By Stormly in Knowledge
Published: Apr 13, 2026
The 7 eCommerce KPIs That Actually Drive Decisions (Not Just Reporting)
You have an eCommerce KPI dashboard. It shows revenue, sessions, and conversion rate. Maybe you have a few more tiles: average order value, bounce rate, traffic by channel.
You look at it every week. And then you make a judgment call anyway, because nothing on that dashboard tells you which product to push in this week’s email, which SKU to pull from ads, or which customer segment is about to stop buying.
That is not a reporting problem. It is a metric selection problem.
Most ecommerce dashboards track the wrong things. Not wrong in the sense of incorrect data, but wrong in the sense of metrics that describe the past without pointing toward a specific next action. This post is about the 7 KPIs that actually do that.
Each one below comes with one specific decision it enables. If a metric cannot tell you what to do, it is not worth tracking weekly.
Why Most eCommerce KPI Dashboards Are Filled with the Wrong Numbers
Improvado’s guide to ecommerce analytics lists 14 metrics. Triple Whale’s retention guide covers customer retention rate, churn rate, revenue churn, CLTV, repeat purchase rate, and NPS. ContentSquare focuses on bounce rates and UX flow.
All of those metrics have their place. But none of them answer the question you actually have on Monday morning: what do I do this week?
The problem is that most analytics tools were built for marketing attribution or session reporting. They show you what happened at the traffic and revenue level. They cannot show you what happened at the product, SKU, or category level. And for an ecommerce operator with 200 products, the session-level view is almost useless for making product decisions.
The 7 KPIs below work because they are product-level metrics, not store-level averages.
The 7 eCommerce KPIs That Drive Decisions
1. Product-Level Conversion Rate
Not your store’s overall CVR. The CVR for each individual product page.
This distinction matters more than most operators realize. Your store might convert at 3.1%. But product A converts at 14% and product B converts at 0.3%, and both get the same ad spend.
A real Stormly example: a store with 400 SKUs had an overall CVR of 2.9%. The product-level CVR report showed a candle set converting at 16% while a comparable premium candle was converting at 0.6%. Same category, same price point, wildly different conversion. The 0.6% product had a single image and no size information. Fixed in one afternoon.
The decision it enables: Which products to scale in ads, which to fix before promoting, and which to deprioritize entirely. Without product-level CVR, you are optimizing a store average that masks everything meaningful underneath it.
Most analytics tools, including GA4, report session-level or goal-level conversion rates. To get product-level CVR broken out by SKU, you need a tool that tracks at the product layer natively. Stormly surfaces this in its product performance report without custom event setup.
2. Cart Abandonment by SKU
Not your overall cart abandonment rate. Which specific products appear most frequently in abandoned carts.
Your overall cart abandonment rate might be 68%. That number tells you nothing about which product is causing it. Cart abandonment by SKU tells you something actionable.
A concrete example: a store running a weekly flash sale found that one product appeared in 42% of all abandoned carts during the sale. The product had a shipping estimate that showed “5-7 business days” for a sale item most customers expected to arrive sooner. Adding “ships today before 2pm” to the product page dropped abandonment for that SKU by 28% within one week.
That insight is completely invisible in a blended cart abandonment rate. You would never find it in GA4. You would never find it in Shopify Analytics.
The decision it enables: Whether cart abandonment is a store-wide funnel problem or a product-specific issue (and therefore whether to fix checkout, pricing, or product presentation). These require very different responses.
Stormly’s cart abandonment report breaks this down by SKU, brand, and category natively. Other tools approximate this only with custom event tracking.
3. New vs. Returning Customer Split by Product Category
Not just your overall new vs. returning ratio. Which product categories are primarily acquisition drivers versus retention drivers.
Some product categories bring in first-time buyers. Others bring in repeat customers. Knowing which is which changes how you allocate acquisition spend.
A practical example: a home goods store discovered through this breakdown that their candles and diffusers drove 74% first-time buyers, while their refill packs drove 88% returning customers. They had been running paid ads evenly across both categories. The insight led them to cut paid ads on refills (already a retention product) and double down on candle acquisition. CAC dropped across the account.
The decision it enables: Where to allocate acquisition budget versus where to invest in retention tactics. Blending these into a single new/returning percentage obscures a meaningful strategic split.
4. 30-Day Cohort Retention by First Purchase Category
Which product category a customer buys first determines how likely they are to come back within 30 days.
This is the most underused metric in ecommerce, and it is where Stormly’s product analytics approach diverges most sharply from general-purpose tools. A standard cohort analysis shows you how a group of customers acquired in a given month retained over time. A product-level cohort shows you how customers who first bought product category X retained compared to customers who first bought category Y.
The numbers here are often dramatic. In one example, customers whose first purchase was a skincare kit had a 41% 30-day retention rate. Customers whose first purchase was a one-time promo item had 9% retention. Same acquisition month, same channels, completely different LTV trajectories based on first product purchased.
