Analytics
Google Analytics Audit: Which Numbers to Trust and What to Fix First
I go through your existing Google Analytics 4 property setting by setting and event by event, test it in the browser, and compare what it reports with what your CRM or store says actually happened. You get a clear verdict on data quality and a fix list in the order that matters.
- years in SEO
- 6+
- client websites worked on
- 60+
- person doing the work
- 1
Signs your property needs a Google Analytics audit
A Google Analytics audit is a manual review of the GA4 property you already have - settings, events, key events, consent behaviour and attribution - tested in a real browser, checked against your own sales records and finished with a fix list ranked by impact. You probably do not need anyone to tell you the data is off. The usual trigger is a moment of doubt that nobody can resolve: the sales team says enquiries are down while GA4 says they are up, a board member asks why revenue in the report is lower than the finance figure, or a new marketing manager opens the property and finds dozens of events with names like test_click_2.
Other signs are quieter. A big slice of traffic sits in Unassigned or (not set). Your payment provider appears as a top referral source. Explorations stop at two months back. Organic search looks flat in GA4 while Search Console shows clicks climbing. Each of these has a known cause, and most take minutes to fix once someone finds them.
The audit suits businesses that make budget or channel decisions from Google Analytics - service firms counting leads, online stores counting revenue, B2B companies judging which campaigns start conversations. If GA4 is installed but nobody uses it for anything, the more useful conversation is about what you want it to answer, and that is closer to a GA4 setup project.
Key takeaways
- An audit is for a property that already exists and already collects data. If there is nothing worth keeping, a rebuild is cheaper than a review.
- The most useful check is reconciliation: GA4 figures compared against orders, enquiries or bookings recorded somewhere else.
- Most findings are configuration, not code - retention, filters, referrals, key event choices - and can be fixed quickly once found.
- Every issue in the report carries evidence, an estimate of how much it distorts decisions, and the specific fix.
- Automated checkers read settings. They cannot tell whether a key event fires on the right action, which is where most of the damage hides.
What's included in my Google Analytics audit
The audit is organised into seven areas. Each is reviewed in the admin settings, in the reports and, where behaviour matters, live in the browser with Tag Assistant and DebugView. The report only shows the checks that failed.
Property settings and internal traffic filters checked against how you sell
Time zone and currency have to match your finance reports, data retention decides how far back explorations can reach, and session timeout should suit how long your buyers take to decide. Internal traffic rules only work if the data filter is set to Active; a filter left in Testing removes nothing, so staff visits keep inflating the pages they open every day.
- Time zone, currency and session timeout
- Event data retention and what it limits
- Internal and developer traffic rules and filter state
- Enhanced measurement options per data stream
Event inventory and key events tested on the action they claim to count
Every event name and parameter goes into an inventory, so duplicates and one action tracked under three names show up straight away. Then each key event is triggered in a real browser. The usual faults are GA4 installed twice - once by a plugin, once through Tag Manager - key events firing on a button click even when the form fails, and automatic form events catching newsletter boxes and search fields. For stores, the purchase event is checked on each payment method you offer.
- Duplicate tags and double-counted page views
- Key events fired on confirmed outcomes, not clicks
- What automatic form events catch on your templates
- Purchase event and revenue on every payment route
- Counting method per key event
Source and cross-domain checks that stop sales losing their true origin
When a visitor moves to a payment page, booking widget or third-party form on another domain, the session can restart and the sale gets credited to the provider or to direct. I trace each of those hand-offs, review how much traffic sits in Unassigned and (not set), and check your campaign links for naming drift that pushes email or social visits into the wrong channel.
- Unwanted referral list and self-referrals
- Cross-domain settings and linker behaviour
- UTM conventions and Unassigned traffic
- Booking, checkout and portal subdomains
How consent choices change what your tags collect, tested both ways
With the cookie banner live, I accept and then reject consent on desktop and mobile and watch what fires. The common faults are analytics loading before the visitor has chosen anything, or never loading after they accept. I also check whether consent signals reach the Google tags at all. The audit reports what happens and how big the resulting gap looks; whether your banner and policy meet the law is for your legal adviser.
