Grow Your Online Shop Revenue with Precise Web Analytics

Web analytics for ecommerce: the metrics that really count, and the ones you should ignore

Every ecommerce shop has analytics. Google Analytics, Shopify Analytics, Adobe, Matomo. Most shops have more data than they can handle. But data without interpretation is just numbers. And the wrong numbers lead to the wrong decisions.

Shops easily spend 20 hours a month on analytics reports without a single measurable improvement to show for it. They track the wrong metrics, read correlation as causation and optimise vanity metrics instead of revenue. This article shows which metrics really count, and how to turn data into decisions that bring in revenue.

The three categories of ecommerce metrics

1. Vanity metrics: they look good and achieve nothing

Vanity metrics are numbers that sound good but have no effect on revenue. They are a distraction.

Page views: a shop with 1 million page views a month sounds impressive. But if the conversion rate is 0.1%, that is 1,000 purchases. A shop with 100,000 views and 2% conversion makes 2,000 purchases, with a tenth of the traffic. Page views are not a success metric. They are a context metric.

Time on site: a longer visit does not mean more engagement. Often it means more confusion. A customer who finds what they want in 30 seconds and buys is better than one who searches for 5 minutes and then leaves.

Example: a shop sees time on site rise by 40% after a redesign while its conversion rate falls by 15%. The new navigation looks nicer but is harder to follow, so people spend longer looking for what they want.

Social shares: a blog post with 500 shares sounds good. But if those shares come from people who never buy and generate no conversions, they are worthless. Social shares are branding, not revenue.

2. Context metrics: useful for diagnosis, not as goals

Context metrics help you understand WHAT is happening. They should not be goals. They should answer questions.

Bounce rate: a high bounce rate on a landing page is bad. A high bounce rate on a blog page is normal. A high bounce rate on a product page means the customer didn’t find what they expected. Bounce rate alone tells you nothing. It only means something in the context of the page.

Traffic sources: knowing where your visitors come from matters. But judging which channel is “better” from the traffic source alone leads to wrong decisions. An organic visitor who converts is not “better” than a paid visitor who converts, as long as the cost per acquisition (CPA) stays below the lifetime value (LTV).

3. Revenue metrics: the only numbers that count

These metrics are directly linked to your revenue. Every improvement here means more revenue.

Revenue per user (RPU): revenue divided by visitors. This is the single most important metric. It combines traffic quality, conversion rate and average order value in one number. If RPU goes up, you earn more, whatever else happens.

A real example: a fashion shop optimised its homepage. The conversion rate rose from 1.8% to 2.1%. The average order value fell from €85 to €78 (because the new homepage pushed bargain products). The conversion rate went up, but RPU went down. The shop thought it had won. In fact it had lost.

Conversion rate by channel: not the overall conversion rate, the conversion rate per channel. A channel with 5% conversion and 1,000 visitors brings 50 purchases. A channel with 1% conversion and 10,000 visitors brings 100 purchases. The first channel is more efficient, but the second brings more revenue. The decision: invest more in the second channel without neglecting the first.

Basket-to-checkout rate: how many of the people who add something to the basket reach the checkout? This metric shows friction in the basket. A shop with 40% basket-to-checkout has a basket problem. A shop with 80% has a checkout problem. The fixes are completely different.

The analytics setups every shop needs

1. Ecommerce tracking in Google Analytics 4

Without ecommerce tracking you don’t know which products sell, which categories perform and which marketing channels bring in revenue. The setup is a one-off and takes 2–4 hours. The ROI is huge.

What you need to track: add_to_cart, begin_checkout, purchase, refund, view_item, view_item_list. Without these events your analytics data is useless for ecommerce decisions.

2. Checkout funnel analysis

The checkout funnel shows you WHERE customers abandon. Not just that they abandon, but at which step.

An electronics shop had a checkout abandonment rate of 65%. The funnel showed: step 1 (address), 92% pass-through. Step 2 (shipping), 78%. Step 3 (payment), 61%. The problem was not the checkout as a whole. It was the payment step.

The cause: payment methods only appeared in step 3. Customers entered all their details and then found their preferred payment method missing. Once the payment methods were shown on the basket page, checkout pass-through rose to 74%.

3. Cohort analysis for customer loyalty

Cohort analysis groups customers by the date of their first purchase and shows how many of them buy again after 30, 60 and 90 days. This is the most important metric for long-term growth.

A beauty shop had strong first-purchase conversion, but cohort analysis showed that only 8% of first-time buyers bought again within 90 days. The answer was not more first-time buyer acquisition. It was better onboarding. A welcome email flow with care tips, product help and a reorder reminder raised the 90-day repurchase rate to 23%. LTV doubled.

From analytics to action: the workflow

This is how we turn data into decisions:

1. 1
   

   Ask a question.

   “Why is our mobile conversion 40% lower than desktop?”

2. 2
   

   Form a hypothesis.

   “The mobile product page has too many elements above the fold. Customers don’t scroll down to the CTA.”

3. 3
   

   Collect data.

   Heatmaps, scroll maps, session recordings. Not just analytics, but qualitative data too.

4. 4
   

   Test.

   A/B test: variant A (current) vs. variant B (CTA placed higher).

5. 5
   

   Decide.

   If B wins, roll it out. If A wins, form a new hypothesis.

6. 6
   

   Document.

   What did we learn? Every insight feeds into the next test.

Frequently asked questions about web analytics

Which analytics tool should I use?

Google Analytics 4 is the standard: comprehensive and well documented. For privacy-conscious shops: Matomo (self-hosted). For enterprise: Adobe Analytics. The setup matters more than the tool. A badly configured premium tool is worse than a well configured basic one.

How often should I review analytics?

Daily for operational metrics (revenue, conversion, abandonment rate). Weekly for strategic metrics (cohort analysis, channel performance). Monthly for strategic reviews (LTV, CAC, RPU trends). Never less than monthly, or you will miss problems until they get expensive.

What is the most important metric?

Revenue per user (RPU). It brings together everything that counts: traffic quality, conversion rate, AOV. If you only track one metric, make it RPU.

How do I avoid analysis paralysis?

Tie every metric to an action. If a metric doesn’t trigger a concrete action, don’t track it. “The bounce rate is 45%” is a fact. “The bounce rate is 45%, so we’re testing a new headline” is an action. No action, no metric.

Should I hire a data analyst?

For shops under €1M revenue a year: no. Use a CRO tool or a freelancer for the setup and weekly reviews. For shops above €1M: yes, but hire someone who tests, not someone who only reports. An analyst who doesn’t run tests is an expensive reporter.

You will find more on this in the Academy track Data-Driven Decision Making ↗.