5 Ecommerce analytics metrics that actually reveal profitability and growth

Ecommerce businesses track dozens of metrics, and while most measure something real, only a handful reveal whether the business is profitable and growing. Traffic, conversion rate, average order value, and even revenue describe activity, and activity can rise while unit economics quietly decline.

That blind spot is where fast-growing ecommerce brands quietly go broke. A store can post record traffic, healthy conversion, and rising revenue for consecutive quarters while every incremental order shrinks its contribution margin. Catching it requires a metric set built around cost, margin, and repeat behavior, reviewed on the same cadence as revenue.

This article covers the three commonly tracked metrics that look strong but don't reveal profitability, the five that do, and the operating practices that turn those metrics into decisions.

What is ecommerce analytics?

Ecommerce analytics is the discipline of measuring an online store's commercial performance by combining customer, order, marketing, and cost data into a single analytical view. It ties shopper events to customer identity and revenue, then joins them to spend data from ad platforms, payment processors, and fulfillment systems. That join is what makes cohort retention, lifetime value, and channel-level contribution margin possible in the first place.

It's easy to conflate ecommerce analytics with web analytics, but the two are distinct. Web analytics covers on-site behavior such as sessions, pageviews, bounce rate, and user flows, and is relevant to any business with a website.

Ecommerce analytics is narrower and deeper: it connects those web activity events to who the buyer is, what they spent, what it cost to acquire them, and whether they came back. It also goes beyond performance reporting, which summarizes what happened. Analysis creates the most value when the business learns why results changed and what to do about it.

Why ecommerce analytics matters

Ecommerce analytics earns its keep by turning data that would otherwise sit in disconnected systems into decisions about where to spend, what to sell, and which customers to keep investing in. Any store can pull a revenue number. Fewer can tell you which channels, cohorts, and SKUs produced that revenue profitably, and answering that second question is where ecommerce analytics creates most of its value.

Specifically, ecommerce analytics gives a store:

• A unified view of the customer. Order data, marketing spend, and customer records live in separate systems by default. Ecommerce analytics joins them so acquisition cost, order history, and repeat behavior can be evaluated together.
• Unit economics that can be acted on. Contribution margin, CAC, and LTV get exposed at the order, SKU, cohort, and channel level, which are the units spending decisions actually get made against.
• Cohort and retention insight. Once orders are tied to identity over time, cohort analysis reveals whether newer customers are more or less valuable than older ones.
• Clarity on the acquisition-retention tradeoff. Combining CAC and LTV across cohorts shows whether the next marginal dollar is better spent acquiring new customers or reactivating existing ones.
• Channel-level profitability. Return on ad spend (ROAS) by channel is easy to calculate. Contribution margin by channel is what tells you which channels to scale, which to fix, and which to cut.

All of this value is gated on data plumbing, though. Data teams routinely lose the majority of their time to preparation work rather than analysis, and in ecommerce that work is largely the CAC-to-order join. Teams that solve it early get to spend their analytical time on decisions.

3 ecommerce analytics metrics that look strong but don't reveal profitability

Some of the most commonly tracked ecommerce metrics are also the most misleading. They describe activity rather than economics, and activity can rise for reasons unrelated to improving profitability. Sometimes it rises precisely as profitability deteriorates.

• Traffic and sessions. Traffic is the easiest metric to move because any team with a budget can buy more sessions. A record traffic quarter built on paid sessions only confirms that the media plan was executed. Profitability depends on what the business earned from those visits. Traffic only becomes meaningful when read alongside acquisition cost and the share of first-time visitors who become repeat customers.
• Conversion rate. Conversion rate looks like pure efficiency, but discounting routinely improves it, and every discount surrenders margin on every order it accelerates. Rising conversion alone doesn't prove economics improved. Revenue growth built on heavy acquisition and promotional spend can mask deep losses.
• Average order value. Free-shipping thresholds or bundles usually lift average order value by nudging shoppers to add one more item. The problem is that the added item may generate less gross profit than the shipping subsidy costs the business. Warby Parker's Q1 2023 captured the split: average revenue per customer rose 8.4% to $270 while gross margin fell from 58.5% to 55.1% and the company lost $10.8 million.

These metrics have their place as diagnostics for what's happening inside the funnel. The mistake is treating them as evidence of profitability when they only measure motion. The metrics that actually answer the profitability question sit one layer deeper, in the economics of the customer.

5 ecommerce analytics metrics that reveal profitability and growth

The five metrics below share a common property: they measure the economics of the customer relationship end to end, from acquisition cost through profit contribution to repeat behavior. When read together, they tell an operator whether the business is acquiring profitable customers, keeping them, and earning more from each cohort over time.

Customer acquisition cost

Customer acquisition cost (CAC) is the total spend required to acquire a new customer. It reveals whether growth is being bought at a sustainable price. Rising CAC without a corresponding rise in customer value means every new customer is compressing margin, while falling CAC on flat volume means the acquisition engine is getting more efficient. It's the denominator behind every profitability question that follows.

The pitfall to avoid is blended CAC. Pooling organic and paid customers together hides the true cost of paid acquisition, and averaging across channels hides the one channel that's dragging the rest down. Calculate paid and unpaid sources separately, break the number out by channel and cohort, and re-check it over time rather than assuming last quarter's efficiency will hold.