- Tag timing before and after the banner
- Consent signals passed to Google tags
- Accept and reject journeys on both devices
- Estimated size of the gap consent leaves
Product links, attribution settings and custom definitions inside the property
Linked products shape what GA4 can tell you. I confirm the Google Ads link and auto-tagging, the Search Console link and whether its reports are actually published to the report library, and any BigQuery export. Custom dimensions are checked for whether they are registered at all and whether high-cardinality values are pushing useful rows into (other). User-ID, Google signals and reporting identity are reviewed against your consent approach.
- Google Ads link and auto-tagging
- Search Console link and published reports
- BigQuery export where one exists
- Registered custom dimensions and cardinality
- Identity and attribution settings
Ownership, user access and personal data that should never reach GA4
Two findings here can turn urgent. One is admin rights held by a former employee's personal login or an agency you stopped paying years ago. The other is personal data stored in the property - typically a form plugin that puts names or email addresses into the thank-you URL, which GA4 records as page location. Google's policies prohibit sending personal data to Analytics, so I flag the exact source and how to stop it.
- User list, roles and account ownership
- Former staff and agency access
- Email addresses or names in URLs and parameters
- Data redaction settings on web streams
Reconciliation with your own records and a fix list ranked by impact
The most telling number in the report is usually the gap between GA4 and your CRM, store or booking system, month by month. Every failed check is written up in plain language with its evidence, a severity rating based on how much it distorts decisions rather than how technical it sounds, and the specific fix. We go through it together so your team can decide what to handle in-house.
- Reconciliation sheet: GA4 against your own figures
- Evidence per finding: screenshots, DebugView captures, report extracts
- Current-state inventory of events, key events and links
- Recorded or live walkthrough of the findings
Experience
Google Analytics Audit from someone who has done it many times
You work with me from the first call to the last report. No account managers, no hand-offs to a junior team.
Book a free call- years in SEO
- 6+
- client websites worked on
- 60+
- documented case studies
- 35
Typical GA4 audit findings and where each one gets fixed
The tabs above cover what gets checked. This table covers what happens next: where common findings are fixed and roughly how much work each is.
| Finding | Fixed in | Usual effort |
|---|---|---|
| Retention left at the two-month default | GA4 admin | Minutes; only helps data collected from that day on |
| Internal traffic filter left in Testing | GA4 admin | Minutes, once the office IP or traffic rule is confirmed |
| One action tracked under three event names | Tag Manager | A few hours, plus updating any reports built on the old names |
| Key event fires on click even when the form fails validation | Tag Manager, often with a data layer push from a developer | A small developer task with a written brief |
| Checkout provider credited with sales that started in search | GA4 admin: unwanted referrals and cross-domain settings | An hour or two, then a test purchase to confirm |
| Analytics loads before the cookie banner | Tag Manager consent settings or the banner configuration | Short, but needs sign-off from whoever owns the policy |
| Email addresses in page URLs after a form submission | Form settings or a developer change, plus redaction | Small change; past data is hard to clean |
| A former agency still holds admin access | GA4 account access | Minutes, once ownership moves to a company login |
Effort is indicative, not a quote. The report marks which fixes your team can make in the admin today and which need a developer.
How the audit runs, start to finish
Elapsed time is usually one to two weeks, mostly decided by how fast access and the comparison data arrive. The work itself happens in this order.
- 1
Agree what the data is for
A short call about the decisions GA4 informs today: channel budgets, landing page changes, sales targets. An audit without this turns into a list of technicalities with no sense of which ones matter.
- 2
Get access and comparison data
Viewer or analyst access to GA4, read access to Tag Manager, Search Console and ad accounts, plus a monthly export of enquiries or orders from your CRM, store or booking system for the same period.
- 3
Review configuration
Every admin setting and product link checked against how your business works - not against a generic best-practice list. A setting that is wrong for a store can be right for a law firm.
- 4
Test behaviour in the browser
I walk the key journeys on desktop and mobile - submitting forms, completing a test purchase where possible, accepting and rejecting consent - and watch exactly what fires and when.
- 5
Reconcile against reality
GA4 key events and revenue lined up against your own records month by month. The size and direction of the gap is often the single most useful number in the report.
- 6
Write up and walk through
Findings ranked by impact, each with evidence and a fix. We go through them on a call so your team can ask questions, decide what to do in-house and what to hand over.
