orders in a month, conversion is 2%. Watch it by device and by traffic source. A store that converts at 3% on desktop and 0.8% on mobile has a mobile checkout problem, not a traffic problem. The post on e-commerce funnel analysis shows how to find the step where visitors drop out.

Traffic by source

Sessions and orders by channel: organic search, paid, email, social, direct, referral. At launch the useful question is which one or two channels produce buyers, not which produces clicks.

Average order value (AOV)

AOV = revenue / orders

$30,000 from 400 orders is an AOV of $75. AOV tells you how much room you have to pay for traffic. It is also the easiest number to move early, through bundles, free-shipping thresholds and sensible add-ons.

What to skip for now

Lifetime value models, attribution models and detailed cohort tables. With a few hundred customers they are mostly noise. Keep the raw data clean so you can build them later.

Stage 2: about $50K to $200K a month

The question now: can you buy customers profitably, and do they come back?

Customer acquisition cost (CAC)

CAC = sales and marketing spend / new customers

$20,000 of spend and 500 new customers is a CAC of $40. Split it by channel as soon as you can, because a blended figure can hide one channel that never pays back. The post on customer acquisition cost beyond the basic formula covers blended against paid CAC and what to include.

Return on ad spend (ROAS)

ROAS = revenue attributed to the ads / ad spend

A ROAS of 3 means $3 of revenue for each $1 spent. Revenue is not profit, though: with a 40% gross margin, a ROAS of 3 leaves $1.20 of gross profit per $1 spent, which is barely positive once shipping and payment fees come out. Know the break-even ROAS for your margin: 1 / gross margin, so 2.5 at 40%.

Repeat purchase rate

Repeat purchase rate = customers with two or more orders / all customers (in the same window)

If 1,000 customers bought in the last 12 months and 250 bought more than once, the rate is 25%. Repeat buyers do not need to be acquired again, so this number decides how much you can afford to pay for the first order.

Contribution margin

Contribution margin per order = revenue - cost of goods - shipping - payment fees - discounts - marketing cost for the order

Say an order is $75. Goods cost $30, shipping $8, payment fees $2.50, a discount $5, and the ad cost per order $20. Contribution is $75 - $65.50 = $9.50, or about 13%. Plenty of stores grow revenue at this stage while contribution per order quietly falls. Track both.

Stage 3: about $200K to $1M a month

The question now: which customers and products actually make money?

Customer lifetime value (CLV)

For repeat-purchase businesses, a simple version:

CLV = AOV × orders per customer per year × gross margin × expected years as a customer

$75 × 2 orders × 40% margin × 3 years = $180 of gross profit per customer.

CLV to CAC

With a CLV of $180 and a CAC of $40, the ratio is 4.5. The usual guide is to stay above 3. Below 1, each customer costs more than they return. Work it out by channel: a channel with a $90 CAC may look fine on first-order ROAS and still lose money over the customer's life, or the reverse.

RFM segmentation

Score customers by recency (days since last order), frequency (number of orders) and monetary value (total spent), usually 1 to 5 on each. It turns a customer list into groups you can act on: recent high spenders for early access, lapsed frequent buyers for a win-back flow, one-time bargain buyers left out of full-price campaigns. The customer segmentation guide has the scoring steps.

Product-level economics

Margin, return rate and repeat rate by product. A best seller with a 25% return rate may earn less than a quiet product with 2% returns. Some products are good first purchases that lead to repeat orders even at a thin margin; others attract one-time buyers only.

Stage 4: about $1M to $5M a month

The question now: where is money leaking at scale?

• Inventory. Weeks of cover by product (stock / average weekly units sold), sell-through rate, and stock-outs on your best sellers. Running out of a top product for two weeks costs more than most marketing tests will ever win.
• Blended and marginal CAC. The next $10,000 of spend usually costs more per customer than the last $10,000. Track the cost of the extra customers, not only the average.
• Operations. Fulfilment cost per order, days to ship, support tickets per 100 orders.
• Attribution. Platforms each claim credit for the same order. Compare what the ad platforms report with your own order data, and use holdout or geography tests where you can.

Stage 5: beyond $5M a month

At this size you typically add a data team and the work shifts to predictive lifetime value, incrementality testing for each major channel, price elasticity studies and supply chain forecasting. Most small and mid-size stores do not need these yet, and doing them badly is worse than not doing them.

Mistakes at every stage

• Revenue without margin. Growth with falling contribution is not progress.
• Ignoring returns. In some categories returns take a large share of gross revenue. Report net revenue after returns.
• Only acquisition. Repeat customers are usually cheaper to keep than new ones are to find. Give retention the same attention.
• Vanity metrics. Followers and total visits are not revenue. Keep them off the main dashboard.
• Wrong comparisons. Comparing December with January tells you about the holidays. Compare with the same period last year.

Where clariBI fits

clariBI puts store, ad and email numbers side by side so the formulas above can use real figures:

• Sales: Stripe, Square and PayPal through the MCP catalog, or order exports from your store as CSV or Excel. For Shopify, the catalog's Storefront entry reads the product catalog only, not orders or customers, so sales come in through an export or a sheet your store app keeps up to date.
• Traffic: Google Analytics 4, including its e-commerce report (revenue, transactions, purchasers), and Search Console. The help article on connecting GA4 lists the reports it syncs.
• Ads and email: Google Ads and Meta Ads, Klaviyo, Mailchimp and Brevo.

With those connected you can ask "What was revenue by week, next to ad spend?" or "Which campaigns returned the most revenue per dollar last month?". The answers are computed with code over your synced rows. On Starter and above, a weekly revenue forecast with 50%, 80% and 95% ranges helps plan stock and spend. There are no inventory or warehouse connectors, so stock levels come from an export or a sheet. The e-commerce use case lists what to connect first.

Next step

Work out your contribution margin per order for last month using the formula above, with real numbers for shipping, fees and discounts. If it is below 10%, fix that before you raise ad spend.