Ecommerce KPIs — Key Performance Indicators — are the measurable metrics that tell you whether your business is growing, stalling, or declining. They help you understand overall commercial health, customer behaviour, and financial performance. Without them, decisions are based on gut feel, historical assumption, or whoever shouts loudest in the weekly trading meeting.

I track seven core KPIs in every ecommerce business I work with. Not because there's anything magical about the number — there are dozens of metrics available — but because these seven, tracked consistently and reviewed honestly, give you a complete picture of what's actually happening to your customer base and your revenue.

"Every business should track KPIs to ensure it's growing and to address any issues identified. The goal is to make the picture clear enough that you can act on it — not to have a dashboard full of numbers nobody looks at."
KPI 1
New Customers
KPI 2
Returning Customers
KPI 3
Second Orders
KPI 4
At-Risk Customers
KPI 5
Lapsing Customers
KPI 6
Conversion Rate
KPI 7
Traffic by Channel
+
AOV · Orders · Revenue
1
Acquisition
New Customers Acquired

A fundamental metric for any business. Continuously identifying new opportunities for customer acquisition is essential — and understanding the cost of that acquisition is equally critical. Your marketing budget has been approved on the expectation of a return. Monitoring cost per acquisition (CPA) by channel tells you which acquisition investment is actually delivering value and which is generating volume at the wrong price.

After acquisition, new customers should immediately enter an email welcome programme designed to encourage that crucial second purchase. Without this, new customer acquisition is a leaky bucket — you fill it, but it drains faster than you realise.

Real example A retailer I worked with saw a 6% YOY increase in new customers for the year — but there was a sharp drop in the early months, with recovery only in the final five months. Without tracking this monthly, the underlying pattern was invisible and the recovery couldn't be linked to any specific intervention.
2
Retention
Returning Customers

Are customers continuing to engage with your brand? This metric should remain stable or increase over time as you acquire new customers and retain existing ones. A drop in returning customers is one of the earliest and most reliable warning signals in ecommerce — it typically surfaces before revenue decline becomes visible, giving you time to act if you're watching it closely.

Common causes of a returning customer decline include delivery price increases that pushed customers to competitors, product availability issues, or a deterioration in service quality that has eroded trust without generating immediate complaints.

Real example One retailer experienced a sharp decline in returning customers that was eventually traced back to stock issues and product availability. The decline had been running for two months before it became visible in the revenue line — those two months would have been recoverable with earlier intervention.
3
Lifecycle
Customers Making Second Purchases

The second purchase is the most important conversion in ecommerce. A customer who buys once might be a one-off transactor — driven by a promotion, a Google Shopping click, or an affiliate voucher. A customer who buys twice has demonstrated intent to have an ongoing relationship with the brand. Second-order rate is the strongest leading indicator of long-term customer value and the most direct measure of whether your post-purchase CRM is working.

Encouraging that second purchase requires relevant engagement — sharing your brand story, offering upsell and cross-sell opportunities based on the first purchase, and timing communications to the right moment in the replenishment or consideration cycle.

Real example One retailer faced consistent difficulty getting customers to make a second purchase online — not because they were dissatisfied, but because they were defaulting to the physical store for subsequent orders. Understanding this channel behaviour was essential to designing the right post-purchase experience.
4
Retention
At-Risk Customers

At-risk customers are those who have purchased before but whose recency and frequency metrics suggest they are beginning to disengage — they haven't reached full lapse status yet, but they're heading there. This is your last, most cost-effective opportunity to retain them before they cross the threshold into lapsed.

Every business has a different purchase cycle, so the threshold for "at-risk" will vary. A subscription food business might define at-risk as no purchase in 45 days. A garden furniture retailer might use 18 months. Define the threshold for your specific business and purchase cycle — then track it consistently.

Real example A retailer I worked with had a high and growing volume of at-risk customers that hadn't been identified because nobody was tracking the metric. By analysing their purchase behaviour and implementing earlier, more relevant re-engagement communications, we materially reduced the at-risk volume over the following quarter.
5
Churn
Lapsing Customers

Lapsing customers are those who were once active but have stopped purchasing — they've passed through at-risk status and have effectively churned. Identifying and tracking this cohort is critical because it allows you to implement retention strategies before the cost of recovery becomes prohibitive, and to understand the true churn dynamics of your customer base.