Free review
Get a free Google Analytics Audit review
Send me your site. Before the call I look at your rankings and pages, then tell you what I would fix first and what it is likely worth.
A manual audit versus an automated GA4 checker
Free audit tools have a place. They read the admin API and flag settings that differ from a default. What they cannot do is see your site.
Automated checker report
- Score out of a hundred based on settings it can read
- Flags retention, filters and links as present or absent
- Cannot tell whether a key event fires on success or on click
- No comparison with your CRM, store or booking data
- Same recommendations for every business
- No view on which issues actually change decisions
What my audit adds
- Every key journey tested in a real browser, consent states included
- Each finding backed by a screenshot, event payload or report view
- Key events checked against the outcome they claim to represent
- Month-by-month reconciliation with your own sales records
- Settings judged against how your business sells
- Fixes ranked by how badly each issue distorts reporting
Where SEO meets CRO
Organic search gets audited too, not just the ads
Part of the audit looks specifically at how search traffic is measured: whether the Search Console link works and its reports are published, whether organic landing pages can be tied to key events, and whether payment or booking steps are stealing credit from search. Those are the numbers an SEO audit or any conversion work needs before it can say which pages earn money.
What is in scope and what is not
The audit reviews an existing GA4 property and the tags that feed it. It includes a look at the Tag Manager container insofar as it affects GA4 data, but a full container cleanup is its own job under Google Tag Manager. Consent works the same way: I test whether tags respect the visitor's choice and report the gap, fixing consent settings tag by tag happens in the container work, and whether your banner meets the law is a question for your adviser, not me.
It is not a rebuild. If the audit shows the property cannot be rescued - wrong account structure, years of polluted data, no event worth keeping - I will say so and recommend starting fresh with a GA4 setup rather than patching. It is also not a technical SEO review; crawl, index and speed issues belong to technical SEO.
From you I need access, the comparison export and around an hour of someone's time across the project for questions. Where a developer is involved in fixes, a named contact helps.
After the audit
Some clients take the fix list and work through it in-house; it is written so they can. Others hand parts of it back to me. The most common follow-on is a Google Tag Manager project, because so many GA4 faults trace back to a container that several agencies have worked in. Where the audit shows the property is not worth repairing, the next step is a GA4 setup built from a written measurement plan.
Repairs to lead and purchase events, cross-domain journeys and consent settings sit inside those two services rather than being sold separately. Once the numbers hold up, they become the baseline for judging SEO work by enquiries and revenue rather than traffic. For how the three services connect, see the Google Analytics consulting overview.
Growth from real projects
Learn more about google analytics audit
Google Analytics Audit FAQs
How long does a GA4 audit take?
Usually one to two weeks from the day access arrives. The checking itself is a few days of work; the rest is waiting on comparison data, test purchases or answers about how a particular form or booking flow is supposed to behave. Very large properties with several data streams take longer.
What access do you need for a Google Analytics audit?
Viewer or analyst access to the GA4 property is enough for the review itself. Read access to Tag Manager, Search Console and any linked ad accounts helps me trace problems to their source. I do not need edit rights unless you want me to make fixes afterwards.
Will the audit fix historical data?
No. GA4 does not let you reprocess data that has already been collected, so fixes apply from the day they go live. The audit does tell you which historical figures are reliable, which are inflated or understated, and roughly by how much, so you can read old reports with that in mind.
Do you audit Tag Manager at the same time?
I review the tags, triggers and variables that send data to GA4, because many GA4 faults start there. A full review of every tag in the container, including ad pixels and third-party scripts, plus restructuring and cleanup, is a separate Tag Manager project.
Can you audit a property used by an agency we still work with?
Yes. The audit is evidence-based, so it is useful to the agency as well as to you. Most agencies welcome a clear list of issues with proof attached. I share findings with whoever you choose and keep the tone factual, since the aim is accurate data rather than blame.
Is the audit worth it if we only have a small site?
It depends on what rides on the numbers. If you spend on ads or make hiring decisions from enquiry volume, inaccurate tracking costs more than the audit. If GA4 is something you glance at occasionally, a short call to check the basics is probably enough, and I will tell you that.
Not sure your data would pass a Google Analytics audit?
Grant viewer access and book a free 30-minute call. I will spend time in the property beforehand and bring the most serious issues I can see, so you can judge whether a full audit is worth doing.