Most online stores that close in their first year don’t do so because of a lack of traffic, but because nobody was looking at the right numbers. This guide is intended for those who have just set up their store and need to know what to measure, what to ignore, and why some metrics that sound good are actually misleading.
Start with three numbers, not twenty
Before opening a dashboard with twelve charts, confirm that you can answer these questions: how many people visit?, how many buy? and how much does each buyer leave behind? Those three answers correspond to sessions, conversion rate, and average order value. If these values aren’t clear to you, any additional KPI is noise.
A common mistake in newly launched stores: setting up Google Analytics 4 and assuming that ecommerce works. GA4 doesn’t record revenue automatically if you haven’t enabled purchase event tracking and sent the correct parameters. Many small stores spend weeks seeing “0 revenue” when there are actually sales. Before obsessing over advanced metrics, verify with a real test order that the purchase event arrives with value and currency. In Shopify, WooCommerce, and PrestaShop you have to check that configuration manually; it isn’t always enabled by default.
The KPIs that actually drive decisions
Conversion rate: the most misunderstood figure
A store’s overall conversion rate mixes visitors from brand campaigns with discovery traffic. A store selling accessories can have an overall rate of 0.8% and a 4% rate on brand search traffic. Looking only at the average leads to false conclusions. Segment by channel, device, and product type before drawing conclusions.
According to industry studies, average ecommerce conversion is around 2-3%, but that figure varies enormously depending on the sector and the product price. A 15-euro item is bought with less friction than an 800-euro one.
Average order value (AOV) and margin
Raising the average order value with bundles or free shipping above a threshold works, but it can destroy margin if the free product has a high cost. Always cross-reference AOV with gross margin. An AOV of 60 euros with a 15% margin is worse business than one of 45 with a 40% margin.
Cart abandonment
It’s normal for 60% to 80% of carts not to be completed, according to typical industry data. It’s not a failure: part of the abandonment is people comparing or just checking the shipping cost. What matters is reviewing how many abandonments occur at the shipping or payment step. That’s where there’s real room for improvement.
Customer acquisition cost (CAC) versus LTV
If you pay 25 euros per customer and their first purchase leaves 30, you’re not making a profit: you’re financing acquisition. The useful metric is 12-month LTV compared with CAC. In subscription or high-repeat-purchase stores, CAC can take months to recover. For one-time purchase products, the margin on the first sale must cover the CAC or the model doesn’t hold up.
What these KPIs don’t tell you
No analytics metric tells you why someone left. Cart abandonment measured in GA4 doesn’t explain whether the user distrusted the payment gateway, whether the page took too long to load on their mobile, or whether the coupon didn’t apply. For that you need session recording tools or user testing, and even then you don’t get absolute certainties. Quantitative analytics points to where to look; qualitative analytics explains the reason.
Another practical limit: cookie consent reduces the real visibility of traffic. With a strict consent banner, GA4 may record fewer sessions than actually occur. Comparing the total orders in the store dashboard with those in analytics and seeing a 15-30% difference is common; it doesn’t always mean something is broken.
A minimum viable dashboard to get started
With a spreadsheet or your platform’s native dashboard you can build a weekly scorecard with: sessions by channel, conversion rate by device, revenue, AOV, estimated gross margin, orders, and percentage of orders with a coupon. Review it every Monday with the same time window and note down what you changed that week. Week-over-week comparison with context is worth more than any automatic chart without explanation.
Honest conclusion
There’s no combination of KPIs that guarantees growth. These indicators serve to detect problems and test hypotheses, not to predict sales with precision. If your store sells little, analytics will tell you where the funnel breaks, but it won’t fix a poorly positioned catalog or prices that are out of line with the market. Start by measuring three figures well, validate purchase tracking, and expand the dashboard only when decisions depend on new data.