The decision it enables: Which products to push in acquisition campaigns (not just which ones sell, but which ones build long-term customers). This changes email welcome sequences, ad creative selection, and even product bundling strategy.
See your own eCommerce KPI dashboard in 5 minutes – try Stormly free
5. AOV Trend by Product Mix
Not just average order value as a number. The trend over time, broken down by which products or categories are contributing to it.
If your AOV drops from $78 to $64 over six weeks, you have a problem. But what kind of problem? Three very different things could be causing it:
- A high-AOV product lost traction and stopped selling
- Customers are shifting to lower-price items in your catalog
- A promotional discount is pulling revenue down artificially
Each cause has a different response. The first is a product visibility issue. The second might signal a positioning problem. The third is not a problem at all. You cannot know which without the product-level breakdown.
The decision it enables: Whether to change pricing strategy, run a bundling promotion, investigate a specific product’s performance, or ignore the drop entirely. An aggregate AOV trend without the product-mix layer attached is mostly noise.
6. Anomaly Alerts on Product Performance
Weekly review is valuable. Catching a sudden 40% drop in product CVR on Wednesday is more valuable.
Anomaly detection is not about setting manual thresholds and waiting for email alerts. It is about having a system that watches all your products simultaneously and flags the ones that changed meaningfully from baseline, without you having to check each one individually.
A real scenario: a store owner found out on Friday that a product had dropped 60% in add-to-cart rate over the prior three days. The cause: a third-party review app had inserted a broken widget on that product page that was invisible on desktop but covered the “Add to Cart” button on mobile. The product was responsible for 18% of weekly revenue. Three days of broken mobile experience, caught only because an anomaly alert flagged it.
The decision it enables: Catching revenue leaks before they compound into week-over-week reporting. Manual dashboards do not catch anomalies. Automated anomaly detection does.
Stormly’s AI surfaces these automatically, flagging products that deviate meaningfully from their baseline without any manual alert configuration. You do not need to decide in advance which products to watch.
7. Churn Risk Score by Customer Segment
Not how many customers churned last month. Which specific customer segments are at high risk of churning in the next 30 days, based on product behavior signals.
A churn rate tells you what already happened. A churn risk score gives you time to act.
For a subscription merchant with 4,200 active subscribers, the churn rate as a number is nearly useless on its own. But if you can see that customers who bought a specific product bundle in their first order and have not repurchased within 45 days have a 73% churn rate historically, you can design a specific intervention for that segment while they are still active.
The decision it enables: Which customer segments to target with retention campaigns this week, what to offer them, and which product behaviors are the leading indicator of churn in your specific store. This is not a metric you can derive from Shopify Analytics or GA4. It requires product-level behavioral data connected to customer history.
Stormly flags at-risk segments automatically using this behavioral pattern. The report shows you the segment, the size, and the signals driving the prediction, without requiring you to build a retention model from scratch.
What These 7 KPIs Have in Common
All seven are product-level metrics, not store-level averages. That distinction is the entire point.
GA4 gives you session data. Shopify Analytics gives you order data. Marketing attribution tools give you channel data. None of them give you the product-catalog layer that connects sessions to specific SKUs, shows you which categories drive retention, or flags which customer segments are about to leave.
This is not a criticism of those tools. GA4 was built for session reporting. Shopify Analytics was built to show you what sold. They do those things well. They were never designed to answer: “Which product should I scale this week?”
Most tools require custom event tracking, BI tools, or data exports to get to the product-level view. Stormly delivers all 7 of these KPIs out of the box for ecommerce stores. The dashboard is built specifically for product decisions, not web analytics or marketing attribution.
Setting Up Your eCommerce KPI Dashboard
The goal is not a dashboard with 25 tiles. It is a dashboard where every metric points toward a specific action.
If you are starting from scratch, the order of priority for these 7 KPIs is roughly:
- Product-level CVR (find underperformers immediately)
- Cart abandonment by SKU (find revenue leaks)
- Anomaly alerts (catch drops before they compound)
- 30-day cohort retention by first purchase (understand which products build long-term customers)
- Churn risk score (act on at-risk segments before they leave)
- New vs. returning by category (align acquisition and retention spend)
- AOV trend by product mix (contextualize revenue changes)
The first three are operational metrics. They help you catch problems and fix them this week. The next four are strategic metrics. They help you make better decisions over the next quarter.
Running both layers together is what separates ecommerce operators who grow systematically from those who are always reacting.
See your own eCommerce KPI dashboard in 5 minutes – try Stormly free
If you have been looking at your analytics and still not knowing what to do with it, the issue is probably not the data quality. It is that you are tracking store-level averages instead of product-level signals. These 7 metrics will not tell you everything. But they will tell you what to do next week.