GA4 for business owners: the 8 reports that matter

GA4 calls them key events. A key event is any event you flag as important, and the flag is what makes a report meaningful. The most common mistake in US small and mid-size accounts is marking the wrong events, or marking too many.

Follow these rules:

1. Flag outcomes, not activity. A purchase, a submitted lead form, a booked call, a phone click from a mobile device, a qualified sign-up. Not scroll depth, not video plays, not page views. If you flag "page_view" or "session_start" as a key event, your conversion rate becomes meaningless and Google Ads will optimize toward noise.
2. Use three to six key events, with one primary. The primary one is what you report to the business and what your ad campaigns bid toward. Secondary ones help you diagnose the funnel.
3. Send value with the event. For ecommerce, the purchase event must carry value, currency and transaction ID; without them the revenue and ROAS views are empty. For lead generation, assign a dollar value to each lead type based on your close rate and average deal size. A $5,000 average sale with a 10 percent close rate means a lead is worth about $500. Approximate is far better than blank.
4. Fire events from one place. Use Google Tag Manager, not a mix of plugins, theme code and hardcoded snippets. Duplicate tags are the leading cause of double-counted conversions.
5. Confirm with DebugView and a test transaction. Watch the event arrive with its parameters before you trust a single report.
6. Match event names across Google Ads. Import key events into Google Ads as the conversion actions you bid on, and stop using the separate Google Ads tag for the same action, or you will count each conversion twice.

Expected outcome once this is correct: a US ecommerce store typically sees a purchase conversion rate of 1 to 3 percent and a lead-generation site 2 to 5 percent on form submissions. If you see 20 percent, you are flagging the wrong event.

GA4 has dozens of reports. These eight cover what a business owner or marketing lead actually needs, and each has a decision attached.

1. Traffic acquisition (by session default channel group)

Found under Reports, Acquisition. It shows sessions, engaged sessions, key events and revenue by channel: Organic Search, Paid Search, Paid Social, Email, Direct, Referral. The decision: which channels produce key events, not just visits. A channel with 40 percent of sessions and 5 percent of conversions is a budget problem.

Watch for a large "Unassigned" or "Direct" share. Over 25 percent usually means missing UTM parameters on email, SMS and social links, or a consent setup that is dropping attribution.

2. Landing page report

Reports, Engagement, Landing page. Which pages people enter on, their engagement rate, and key events per entry. The decision: which pages deserve conversion work and which content is pulling qualified traffic. A landing page with high entries and an engagement rate under 40 percent is a candidate for redesign; see how we approach that in UX, UI and conversion design.

3. Pages and screens

Same section. Views, users, average engagement time and key events per page. Use it to find your highest-value pages (pricing, services, product categories) and check that they are linked from the home page and navigation.

4. Key events report

Reports, Engagement, Key events. Each flagged event, its count, its value and the channel breakdown when you add a dimension. Review it weekly. Any sudden drop usually means a broken form, a changed thank-you page or a tag that stopped firing after a site update.

5. Ecommerce purchases (or lead funnel exploration)

For stores: Reports, Monetization, Ecommerce purchases shows items viewed, added to cart and purchased with revenue by item. Typical US view-to-cart rates are 5 to 10 percent and cart-to-purchase 25 to 45 percent. For lead-gen businesses, build a funnel exploration instead: landing page, service page, form start, form submit. The drop between form start and submit is where the money is.

6. User acquisition, first user source

Different from report one. It credits the channel that first brought a user, not the channel of the converting session. Compare the two: if Paid Social is high on first touch and low on last touch, it is doing introduction work that last-click reporting hides. This is the beginning of attribution thinking, which we cover in marketing attribution models explained.

7. Tech details: device category and browser

Reports, Tech. Mobile is 55 to 75 percent of US consumer traffic in most categories, yet converts at roughly half the desktop rate. If your mobile conversion rate is below a third of desktop, you have a mobile experience problem, not a traffic problem. Check browser too: a sudden Safari drop often means a cookie or consent change broke tracking.

8. Advertising, attribution paths

Under Advertising. Conversion paths show how many touches precede a conversion and which channels appear early versus late. Most considered purchases in the US take two to five touches over one to three weeks. If your budget is 90 percent on channels that only appear last, you are underfunding demand creation.

Spend thirty minutes in Admin and fix these:

• Data retention is set to 2 months by default. Change it to 14 months, or your year-over-year explorations will have no data.
• Internal traffic is not filtered. Define your office and agency IP addresses and activate the filter. Your own team can be 5 to 15 percent of a small site's traffic.
• Unwanted referrals. Add your payment processor domains (for example the hosted checkout you use) so a returning buyer is not credited to "Referral."
• Google Signals and consent. Enable Consent Mode v2 if you run Google Ads, so modeled conversions fill gaps left by opt-outs. Review Google Signals carefully under state privacy laws such as CCPA/CPRA; many US businesses now keep it off.
• Cross-domain tracking. If your booking tool, store or quote form lives on another domain, configure the domain list. Otherwise every conversion appears as a referral from your own vendor.
• Link Google Ads, Search Console and BigQuery. The BigQuery export is free, and it is your insurance policy for the day you need to join analytics data with CRM or order data.
• Session timeout and engaged session definition. The 10-second engaged session default is fine; just do not compare your engagement rate to the old bounce rate.

Decisions come from rhythm, not depth. Every Monday:

1. Key events, last 7 days versus previous 7 days. Investigate any swing over 20 percent.
2. Traffic acquisition with key events and revenue columns. Note which channel moved.
3. Landing pages, sorted by entries. Spot new pages gaining traffic and pages whose engagement fell.
4. Tech details, mobile versus desktop conversion rate.
5. One question for the team, written down, that the data raised.

Monthly, add attribution paths and the first-user versus session-source comparison. Quarterly, compare GA4 revenue to your commerce platform or CRM and recalibrate the lead values.

• Treating GA4 as the source of truth for revenue. It is directional. Your order system or CRM is the ledger.
• Chasing total users. Sessions and users are inputs, not results.
• Changing key events without documenting the date. Every change creates a step in your trends that someone will misread six months later.
• Letting a web agency and an ad agency each install their own tags. One container, one owner.
• Ignoring the consent banner's effect. After a compliant banner goes live, reported traffic typically falls 10 to 30 percent while real traffic does not change. Annotate it.

We treat analytics as part of the growth system, not a reporting chore. The sequence is discover, plan, build, improve.

Discover. We audit the existing property: key events, tag container, consent setup, data retention, cross-domain configuration and the gap between GA4 and the commerce or CRM numbers. We interview the owner about the three decisions they need the data to support.

Plan. We write a measurement plan on one page: business goals, the key events that represent them, their dollar values, the parameters each event carries and the reports that will be reviewed weekly and monthly.

Build. We rebuild the tag container, implement the events, configure Consent Mode and the privacy settings that US state laws require, link Google Ads and BigQuery, and set up a dashboard that pulls GA4 alongside platform spend and CRM outcomes. The engineering side of this connects to our systems integration work when CRM or order data is involved.

Improve. We run the weekly routine with the client's team for the first quarter, then hand it over with documentation. Insights feed straight into performance marketing decisions, because measurement that does not change spend is just overhead.

A representative engagement is our Performance Growth Sprint. The client's paid media was being judged on platform-reported conversions that GA4 contradicted, and nobody trusted either number. We reset measurement first: defined a small set of key events with values, rebuilt the tag container, reconciled GA4 against the order system and only then restructured campaigns and creative testing around the agreed numbers. The pattern is consistent across our work: fix what is counted before changing what is spent.