Google Analytics audit: what one checks, what it finds and what it costs
A buyer's guide to the GA4 audit. The checks that matter, a real account where five of seven recorded leads were staff testing the form, what UK providers charge and how to tell if yours needs one.
The short version
- Published UK prices for a one-off GA4 audit run £850 to £1,500 plus VAT. Almost nobody else publishes a number.
- The findings that change decisions are rarely settings. They are about who is in your data.
- In one live account, five of seven recorded leads were internal QA tests and two-thirds of "direct" was three office devices.
- Plenty of properties do not need a paid audit. The thirty minute self-check at the bottom will tell you.
A google analytics audit is a structured read of your GA4 property to settle one question: are these numbers true enough to make decisions with. The honest answer, before you read any further, is that most live properties contain at least one finding that would change a decision, and that a fair number of businesses do not need to pay anyone to find it.
Price is the other half of the question. Two UK providers publish a figure for a one-off audit, at £850 and £1,500 plus VAT. The rest ask you to fill in a form, and a large group give the audit away because it is a lead magnet for a retainer rather than a piece of work.
Below: what an audit checks, what one found in a real account, what the market charges and how to tell whether yours is worth buying. If you decide by the end that you do not need one, that is a fine outcome.
What a Google Analytics audit actually checks
Most audit checklists online are settings lists. Timezone, currency, data retention, internal filters. Those matter and they are quick, but they are the easy half. The half that changes decisions is about people: who is in your data, where they came from and whether the events you call leads are human beings who want to buy. A GA4 audit worth paying for covers both. Six areas, roughly in the order they cause damage.
01
Who is actually in the data
Internal traffic filters, office IPs, developer traffic, bot filtering, agency sessions. If your own team is a measurable share of traffic, every rate on every report is wrong.
02
Where visits really came from
Referral exclusions, cross-domain setup, payment gateways breaking sessions, UTM discipline and how much of "direct" is really unattributed traffic.
03
Whether key events are real
Key event definitions, duplicate firing, thank-you-page-only tracking, test submissions counted as leads and whether the count matches your inbox.
04
Consent and collection
Consent Mode v2, whether the banner blocks the tag outright rather than degrading it, and how much of your data is modelled rather than measured.
05
Configuration and retention
Data retention, filters left in Testing, unwanted referrals, session timeout, currency and whether anything changed since setup unrecorded.
06
Whether the reporting can be trusted
The Google Ads link, attribution model, BigQuery export and whether the dashboard is reading the fields you think it is.
The settings half takes about an hour. The people half takes reading, and it is the part usually missing from a free scan.
The audit is not the deliverable. The decision you make afterwards is.
what you are actually buyingWhat one audit found in a live account
We built and ran the measurement for an eco-education client: measurement plan first, GTM container, consent mode, server-side verified conversions, the gclid stored against the actual enquiry and a written monthly report. The property looked healthy. Traffic up, leads recorded, ads spending. Three findings changed how that business read every report afterwards.
Five of the seven recorded leads were internal QA tests. Someone had been checking the form still worked, and every check was filed as a lead. Two-thirds of the traffic labelled direct came from three office devices. And the one enquiry attributed to paid search was a customer who had already bought and paid, so that month the ads acquired nobody new.
None of that is exotic. It is the ordinary result of a property that was installed and never read. What made it findable was checking each recorded conversion against the real enquiry instead of trusting the total, the same by-hand method we use on paid search, where every costed search term is judged individually with a written reason. That discipline is in advanced conversion tracking: what it finds that your analytics misses.
A lead in GA4 is a claim, not a fact. Nobody had asked who those seven people were.
totals hide the peopleWhat a GA4 audit costs in the UK
Two UK providers publish a number. KRM Digital lists a GA4 audit at £850 per property, plus £200 to fix what it finds. Its consulting day rate is £1,000. Vixen Digital lists £1,500 plus VAT for a 40 check audit of one property, £2,500 for a 68 check version and £3,000 for the ecommerce version.
Every other operator we checked, including agencies ranking on page one for this term, asks you to fill in a form. That is the market in one line: a small published tier, a large silent tier and a free tier.
| What you are buying | What you actually get | Typical UK price |
|---|---|---|
| Free audit | An automated checklist scan, offered as a lead magnet for a retainer | £0 |
| One-off GA4 audit | A structured check of one property, a written report and a call to walk through it | £850 to £1,500 |
| Extended or ecommerce audit | More checks, ecommerce data health, a dashboard rather than a document | £2,500 to £3,000 |
| Consulting day rate | A day of a senior analyst's time, scoped by you, nothing guaranteed | around £1,000 |
| Audit plus the fixes | The findings and the implementation, not just a list of faults | £1,050 upwards |
| Cellsior GA4 audit | Findings verified against raw data before they ship, in plain English | £495 |
Our own price is £495. What moves it: how many properties are in scope, whether paid media is running and whether you want the findings only or the findings fixed. Auditing one property with no ads is a different job from auditing three feeding a Google Ads account, and any google analytics audit agency quoting one number for both is not looking properly.
