Ecommerce KPIs and analytics: the numbers actually worth checking every week

It's easy to fixate on a single number: how much we sold this month. It's the figure at the top of the dashboard, the one mentioned in every meeting, the one that decides whether the month "went well" or "went badly" in casual conversation. The problem is that figure, on its own, says nothing about why what happened happened, and without that information it's impossible to know which lever to pull to make next month better. An ecommerce store automatically generates dozens of useful data points; the difference between a business that improves steadily and one that lurches around usually comes down to which part of that data gets looked at regularly, and what gets done with what's seen.

The metrics that matter more than total sales

• Conversion rate. What percentage of visits ends in a purchase. It's the most direct gauge of whether the store is doing its job of convincing people who've already arrived, regardless of how much traffic there is.
• Average order value (AOV). How much each customer spends on average per purchase. Raising this figure (with product recommendations, free shipping above a certain amount, product kits) has a direct impact on revenue without needing more traffic or more customers.
• Customer acquisition cost (CAC). How much it costs, on average, to get someone new to buy for the first time, adding up all marketing spend divided by new customers gained. Without this figure, it's impossible to know if an ad campaign is genuinely profitable.
• Customer lifetime value (LTV). How much a customer generates on average across their entire relationship with the store, not just the first purchase. Comparing LTV with CAC is the most honest comparison there is for whether the business is sustainable long term.
• Cart abandonment rate. What percentage of people who add something to their cart don't complete the purchase. A high figure points to friction in the checkout process or unresolved doubts before paying.
• Repeat purchase rate. What percentage of customers buys again within a given period. It's the most direct indicator of real loyalty, beyond stated good intentions.

Why the relationship between CAC and LTV should guide the most decisions

Of all the metrics, comparing what it costs to acquire a customer (CAC) against what that customer generates over time (LTV) is probably the most important and the least watched by small businesses. A business that spends 20 euros acquiring a customer who only buys once for 25 euros, never returning, barely survives; the same business, with an LTV of 150 euros because the average customer buys six times a year, has far more room to invest aggressively in advertising. Without knowing this relationship, it's impossible to know whether it's worth raising the ad budget or, on the contrary, whether every euro spent attracting traffic is quietly burning money.

A worked example to see it clearly

A store sells a product for 40 euros with a 50% margin. It spends an average of 15 euros on advertising per new customer acquired (CAC of 15 euros). If that customer buys only once, the business earns 20 euros of margin and spent 15 acquiring them: a profit of just 5 euros per customer, a very tight business. If, thanks to good email marketing and loyalty work, that same customer buys on average 2.5 times a year for two years (a real LTV of 100 euros in accumulated margin), that same 15-euro CAC stops being a problem and becomes a clearly profitable investment. The difference between a tight business and a healthy one wasn't in the acquisition cost, it was in what happened after the first sale.

How to build a dashboard without getting lost in data

The temptation, as soon as you discover analytics, is to want to watch forty metrics at once, which paralyses more than it helps. It's better to build a short dashboard (five to seven key metrics) reviewed regularly (weekly for operational ones like conversion and cart abandonment, monthly for more strategic ones like LTV and CAC), rather than a lengthy report nobody reads closely. Tools like Google Analytics 4, Shopify's or WooCommerce's native dashboard, and specific ecommerce analytics plugins already calculate most of these metrics automatically; the real work is deciding which ones to watch, not calculating them by hand.

A real case: the store that discovered its best campaign was its worst

An accessories store ran two ad campaigns in parallel, and for months prioritised the one generating more sales in total volume, assuming it was the more profitable. When they calculated the real CAC of each campaign separately, they discovered that "winning" campaign had an acquisition cost almost three times higher than the second campaign, which generated fewer sales in absolute number but with a far healthier margin. Reallocating budget from the "big but expensive" campaign to the "small but profitable" one improved the business's net profit without selling a single extra unit.

Segment before drawing conclusions

Looking at business-wide metrics without segmenting hides important differences. The overall conversion rate can look acceptable while, within it, one channel converts extraordinarily well and another barely converts at all, cancelling each other out in the average. Segmenting by acquisition channel (advertising, organic, email, social media), by device type (mobile versus desktop, where conversion rates often differ substantially) and by product category almost always reveals improvement opportunities the overall figure hides.

The mistake of comparing your store to generic sector benchmarks

It's tempting to search online for "average ecommerce conversion rate" and use that figure as a yardstick for your own business, but that kind of generic benchmark mixes sectors, average order values and business models so different from each other that the comparison offers little real value. A far more useful benchmark is your own history: comparing the current month against the same month last year, or against the average of the last three months, filters out much of the seasonal noise and gives a much more honest picture of whether the business is improving or worsening on terms that genuinely apply to it.

How to present these metrics if you need to report them to a partner or investor

When these metrics need to be communicated to someone who doesn't live the business day to day (a partner, an investor, a bank for a financing line), it's easy to fall into the temptation of showing only the figures that look good. The more professional approach, and in the long run more useful for yourself, is to always present the CAC-to-LTV relationship alongside the sales figure, not just the latter, because it's what genuinely explains whether growth is sustainable or is being bought by spending more and more on advertising without improving the business's real profitability.

When to trust a low-volume data point less

A common mistake when starting to look closely at metrics is drawing firm conclusions from samples that are too small: a campaign that generated only eight sales can look excellent or terrible depending on the luck of those eight specific orders, without that saying anything reliable about its real performance at greater scale. Before making important decisions based on a metric, it's worth asking whether the volume of data behind that figure is enough to be representative, or whether you're still within the range where a handful of specific orders can completely distort the conclusion.

Common mistakes when interpreting metrics

The first frequent mistake is confusing correlation with causation: seeing sales rise the same month a new feature launched and automatically assuming that feature caused it, without ruling out equally plausible explanations (seasonality, an ad campaign that coincided in timing, a competitor raising prices). Before attributing an improvement to a specific change, it's worth asking what else happened during that same period that could also explain the result.

The second mistake is looking at metrics only in absolute value with no trend context. A 2% conversion rate might sound reasonable in isolation, but if it was 3% six months ago, that same figure actually hides a decline deserving immediate attention. The third, more subtle mistake is optimising one metric in isolation without watching the effect that change has on the others: aggressively raising average order value with very high free-shipping thresholds can, in parallel, tank the conversion rate if the threshold discourages a significant share of customers, leaving the business worse off overall even though the metric being closely watched improved.

The fourth mistake is not distinguishing between a metric the business can directly influence and one it can only observe. Total traffic, for example, depends on many external factors (platform algorithms, sector seasonality) beyond the business's direct control, while the website's own conversion rate is something that can be directly acted on by improving the product page, price or checkout. Focusing most attention on genuinely actionable metrics, instead of obsessing over figures largely driven by external factors, usually gives a better return on the time spent analysing.

Finally, it's worth checking how often a specific metric actually changes meaningfully, because not every one moves at the same pace. Checking the conversion rate daily, when daily order volume is low, generates more statistical noise than useful information; adjusting review frequency to the actual volume of available data avoids drawing conclusions from variations that are nothing more than chance.

Frequently asked questions