Is Your GA4 Setup Correct? 50 Checks to Verify It

Short answer
Is your GA4 setup correct, and how can you tell?
A correct GA4 setup means the tag fires once on every page, key actions record as one event with the right value, campaign sources carry consistent labels and the ad account connection points to the right property. The 50 checks below test all of that, one area at a time.
So why audit at all? Because a broken GA4 setup does not throw an error. It quietly collects too much or too little data. As a result, the numbers look tidy, and your decisions rest on bad input. Our team sees this often in the field: the GA4 setup "works" but cannot be trusted.
First, we split the audit into eight areas: account, tag, data settings, filters, events, campaign tags, links, and consent plus access. Each area has a numbered list, so you can tick off your progress.
For campaign links, we also built a free helper. You can keep our utm builder open in another tab for the campaign section.
Where do you start a GA4 setup audit?
Start with three preparations. First, confirm that you have editor or administrator access to the property. Second, talk to the person who knows your site code and tag manager. Third, open a sheet where you will record findings.
In the sheet, give every check a status, a proof note and an owner. This simple layout also turns the audit from a chat into a working document. It also makes the next audit much faster.
For live checks, Google's DebugView shows events and user properties as Analytics collects them. However, Google notes that you must turn on debug mode before DebugView displays data.
The official reference for the basics is set up Analytics for a website. It also explains the property, data stream and tag trio.
If you spend money on ads, tracking errors damage bidding directly. So treat this audit as the first step of any google ads management routine. If you prefer to work alone, follow the lists below in order.
1-6: Is your account and property structure correct?
Account and property structure is the base of every report. So a mistake here splits your history when you fix it later.
- Pick one official property per website, because several properties on the same site scatter your numbers.
- Verify the time zone and currency, since mismatches create daily gaps against your ad platform.
- Let a company account own the property instead of one employee's personal login.
- Fill in the industry category and business size completely.
- Use a separate property or a separate data stream for test environments.
- Name the property clearly so anyone can tell which site it measures.
To test this area, first open the admin screen and read the property list. If two properties measure the same site, agree with your team on the official one. Do not delete the old one, because you may need its history later.
The time zone, however, looks small but matters a lot. If your ad account sits in a different zone, daily conversion numbers cut the same day differently. For example, a late-night sale lands on today in one tool and on tomorrow in the other. That is how the question "why do the numbers not match" starts.
7-12: Are your data stream and tag setup correct?
The tag is the eye of GA4. If the eye looks at the wrong place, everything blurs.
- Check that the web data stream carries your real domain.
- Confirm that the measurement ID on the site matches the ID in the data stream.
- Make sure one method loads the tag, because a tag manager plus hard-coded script counts everything twice.
- Test that the tag runs on every template, including login, cart, checkout and thank-you pages.
- Test the same tag on the mobile template.
- Wait about half an hour after setup, then find your own visit in the Realtime report.
A duplicate tag is the most common setup error. You can catch it by searching the page source for your measurement ID, and then counting how often it appears. The network tab of your browser's developer tools also shows how many times the same request fires.
Then open one example of each page type. Specifically, check product, category, contact, cart and thank-you pages. Checkout and thank-you pages often use a separate template, so the tag may be missing there.
13-18: Are your data collection and retention settings correct?
Your GA4 setup also decides privacy and report depth in this area. Settings you never chose on purpose, therefore, tend to surprise you later.
- Choose the data retention setting on purpose, because Google says it affects explorations and funnel reports but not standard reports.
- Review which enhanced measurement events are on, since scrolls and file downloads can crowd your counts.
- If you use Google signals, confirm that they match your consent and privacy notice.
- If you use a user ID, pick one that does not identify a person directly.
- Check that event parameters never carry email addresses, phone numbers or full names.
- Look for personal data inside page URLs and query parameters.
Also, the retention setting sits in the admin area under data settings. You need the editor role to change it, and Google states that a change takes effect after a short waiting period.
Enhanced measurement collects scrolls, outbound clicks, site search, video and file downloads automatically. You will not need all of them. If you never use file downloads, that event only adds noise, so list what you need and review the rest.
Be careful with personal data. If an email address stays in the address bar after a form submission, that value also travels to GA4. In that case, simplify the redirect URL or exclude the parameter instead. This article is not legal advice, so ask a specialist about privacy rules.
19-24: Are your traffic filters and referral settings correct?
Your own visits, developer tests and the trip to a payment provider all distort reports. This area removes that noise.
- Define internal traffic. Google explains that you can mark office IP ranges with a traffic_type parameter.
- Run the data filter in testing state first, then activate it once the result looks right.
- Also keep development and staging traffic out of the live property.
- Add domains that belong to your own process, such as a payment provider or login domain, to the unwanted referrals list.
- If you run more than one domain, set up cross-domain measurement.
- Check whether your own domain shows up as a referral source.
Google's filter out internal traffic page walks through the filter steps. The unwanted referrals page explains that you manage the list in data stream settings and that the list has a size limit per stream.
However, for several domains, read cross-domain measurement. It keeps one person as one user, because it passes the identifiers across your domains.
