Google Analytics 4 made a lot of marketers miserable for two years. The reports were unfamiliar, the data structure was different, and most blog posts explaining it were written before GA4 was actually finished. In 2025 the platform is stable, the reports are usable, and the analytics are genuinely better than what we had before — once you know which reports to trust and which to ignore.

The first thing to know about Google Analytics 4 is that the default reports are kind of bad. Not bad-information bad — the underlying data is good. But the way it's presented out of the box doesn't surface the questions most marketers actually want to answer. Once you build a few custom Explorations, GA4 becomes far more useful than Universal Analytics ever was.

This guide walks through the reports I actually use weekly, the metrics that matter (and the ones that don't), and how to extract real insights instead of just numbers.

Setting up GA4 correctly

Before we discuss reports, the foundation has to be right. The most common GA4 problems come from a botched initial setup, not from the platform itself.

Conversion events

In GA4, "conversions" are just events you've manually marked as conversions. Decide which 3-5 events represent real business value. Typically: form submissions, purchases, demo bookings, key page views (pricing page, contact page), and mark them in Admin → Events → Conversions.

Don't mark every event. If you mark "scroll" or "video_play" as conversions, your conversion totals become meaningless. Be ruthless about what counts.

Cross-domain tracking

If your site has subdomains or you redirect users to a third-party checkout (Stripe, Razorpay, Calendly), configure cross-domain tracking. Otherwise GA4 sees the redirect as a new session and attributes the conversion to "direct" traffic instead of the original source.

Google's official cross-domain tracking documentation walks through it. It's a 15-minute setup that saves weeks of misattributed data.

Enhanced measurement

GA4's enhanced measurement automatically tracks scroll depth, file downloads, outbound clicks, video engagement, and form interactions. By default these are all enabled — go to Admin → Data Streams → Enhanced Measurement to confirm.

These automated events are useful but they generate noise. Most teams turn off "scroll" tracking because every long-form blog post triggers it constantly.

The four reports I check every Monday

GA4 ships with hundreds of report templates. Most are forgettable. These four answer almost every weekly question I get from clients.

1. Traffic acquisition by session source/medium

Where: Reports → Acquisition → Traffic acquisition

What it shows: Where your visitors came from, broken down by channel (Organic Search, Paid Search, Direct, Social, etc.), source (google, facebook, etc.), and medium (organic, cpc, social, etc.).

Why it matters: This is the single most important report for understanding which marketing efforts are working. If you're spending ₹1 lakh per month on PPC and your "Paid Search" channel is bringing in 5% of total sessions, you have a structural problem.

The trap: GA4's default attribution is data-driven, which means it spreads credit across multiple touchpoints. Most marketers want last-non-direct-click attribution for weekly reporting. Switch attribution model in Admin → Attribution Settings.

2. Pages and screens by engagement

Where: Reports → Engagement → Pages and screens

What it shows: Which pages on your site are getting the most traffic and engagement.

Why it matters: Tells you what content is working and what's not. The pages with high traffic but low engagement (low average engagement time, low conversion rate) are candidates for content refreshes or removal.

Custom view I add: Sort by Sessions descending, add Engagement rate as a secondary metric, filter to last 90 days. This surfaces pages that should be your highest-priority for optimization.

The SEO score checker helps prioritise — run it on your top 10 pages by traffic, fix the issues it finds, watch the engagement rate climb.

3. Conversions by event name

Where: Reports → Engagement → Conversions

What it shows: Which conversion events fired, how often, and broken down by source.

Why it matters: This is where you see whether your marketing is producing actual outcomes vs just traffic. A drop in form submissions while traffic stays flat is much more important than a drop in pageviews.

The trap: Some conversion events fire multiple times per session (e.g. Multiple form submissions on the same page). GA4 counts each. For unique-conversion analysis, use the "first time" parameter.

4. User exploration

Where: Explore → User exploration

What it shows: Individual user journeys, every event a specific user fired, in order.

Why it matters: When numbers don't make sense, looking at individual journeys reveals why. "Why is my conversion rate so low?", open user exploration, watch how 10 actual users behave on the site, and the answer becomes obvious.

This is the report that's missing from Universal Analytics and changes how you debug funnels.

Marketing analyst reviewing Google Analytics 4 reports on a laptop with charts
GA4's Explore section is where the real value lives — the default reports are a starting point, not an endpoint.

Custom Explorations worth building

GA4's Explore section is where you build the reports the default templates don't include. Four explorations that pay back the setup time:

Funnel exploration

Build a multi-step funnel from "session start" → "view product page" → "begin checkout" → "purchase" (or whatever your conversion path is). Each step shows drop-off rate.

The single most actionable report in GA4 for ecommerce. Where the drop-off is biggest is where to focus optimization.

Path exploration

Shows the most common sequences of pages users follow. Useful for understanding actual user behavior vs your assumed flow.

