How to read GA4 data to understand where you lose customers

The funnel report: the abandonment map

Funnel Exploration in GA4 (Explore > Funnel Exploration) shows exactly how many users move from one step to the next. Normal abandonment rates: session to product view (40-60% of sessions reach a product page), product view to add to cart (5-15%), add to cart to checkout (40-60% abandon), checkout to completed order (20-40% abandon). If your abandonment on a specific step is 50% higher than these averages, there is a concrete problem there.

How to interpret abandonment rates per step

A 90% abandonment from product view to add to cart is not necessarily a problem: users explore, compare, research. Problematic abandonment is from add_to_cart to begin_checkout (if it exceeds 60%) and from begin_checkout to purchase (if it exceeds 50%). These two transitions are where intent was clear and something prevented completion.

The correct approach: calculate absolute abandonment (real numbers, not percentages) at each step. An 80% abandonment on a funnel with 5,000 cart entries matters more than 95% on a funnel with 200 entries.

Open funnel vs. closed funnel: the difference that matters

GA4 allows configuring the funnel as "open" (users can enter at any step) or "closed" (must go through all steps in order). For realistic analysis, use an open funnel: many users add directly to cart after seeing the product in another context (email, ad). A closed funnel excludes these users and underestimates real conversion.

How to build a correct funnel in GA4

GA4 uses events, not pageviews, as the basis for funnels. Standard eCommerce events: view_item, add_to_cart, begin_checkout, purchase. Quick check: Settings > Events. If you don't see add_to_cart and purchase in the events list, eCommerce tracking is not configured.

The most frequent tracking gap on Romanian stores: begin_checkout is configured, but purchase doesn't fire for all payments (especially for Netopia and cash on delivery).

Verifying data quality before any analysis

Before interpreting any GA4 report, check three things: 1) the purchase event fires for all payment methods (test card payment, cash on delivery, voucher), 2) there are no session duplicates caused by multiple tags, 3) internal traffic is excluded via IP filter. Dirty data produces wrong conclusions regardless of how much time you spend in reports.

Custom events that add context

Standard events (view_item, add_to_cart, purchase) tell you what happened, not why. Custom events that add valuable context: scroll_depth (how many users reached the description, reviews, recommendations section), form_error (which checkout fields produce errors), and size_chart_click (how many users checked the size guide before adding to cart). This data transforms funnel analysis from descriptive to diagnostic.

Path Exploration: what people do before buying

Set the starting point as the purchase event and go backwards: you'll see that 60-70% of buyers visited the returns or FAQ page before ordering. If people look for return information before buying, make your return policy more visible on the product page, not just in the footer.

Another common pattern: users who buy visit an average of 4-7 product pages, not 1-2. If your product pages don't have an efficient navigation system between similar products, you're losing additional orders.

Navigation patterns that indicate purchase intent

Path exploration analysis on users who completed orders reveals consistent patterns: visiting the FAQ or Contact page before purchase indicates need for confirmation (especially for high-value products), repeated visits to the same product page across multiple sessions indicates active deliberation and comparison, and direct navigation to the checkout page (without add_to_cart in the current session) indicates a returning user with products in a saved cart.

How to use Path Exploration to improve navigation

If Path Exploration shows 30% of users hit a "Page not found" (404) on the way to purchase, there are critical broken links. If users need more than 3-4 clicks to get from the homepage to a specific product, the navigation structure is too deep. Both problems are fixable and have direct conversion impact.

Segmenting the data: where is the problem coming from

• Mobile vs desktop: if mobile rate is below 50% of desktop rate, there is a specific mobile UX issue
• New vs returning: new users convert at 0.5-1.5%, returning at 3-8%
• Traffic source: if Meta Ads brings 40% of traffic but only 15% of orders, the audience is poorly qualified
• Geographic location: in Romania, Bucharest and Cluj-Napoca frequently have conversion rates 2-3x higher than smaller cities

Behavioral segmentation: converters vs. non-converters

The most valuable segmentation in GA4: create two segments, "Users who placed an order in the last 30 days" and "Users who added to cart but didn't buy in the last 30 days". Compare their on-site behavior: pages visited, session duration, number of sessions. The differences show exactly what the buyer does differently from the non-buyer with intent.

Attribution analysis: which channel wins the order

GA4 Attribution Reports (Advertising > Attribution) show which channels initiate the order (first touch) and which complete it (last touch). If Meta Ads appears consistently at first touch but rarely at last touch, the platform works as a discovery channel, not direct conversion. This changes how you evaluate ROAS and how you allocate budget.

What to do with the information

Priority order: fix tracking issues first, then identify the step with the highest absolute abandonment, then test one change at that step. Complementary tools: Hotjar or Microsoft Clarity (free) for heatmaps and session recordings — they show what the user does on the high-abandonment page, something GA4 cannot do alone.

A serious GA4 audit on an online store takes 2-3 hours and typically identifies 3-5 concrete problems with direct sales impact. If you don't have time for it, GA4 data remains a decorative report.

Prioritizing actions by impact and effort

Not all problems identified in GA4 deserve the same attention. Correct prioritization: high impact + low effort (enabling Apple Pay, displaying return policy more visibly) gets resolved immediately. High impact + high effort (rebuilding checkout, restructuring navigation) gets planned in sprints. Low impact + low effort (copy adjustments, colors) gets A/B tested. Low impact + high effort gets ignored.

Regular reporting: what to track weekly

Minimum weekly dashboard for an online store: conversion rate per traffic source (versus previous week and same period last year), abandonment rate per funnel step, and revenue per channel. If a metric deviates more than 20% from the average of the last 4 weeks, investigate the cause before making any budget decisions.