The referral list matters most on sites with a payment flow. When a visitor leaves for a payment provider and returns, the provider's domain can look like the source. The real source then disappears. As a result, this mistake inflates "direct" and "referral" traffic.
25-32: Are your events and key events correct?
This area is the heart of the audit, because your business goal gets measured here.
- Make sure the key events list reflects your real business goals. Google, for instance, defines a key event as an action that matters most to your business.
- Do not mark every click as a key event, because too many key events hide the real goal.
- Use one naming rule for events, for example lowercase letters and underscores.
- Fire the form submission event only after a successful send, not on page load.
- Check whether phone and WhatsApp clicks are measured.
- In a store, test that the purchase event fires once per order.
- Confirm that purchase value and currency travel with the event.
- Register the custom dimensions you want in reports, or they will not show up.
To test events, first behave like a real visitor. Fill in and send the form, tap the phone button, add a product to the cart and finish an order. At each step, confirm in DebugView or the Realtime report that the event appears once with the right parameters.
An event that merely "works" is not enough. It also must fire at the right moment. If a form event fires as soon as the page opens, your conversion count climbs far above reality. The opposite happens too: after a successful send, a redirect can stop the event from recording.
You can read the key event logic on Google's key events page. For form events, see our guide on form tracking, which event where.
33-38: Are your campaign tags and UTM structure correct?
UTM parameters tell you where traffic comes from. Inconsistent labels, then, split one source into several rows.
- Use the same UTM pattern on all paid links.
- Keep capitalization consistent in source and medium names, because GA4 treats "Facebook" and "facebook" as separate sources.
- Never add UTM parameters to your own internal links, since they break session source.
- Confirm that auto-tagging is on for Google Ads.
- Tag the links in email newsletters and social posts.
- Write the naming rules in a shared sheet.
Example scenario: for a June newsletter you set source "newsletter", medium "email" and campaign "june_sale". To build that link, use our UTM link builder with your page address. For a wider explanation, read what are UTM parameters.
However, the biggest risk with UTM naming is human error. If three people tag the same campaign with three spellings, you see three rows in the report. So the fix is to build links in one tool and keep the rules in one place.
Also decide how auto-tagging and manual UTM values interact in your setup, and write that down. Otherwise the same ad can show up under different sources in your report.
39-43: Are your Google Ads and other connections correct?
Links let GA4 data flow to other tools. Also, a link to the wrong account wastes your time.
- Check that the Google Ads link points to the correct ad account.
- Decide which conversions the ad account uses for bidding and which it only observes.
- If you use Search Console, confirm that its link shows the right site.
- If you need exports, verify that the export links are active.
- Compare event names in other ad platforms with GA4 event names.
To test a link, open the matching screen in both GA4 and the ad account and confirm that you see the same account. Then check that a conversion event reaches Ads. Google notes that you can create an Ads conversion from a GA4 key event and manage its settings in either place.
The split between bidding conversions and observation-only conversions also matters. According to Google Ads Help, conversion actions must be set as primary or secondary correctly, because Smart Bidding relies on those signals. If you set them wrong, the ad system then learns the wrong goal.
We explain gaps between the two tools in why GA4 and Google Ads numbers don't match. For common Ads side errors, see why conversion tracking goes wrong.
44-46: Are your consent and privacy settings correct?
First, consent management decides when and how the tag runs. Also, a wrong setup means you either lose data or collect data without consent.
- Test the consent banner together with your tags, and try accept and reject states separately.
- Confirm that your privacy policy describes your GA4 use.
- Know how the tag behaves for a visitor who declines, and write it down.
Run two scenarios in the banner test. First, accept cookies and watch events flow. Then reject them in a fresh private window and confirm that the tag behaves as you expect. Repeat this test every time you update your banner provider or theme plugin.
This area needs a legal view. This article is not legal advice, so ask a qualified specialist to review your privacy text.
47-50: Are your access and documentation settings correct?
The last four checks, finally, keep the setup from decaying over time. They make the work sustainable.
- Give user roles on a least-privilege basis.
- Remove access for former employees and agencies you no longer use.
- Write a short setup document that says which event measures what and through which trigger.
- Log the date and a short note after every change.
The access list is also a place people forget. Review who holds which role at least twice a year. A departed agency or employee who still has admin rights is a risk for both security and data integrity.
The setup document does not have to be long. A table with event name, trigger, purpose and owner is enough. That table, for example, saves hours when you hand the site to a new developer or agency.
Which checks are critical and which can wait?
However, not all fifty items weigh the same. First fix what breaks conversion data, then what affects report quality, and last the housekeeping items.
| Priority | Items | Why |
|---|---|---|
| Critical | 7-12, 25-31, 39-40 | They directly affect conversion counts and ad optimization. |
| High | 1-3, 19-24, 33-36, 44-46 | They decide report accuracy and privacy fit. |
| Medium | 13-18, 32, 37-38, 41-43 | They deepen your analysis. |
| Housekeeping | 4-6, 47-50 | They keep the setup sustainable and easy to hand over. |
This order is only a start. For a store, the ecommerce items move up. On a service site, the form items move up instead.
