E-Commerce Analytics: KPIs Every Online Store Should Track

By Grocito16 October 2026

E-commerce analytics is the habit of measuring how visitors become customers and how customers become repeat buyers. Start with the funnel from sessions to purchase, then track a small set of KPIs: conversion rate, average order value, revenue per visitor, repeat purchase rate, acquisition cost, return rate and margin. Segment them by channel, device and customer type, and review them on a fixed routine.

Key takeaways

Begin with the questions you need answered, then pick the metrics. A long list of charts nobody acts on is worse than five numbers reviewed every week.
The funnel (sessions, product views, add to cart, checkout start, purchase) shows where shoppers leave, and each step has a different fix.
Averages hide outliers. Look at the spread, the median and segments before you celebrate or panic.
Revenue is not profit. Track margin and contribution margin per product and per channel.
Good tracking depends on clean events, a consistent data layer and honest handling of consent.

What questions should store analytics answer?

Dashboards fail when built around what a tool can show rather than what the owner must decide. Write the questions down first.

Which sources bring people who actually buy?
Which step of the journey loses the most shoppers?
Which products get views but no sales, or come back as returns?
Do first-time buyers return?
Are we making money after product cost, discounts, shipping, payment fees and advertising?
Are we running out of stock on things people want?

If a metric does not help answer a question or make a decision, leave it off the main dashboard.

What does the e-commerce funnel look like?

A funnel is a sequence of steps that most shoppers move through, with some leaving at each stage.

Sessions: visits to the store, from any source.
Product views: sessions that reached at least one product page.
Add to cart: sessions where something went into the basket.
Checkout start: sessions that began the checkout flow.
Purchase: sessions that completed an order.

Where drop-offs usually hide

Sessions to product views: weak landing pages, confusing navigation or traffic that was never interested, often a traffic quality problem.
Product views to add to cart: unclear pricing, thin descriptions, poor photos or doubts about delivery.
Cart to checkout start: surprise costs, a forced account, or an unclear next step.
Checkout start to purchase: long forms, limited payment options, failed payments or last-moment trust concerns.

Read the funnel by device too, since a healthy desktop funnel with a weak mobile one points to a mobile experience problem. Our guide to e-commerce conversion rate optimisation from product to checkout covers fixes for each stage.

Which KPIs matter most, and how are they calculated?

You do not need numbers to understand the formulas. Define each metric once, write it down, and use it the same way every time.

Conversion rate: orders divided by sessions for the same period. Always say which sessions you mean (all, or only those from a given channel).
Average order value: total order revenue divided by the number of orders. Decide whether it includes shipping and tax, and stay consistent.
Revenue per visitor: total revenue divided by visitors. It combines conversion rate and order value, so it suits comparing channels.
Repeat purchase rate: the share of customers who place a second order within a period you choose.
Customer acquisition cost: marketing and sales spend for a period divided by the number of new customers won. Read it beside lifetime value, as explained in our guide to customer acquisition cost and lifetime value.
Return rate: returned orders or units divided by orders or units sold. Break it down by product, because a few items usually cause most returns.
Cart abandonment rate: carts that did not become orders, divided by carts created. Pair it with abandoned cart recovery.
Gross margin and contribution margin: gross margin is revenue minus product cost. Contribution margin also subtracts the variable costs of the order, such as discounts, shipping, packaging, payment fees and attributed ad spend.
Inventory turnover: how many times you sell through your average stock in a period. Slow turnover ties up cash, while very fast turnover can signal stock-outs.
Stock-out rate: how often demanded products are unavailable. Lost sales from empty shelves rarely show up in a conversion chart, so track it separately.

How do you segment the numbers?

An overall conversion rate is a blend of very different groups. Segment before you draw conclusions.

Channel: organic search, paid search, social, email, direct, referral and marketplace traffic behave differently.
Device: mobile, desktop and tablet shoppers face different pages and different friction.
New versus returning: returning customers usually convert more easily, so a rise in returning traffic can flatter the overall rate.
Category: considered purchases and low-cost repeat items need different expectations.
Location: useful when delivery times or payment preferences differ across regions.

Keep segments few and meaningful, because tiny groups produce noise that looks like insight.

What do cohort and retention views show?

A cohort groups customers by when they first bought, for example by month, and follows what each group does next. It answers what the main dashboard cannot: do customers come back, and does a campaign bring long-term buyers or one-off bargain hunters?