Why most free GA4 audits are not audits
Free is the default here. Most UK operators offering a google analytics audit service give it away, because the audit is not the product. The retainer is. That is not dishonest and some free scans are a useful first filter. It does mean the incentives are worth reading, because every UK operator we found selling an audit also sells the work the audit recommends.
The bait
Free trains buyers to expect nothing
When most providers give it away, buyers reasonably conclude an audit is worth nothing. So the free version is built to be cheap: a tool export, a traffic-light sheet, no reading of your conversions.
The conflict
The auditor is bidding for the fix
An audit written by the firm hoping to win the rebuild has a reason to find a rebuild. Almost nobody in the UK offers a straight second opinion on someone else's work, which is what most unhappy clients want.
If the audit is free, the audit is the sales call.
the price tells you the purposeWhen you do not need a paid audit
We would rather say this than sell you something. Skip the paid audit for now if any of these are true.
- You are not spending on ads. The main financial risk of bad analytics is misdirecting media spend.
- Nobody reads the reports. If the numbers do not feed a decision, making them accurate changes nothing.
- You have low volume. Under a few hundred sessions and a handful of enquiries a month, you can check every lead by hand in an afternoon.
- The property was built to a written measurement plan recently by someone who documented it. Audit in a year.
- You already know it is broken and have decided to rebuild. Then you want a GA4 setup agency, not a report confirming what you know. That fee is paperwork.
The exception is money. If you are running paid media the question stops being analytics and becomes: can you tell which pounds worked. Most accounts cannot.
How to check your own GA4 property in thirty minutes
Five checks, no tools, no cost. If they all pass, you probably do not need to buy an audit yet.
Data retention
Admin, then Data collection and modification, then Data retention. Standard properties hold event data for 2 or 14 months and 26 months is Analytics 360 only. If yours says 2, change it to 14. It is not retrospective, so the sooner the better.
Your internal traffic filter
Admin, then Data filters. GA4 creates new filters in Testing state, and a filter in Testing excludes nothing. A correctly written filter never switched to Active is one of the most common things we see.
Count your key events against reality
Match last month's key event count one by one against your inbox or CRM. Not the totals, the individual records. If they differ by more than one or two, stop reading reports until you know why.
Look hard at "direct"
Traffic acquisition report. Direct is not people typing your URL, it is the bucket for traffic GA4 could not label. If it is a large share, check your own office devices first.
Compare GA4 with the ad platform
Google Ads clicks against GA4 paid sessions for the same dates. A gap of 10 to 20% is normal. A gap of half means something is broken between the click and the tag.
If those five turn up nothing, park it. If two or more fail, the next question is whether your conversions are recorded properly at all, which we went through in conversion tracking in GA4: why your analytics are quietly lying to you. For what a finished setup looks like on the other side of an audit, our tracking and reporting page covers the pipeline we build. If the reporting layer is the real problem, how to build a marketing dashboard in Looker Studio is the hands-on version.
Two more if you are shopping around. The paid search version of this page is our PPC audit case study. And if you are weighing a google analytics consultancy against a full-service agency, we listed the options in the best analytics and reporting agencies in the UK, including what each will and will not say about price.
Want to know whether your GA4 numbers can be trusted before you spend another month acting on them?
Ask about a GA4 auditFAQ
What is a Google Analytics audit?
A structured review of a GA4 property to establish whether its numbers can be trusted. It covers who is in the data, where traffic really came from, whether the recorded key events are genuine enquiries, consent and collection, and the configuration underneath. The output should be a ranked list of what to fix and what each fault is costing you.
How much does a GA4 audit cost in the UK?
The two published prices we found are £850 per property from KRM Digital and £1,500 plus VAT from Vixen Digital, rising to £2,500 or £3,000 for extended and ecommerce versions. Consulting day rates sit around £1,000. Most providers publish nothing and quote on scope. Ours is £495.
How long does a Google Analytics audit take?
For a single property, days rather than weeks. The settings checks take about an hour. The part that takes real time is matching recorded conversions against actual enquiries one by one, which is manual, and which is also where the findings that change decisions come from.
Do you need admin access to our GA4 property?
Viewer access to the GA4 property and read access to the GTM container is enough to audit. Admin is only needed to change things. You grant it from your own account in a couple of clicks and can remove it at any time, and you keep ownership of everything.
Is a free GA4 audit worth having?
As a first filter, sometimes. Most free audits in the UK are automated checklist scans offered as a lead magnet for a retainer, so they check settings rather than reading your conversion data. Take one, then judge it on whether a single finding told you something you could act on.
Do we need a GA4 audit or a rebuild?
If you already know the setup is wrong and have decided to replace it, skip the audit and pay a GA4 setup agency to build it against a written measurement plan. An audit earns its fee when you do not yet know whether the numbers are wrong, or when you want an independent read on work someone else did.
Related: white label PPC measurement.