Contribution margin

Contribution margin (CM) is what an order keeps after every variable cost it incurs, and it's the real test of whether more volume improves the business or damages it. Gross margin stops at cost of goods sold (COGS) and misses the fulfillment, shipping, and processing costs ecommerce specifically incurs, which is why finance teams often work in layers:

• CM1 equals revenue less COGS.
• CM2 equals CM1 less fulfillment, shipping, and payment processing.
• CM3 equals CM2 less CAC.

The most common mistake at CM3 is double-counting: subtract CAC once, and don't also subtract the ad spend already inside it. And remember that the lower the contribution margin, the higher the return on ad spend (ROAS) required just to break even.

Customer lifetime value to CAC ratio

The LTV-to-CAC ratio compares the total value a customer generates over their lifetime to the cost of acquiring them, and it reveals whether acquisition spend is producing customers who justify the investment. The benchmark most teams reach for is 3:1, but the right ratio depends on the business's margin structure and repeat behavior. There's no universal target.

Teams typically fail here by using revenue instead of gross profit for LTV. A customer who generates $300 of lifetime revenue at a 50% gross margin is worth $150, so a $50 CAC yields 3:1. But the revenue version reports 6:1 and misleads every budget decision downstream. Always pair the ratio with a payback period (months until cumulative contribution profit ≥ CAC) so the risk of slow recovery stays visible.

Cohort retention

Cohort retention groups customers by their first-purchase period and tracks the share of each cohort that buys again in each subsequent period. It reveals whether customer quality is improving or decaying over time, and it exposes acquisition quality that blended metrics hide. A cohort acquired through deep discounts, for example, often retains worse than an organic cohort acquired the same month.

Don't trust the blended number. A 30% average can mask a January cohort at 45% and a March cohort at 15%, and holiday-acquired cohorts especially should be compared against organic cohorts before being credited with the same value.

Repeat purchase rate

Repeat purchase rate is the share of customers who order again within a defined measurement window. It shows whether the product and post-purchase experience are strong enough to earn a second order, a leading indicator of LTV and an early warning system for retention problems that might otherwise take a year to show up in cohort curves.

Results vary heavily by category, business model, and window length, so benchmarks are only useful within the same category. During peak holiday periods, separate repeat-buyer revenue from new-buyer revenue so seasonal one-time shoppers don't distort the underlying repeat trend.

How to use ecommerce analytics to guide operating decisions

You can act on the right metrics through three operating mechanisms: how the data is assembled, how it's segmented, and how often it's reviewed.

Join order, marketing, and customer data in one place before analysis

Land order, marketing, payment, fulfillment, and customer data platform (CDP) data in the same cloud data warehouse and model them into a single joined view. Until that view exists, CAC lives in one system, and the orders it produces live in another, and every downstream metric is either an estimate or an export. Build the joined view before starting the analysis.

Segment every metric by cohort and channel before trusting the total

Blended totals hide the decision. Blended CAC and retention totals can conceal an overpriced paid channel or an early-churning cohort. Compare channel ROAS by platform, but always weight it against product margin, and track contribution margin at the channel level so a larger channel with worse economics doesn't get credited as the winner.

Review margin and retention on the same cadence as revenue

Most teams review revenue weekly and margin quarterly, which guarantees that margin problems are caught only after they've compounded. Put contribution profit on the weekly business review next to revenue, and check it before raising budgets after a strong ROAS week, launching a promotion, or shifting spend between channels. Give margin the same decision-making weight revenue already has.

How Sigma supports ecommerce analytics

Sigma is the runtime layer to build and scale analytics, apps, and agents directly on live cloud data warehouse data, using a spreadsheet interface backed by warehouse governance. For ecommerce teams, that means the analyses that reveal profitability and growth, such as CAC by channel and cohort, CM3 by SKU, and retention curves by acquisition source, run against the same governed data model that finance and operations already use.

Once the governed data model is available, business analysts can build and update these analyses without routing every request through the data engineering backlog.

Live queries across order, marketing, and customer data in the warehouse

Sigma runs live against the warehouse, so an analyst working with order, ad spend, and customer tables in Databricks, Snowflake, BigQuery, or Amazon Redshift sees current numbers rather than last week's extract. Formulas, filters, and pivots compile to SQL and execute in the connected warehouse, with row-level security inherited at query time. You can analyze the joined view every metric in this article depends on directly, without pulling data into a separate system that immediately goes stale.

Spreadsheet-native cohort and margin analysis

A cohort table is a pivot: first-purchase month on rows, months since purchase on columns, and repeat buyers as the value. Contribution margin is a set of subtraction formulas across joined columns. In Sigma, a business analyst builds both with the pivot tables and calculations they already use, against billions of rows, while governance and lineage stay where the warehouse enforces them.

Input Tables for margin and forecast corrections

Margin analysis stalls when inputs are wrong, whether that's a third-party logistics (3PL) rate card that changed mid-quarter, a SKU mapped to the wrong COGS, or an under-reported channel spend figure. Input Tables let a finance or marketing analyst correct those assumptions and model forecast scenarios inside the same workbook that displays the margin view. New data writes to a separate schema without overwriting the original warehouse data, and changes are auditable.

Run ecommerce analytics with Sigma

Cost, margin, and retention belong at the center of any serious ecommerce analytics build, with traffic and conversion as diagnostics beneath them. Sigma runs that analysis live on the warehouse where your data already sits, so the analyst who gets the profit question can answer it without an export or a ticket. Get a demo to see cohort, margin, and LTV views built on your own warehouse data, or try Sigma free.