Compare cohorts by first product bought, first channel and whether a discount was used. If customers who start with one category return more often, that is a reason to feature it or adjust which products you advertise first. For subscription or replenishment products, our article on subscription e-commerce and retention goes deeper.

How should tracking be set up?

Poor data makes every KPI untrustworthy, so treat tracking as a build task with testing, not a snippet pasted at the last minute. Our website analytics setup guide covers the basics for any site.

Events and the data layer

Define a standard list of e-commerce events: product view, add to cart, checkout start, payment step, purchase and refund. A data layer is a structured block of information on the page, such as product identifier, category, price and quantity, that your analytics tools read. Agree the naming once so every tool receives the same facts.

Test before you trust

Place test orders and confirm each is counted once, with the right value.
Check that cancelled and refunded orders are reflected, and that staff traffic is excluded.
Compare analytics revenue with your order system, and investigate large differences.

Consent and privacy

Where visitors decline cookies, your analytics will undercount, so treat your order system as the source of truth for revenue and orders. Rules differ by country and change over time, so ask a qualified adviser. Our guide to cookie consent and privacy policy is a good starting point.

What are the limits of attribution?

Attribution assigns credit for a sale to marketing touchpoints. It is helpful but never exact, because shoppers switch devices, decline cookies and see a brand in several places before buying.

Treat each model (last click, first click, data-driven) as a lens, not the truth.
Look at assisted conversions as well as final clicks.
Use consistent campaign tagging so sources are not lost in a vague direct bucket.
Consider asking buyers how they heard about you.

What should dashboards show for each role?

Owner: revenue, orders, conversion rate, average order value, contribution margin, repeat purchase rate and cash tied up in stock.
Marketing: conversion by channel, acquisition cost, revenue per visitor and email performance.
Operations: fulfilment time, stock levels, stock-outs, return rate and reasons, and payment failures.

A BI layer can pull these together from your store, ad platforms and inventory system. Our analytics and BI dashboard suits this joined-up view, and our data analytics services can help if your sources are scattered.

How often should you review the numbers?

Weekly: funnel steps, channel performance, top products, stock-outs and unusual swings, noting what changed that week.
Monthly: margin by product and channel, repeat purchase rate, return reasons, cohorts and acquisition cost.
Quarterly: pricing, range decisions and whether the dashboards still match your questions.

Always end a review with one to three actions and an owner for each.

Which metric traps should you avoid?

Averages hiding outliers: one large wholesale order can lift average order value for a month, so check the median.
Vanity traffic: more sessions mean little if the sources never buy.
Comparing unlike periods: festivals and sales distort week-on-week changes, so compare with the same period last year where you can.
Ignoring returns: revenue before returns overstates performance.
Optimising a ratio in isolation: lifting conversion rate with heavy discounts may reduce margin.
Changing definitions silently: mark any redefinition on the chart.

A worked example

Imagine a small store selling kitchen products that sees a steady rise in sessions but flat sales. The funnel shows plenty of product views and add-to-cart actions, but many shoppers leave at checkout start, mostly on mobile. Splitting by channel shows a new social campaign brings cheap visits with little buying intent, while email buyers convert well and return often. The owner shifts effort towards email, simplifies the mobile checkout form and adds a delivery estimate on the cart page, then watches checkout completion by device and contribution margin by channel rather than raw traffic.

FAQ

What are the most important KPIs for an online store?

Start with conversion rate, average order value, revenue per visitor, repeat purchase rate, customer acquisition cost, return rate and contribution margin. Together they show how well you attract buyers, how much each order is worth and whether the business is actually profitable.

How do I calculate e-commerce conversion rate?

Divide the number of completed orders by the number of sessions in the same period, and state which sessions you counted. For better insight, calculate it separately by channel and device rather than relying on one blended figure.

What is the difference between gross margin and contribution margin?

Gross margin is revenue minus the cost of the products sold. Contribution margin also subtracts variable costs of fulfilling the order, such as discounts, shipping, packaging, payment fees and attributed ad spend, which makes it a better guide to what each sale really earns.

How often should I review my store analytics?

A short weekly check of funnel, channels, top products and stock, plus a deeper monthly review of margin, retention and returns, works for most stores. Keep a fixed routine and finish each review with specific actions.

Next steps

List the five decisions you make most often in your store, then pick the metrics that inform each one. Check that your tracking records purchases correctly, build one simple dashboard per role, and book a weekly review slot. If you would like help with tracking, reporting or dashboards, contact us and we can review your current setup with you.