For example, a site that runs no ads and relies on organic traffic can rank the link items lower. In contrast, an account that spends budget every month should treat items 39 and 40 as the most critical. The ad system magnifies wrong learning every single day.
When you set priority, ask one question: if this item is wrong, which decision breaks? If the answer is "a budget decision", the item is critical. If the answer is "how a report looks", the item can wait.
How do you score the audit result?
Give every item one of three states: done, partial or missing. Then count how many critical items are done. Even one missing critical item moves to the top of your list.
- Done: you tested it and saw it work correctly.
- Partial: it works but breaks under one condition.
- Missing: it was never set up or it works wrongly.
Instead, the goal of scoring is not a grade. It is a sequence. So write an owner and a date next to every missing item.
Add a "how I verified it" column too. A DebugView screenshot, a test order number or a note about the page source will save time when you retest months later. It also builds trust when you report to a manager or a client.
Which GA4 setup mistakes show up most often?
These mistakes, in practice, repeat in the field. Most of them already appear in the lists above.
- Inflated numbers because of a duplicate tag.
- A form event that fires on every page load.
- Internal traffic that nobody filtered.
- Inconsistent UTM names.
- A purchase event without a value parameter.
When conversions come out missing, read GA4 conversions missing. For unexpected source rows, the GA4 unassigned traffic guide helps.
How often do you re-audit your GA4 setup?
Run a short audit in three situations. After a major site change, after a theme or plugin update and after a new campaign structure. Beyond those, scanning the critical items every three or six months is enough.
A re-audit is far shorter than the first one. You only test the critical items and the parts that changed. Add the results to the first sheet so you can see how things drift over time.
Which tools do you use to test a GA4 setup?
Three tools cover most audits. First, GA4's own DebugView and Realtime report. Second, your browser's developer tools, where the network tab lets you count requests. Third, the preview mode of your tag manager.
- DebugView shows events and parameters as they arrive.
- The Realtime report confirms your own visit quickly.
- Browser developer tools reveal how often the same request fires.
- Tag manager preview shows when each trigger runs.
No single tool is enough. If you confirm the same event in two different ways, your chance of a mistake drops sharply.
What do you do when a check fails?
First measure the effect of the error. If a duplicate tag exists, review conversion numbers for recent weeks and note the date of your fix. That way you do not mix data from before and after the repair.
Make the fix in one go, by one person. If several people change tags at once, you cannot tell which change did what. Test each fix, then move on to the next item.
Remember that a fix does not rewrite history. Mark the faulty period in your reports with a note, so year-on-year comparisons do not mislead you.
How does a GA4 setup audit differ for a small business and a store?
Service sites focus on forms, phone and WhatsApp events. Stores focus on product views, add to cart, checkout start and purchase. So the content of items 25-32 changes with the site type.
In a store, check three things in particular: one purchase event per order, cart value traveling with the right currency and the session staying whole when the visitor returns from the payment provider. On a service site, form success and phone clicks come first.
Campaign tags and links work the same for both types. As a result, more than half of the list fits every site. What differs is which items you call critical.
Which reports do you open first during the audit?
Three reports give a fast first read. First, the events list: do the expected events appear, and are their names consistent? Second, the traffic sources table: is the share of "direct" or "unassigned" suspiciously high? Third, the landing pages report: do you see unknown rows or rows with your own domain?
These reports do not give a verdict, but they show where to look. If the "direct" share suddenly jumps, you start searching for a referral list or UTM problem. That way you jump to the right section of the 50-item list quickly.
Tag manager or hard-coded tag: how do you choose?
Both work. A tag manager lets you change events and triggers without a code release. Hard-coded tags can feel simpler on a small site.
Whichever you pick, use only one. Mixing the two is the most common cause of double counting. We collected what you can measure through a tag manager in what to track with Google Tag Manager.
Which reports can you trust after the audit?
When the critical items are done, you can read conversion and source reports with reasonable trust. Still, GA4 like any data source has sampling, delay and privacy limits.
So build two habits. First, do not base decisions on the last day or two, because conversion delays can change numbers later. Second, treat small gaps between GA4 and the ad panel as normal, and investigate only the large ones.
Then move on to your own analysis. The first metrics to read are in the first 10 GA4 metrics for conversion optimization. To find where conversions leak, try GA4 funnel exploration. To compare attribution views, you can open our attribution model comparison tool.
Should you audit alone or get outside help?
With panel access you can do most of the list yourself. Items that need code, such as event triggers and ecommerce parameters, may need a developer.
When you cannot solve an item alone, working with someone who reads tracking and ads together saves time. As Talha Aslan and team, we check measurement together with the ad account, give findings in writing and explain the reason behind each recommendation. If you also want steady campaign work, look at our google ads agency services.
We do not guarantee results, because results depend on your account data, your budget and your market. A correct GA4 setup is simply the first condition for reading those results honestly. Next, to pick the right metric for your goal, read our sibling guide find your most efficient Google Ads campaign. Understanding how sources group is easier with GA4 default channel group.
Core guide: What Is Digital Marketing? Its types, examples and how it is done, on one page with an interactive film that travels through the brain.