Not all lapsing customers are the same. Some have lapsed because they no longer have the need. Others lapsed because of a poor experience. And some lapsed because they were acquired in the first place through a promotion that was never going to generate a loyal customer — and understanding that distinction is what prevents you from chasing the wrong cohort with the wrong message.

The Giveaway Warning — A Real Example

One retailer ran a giveaway campaign that drove 10,000 new customers. It looked like a significant success. But when we tracked the second-order rate for that cohort, 88% never returned to make a second purchase. The campaign had generated volume at significant cost but delivered almost no commercially durable customers. Tracking lapsing customers by acquisition cohort is what reveals this — and prevents repeating the same mistake.

6
Commercial Performance
Conversion Rate

Website conversion rate is the ratio of orders to total sessions — and it's one of the most commercially sensitive metrics in ecommerce. A small movement in conversion rate has a disproportionate impact on revenue: a site converting at 1.5% instead of 1.0% on the same traffic is generating 50% more orders without spending a penny more on acquisition.

Conversion rate benchmarks vary significantly by sector — fashion retailers might typically see 4–5%, while specialist B2B or higher-consideration categories will operate at lower rates. What matters is your trend relative to your own baseline, not a generic industry benchmark.

Always segment conversion rate by traffic source — organic, paid, email, and direct will convert at very different rates, and blending them into a single number hides the most useful information.

Real example A business with a typically high conversion rate saw a sudden dip that initially appeared to be a traffic quality issue. Investigation revealed it was a stock availability problem — customers were reaching product pages and finding items unavailable. Once stock was addressed, conversion returned to its previous level within two weeks.
7
Traffic
Traffic by Channel

Tracking traffic in aggregate tells you very little. Tracking it by channel tells you almost everything. Each channel has a distinct profile — different conversion rates, different customer LTV, different cost structure, and different sensitivity to changes in your marketing activity. When set up properly, channel-level traffic tracking allows you to evaluate precisely what's driving growth and what isn't.

The channels that matter most vary by business, but the ones I always examine first are direct traffic (a proxy for brand health), organic search (a measure of SEO investment and brand strength), and email (a measure of CRM programme health and database quality).

Real example A business I worked with faced a decline in both email and direct traffic simultaneously — email due to regulatory changes that restricted promotional sends for five months, and direct due to years of underinvestment in brand. The one bright spot was organic search, which with proper investment showed a significant increase in qualified traffic. Without channel-level tracking, the organic growth would have been invisible inside an overall traffic decline.

The Supporting Metrics

The seven KPIs above are the core framework. But they work best alongside four supporting metrics that add commercial context to the customer-focused view.

Number of Orders
The volume of transactions — distinct from revenue and useful for understanding whether revenue movements are driven by order volume or by AOV changes. These two causes require very different responses.
Units per Order
How many items customers are buying per transaction. A key lever for revenue growth that requires no additional acquisition — cross-sell, bundles, and free delivery thresholds all move this number.
Average Order Value (AOV)
The average revenue per transaction. Tracks the commercial effectiveness of your merchandising, upsell strategy, and promotional mechanics. A rising AOV with flat order volume is a strong positive signal.
Revenue
The output metric. Revenue tells you the result — the other KPIs tell you why. Revenue tracked in isolation without the customer and channel context is the least useful metric in the set, even though it's often the most reported.
The 7 KPIs
  • New Customers — track monthly with CPA by channel. Feed immediately into a welcome programme
  • Returning Customers — an early warning signal. A drop here surfaces before revenue decline by weeks or months
  • Second Orders — the strongest predictor of long-term customer value. Track the rate, not just the volume
  • At-Risk Customers — define the threshold for your purchase cycle and track it consistently. Earlier intervention is always cheaper
  • Lapsing Customers — track by acquisition cohort. The giveaway that generated 10,000 customers who 88% never returned from is a salutary lesson
  • Conversion Rate — segment by channel. A single blended rate hides the most useful information
  • Traffic by Channel — direct, organic, and email are the ones to watch most closely. Each tells a different part of the commercial story