I once found a client where 30% of users went from blog post → about page → contact page (instead of the assumed blog → product page → contact). Restructured internal links accordingly. Conversions rose 18% in two weeks.

Cohort exploration

Groups users by first-visit date and shows retention over time. Critical for SaaS or subscription businesses. Useful for content sites too. See whether users who arrived in October keep coming back, or whether they were one-off visitors.

Segment overlap

Shows the overlap between user segments — users who are "engaged" AND "purchased" vs those in only one segment. Useful for finding patterns ("most purchasers were also email subscribers" — now you know to invest in email).

Metrics to ignore in GA4

A few default metrics produce more noise than signal:

  • Bounce rate, GA4 redefined this from Universal Analytics. The new definition is "sessions that don't include any engagement." It's not directly comparable to old bounce rate. Use Engagement Rate instead.
  • Sessions per user, affected by cross-device behavior, hard to interpret in absolute terms.
  • Average engagement time per session; heavily skewed by a few power users on long sessions. Median would be more useful but GA4 doesn't expose it.
  • Total events — meaningless without context. Total events going up could mean engagement is improving or it could mean you accidentally added a noisy event.

Focus on: sessions, conversions, conversion rate, engagement rate, average engagement time per active user (the per-user version).

Connecting GA4 to other tools

GA4 becomes more useful when paired with:

  • Google Search Console — link them in Admin → Property Settings → Search Console links. Now your GA4 reports include search query data.
  • Google Ads, link them. Conversion data flows back to Google Ads automatically and you can use GA4 audiences for remarketing.
  • BigQuery. Free for GA4 standard. Export your raw event data to BigQuery for SQL-based analysis. Major for any team with even a junior data analyst.
  • Looker Studio; for unified dashboards combining GA4, Search Console, and Google Ads.

The free link tracker on this site helps generate consistent UTM parameters — without consistent UTMs, GA4's attribution data is noisy regardless of your reports.

The post-cookie attribution challenge

GA4 uses different attribution models than its predecessor:

  • Data-driven attribution (default, most accurate when you have enough data)
  • Last click (default for older accounts)
  • Position-based (40% first touch, 40% last touch, 20% middle)
  • Linear (equal credit across all touchpoints)
  • Time decay (more credit to touchpoints closer to conversion)
  • First click (100% to first touchpoint)

For weekly reporting, last-non-direct-click is usually clearest. For deeper analysis, data-driven attribution is more accurate but harder to explain to stakeholders.

GA4 also handles iOS-related attribution gaps with modeled conversions — Google estimates conversions that couldn't be tracked due to consent or browser restrictions. These are reported separately so you can see modeled vs measured conversions.

What a typical weekly review looks like

My Monday morning GA4 routine for a typical client takes about 30 minutes:

  1. Open Traffic Acquisition. Compare last 7 days to previous 7 days. Note any 20%+ moves up or down.
  2. Open Conversions. Same comparison. Investigate any drops.
  3. Open Search Console (linked in GA4). Check for new queries appearing in Performance, large CTR drops on important queries.
  4. Run a saved Funnel Exploration. Check if drop-off rates have shifted.
  5. Check the User Exploration. Look at 5-10 random users from the previous week. Spot anything weird in their journey.
  6. Note 2-3 questions for the team. Why did organic traffic drop on this specific day? Why did paid CPA spike on Wednesday?

The point isn't to read every chart. It's to spot patterns that warrant deeper investigation. Most weeks, nothing meaningful changed and the review is fast. Some weeks it surfaces issues worth a half-day of investigation.

Common GA4 mistakes

A few patterns I see in client audits:

  • Multiple GA4 properties on the same site — leftover from migration. Pick one as canonical.
  • No conversion events configured. Site shows traffic but no business outcomes.
  • Wrong timezone. Affects daily comparison data. Set in Admin → Reporting Identity.
  • Internal traffic not filtered — your team's visits inflate metrics. Set up internal traffic filters under Admin → Data Streams → Tagging Settings.
  • Cross-domain not configured — major attribution leak.
  • Server-side GTM not used — for serious analytics, server-side tracking via Google Cloud Run or Firebase is becoming standard.
GA4 isn't difficult. It's different. Spend a day reading the official docs, build the four custom explorations above, set up your conversion events properly, and you'll get more actionable data than Universal Analytics ever provided. The teams that complain about GA4 the most are usually the ones who never moved past the default reports.

Resources to learn deeper

If GA4 will be a major part of your work, these are worth your time:

Final thoughts

GA4 rewards a particular kind of analyst — one who's willing to ask better questions instead of just reading default reports. The platform has more useful data than Universal Analytics ever did, but it doesn't hand it to you. Build the four custom reports above, look at them every Monday morning, and you'll spot trends weeks before competitors who only look at the default Acquisition report.

Need help applying this to your own site? I'm Shani Maurya — a freelance web developer and digital marketer based in Delhi. If you'd like a hands-on audit or full implementation, get in touch — I usually reply within a few hours.