Home / Web analytics audit
Web Analytics Audit / Second Opinion
Your agency's reports:
can you just trust them?
Your GA4/GTM setup, the numbers in your reports, and explanations like “this is as far as the analysis can go.”
A Web analytics audit is an independent analyst checking whether you can take them at face value.
Your first consultation is free, by email
TRY
Your agency’s work: what’s wrong?
Samples of what an agency might send you. Try the monthly report or the chat reply, in either order.
Separate record (Sample Co.)
Inquiry log, September: 620
Inquiry form runs on example.jp
10 things worth questioning. Tap or click anything that looks off
Monthly Analytics Report p.1 Summary
Prepared by XYZ Marketing
Sessions
48,210
+12.4% MoM
Users
31,580
+9.8% MoM
Conversions
1,284
+38.2% MoM
CVR
2.66%
+0.50pt MoM
Bounce rate
41.2%
-2.1pt MoM
*Bounce rate: share of visitors who left after one page
Conversions by day
Conversions vs. last month
User type
| Type | Users | Share |
|---|---|---|
| New users | 24,900 | 78.8% |
| Returning | 9,840 | 31.2% |
| Total | 31,580 | 100% |
Conversions by source (top 6)
| Source / medium | Conversions |
|---|---|
| google / organic | 352 |
| google / cpc | 300 |
| example.jp / referral | 276 |
| (direct) / (none) | 138 |
| yahoo / organic | 60 |
| instagram.com / referral | 33 |
Sessions and CVR by channel
| Channel | Share of sessions | CVR |
|---|---|---|
| Organic Search | 36% | 2.4% |
| Paid Search | 21% | 3.1% |
| Unassigned | 18% | 1.2% |
| Direct | 13% | 2.2% |
| Referral | 8% | 7.2% |
| Organic Social | 4% | 1.7% |
| Average | – | 2.97% |
Users by age⚠ Thresholds applied
| 25-34 | 32.4% |
| 35-44 | 27.9% |
| 45-54 | 20.6% |
*A fictional report created by K.Works
8 things worth questioning, one per paragraph. Tap or click a suspicious sentence
Sample Co., web manager
Thanks as always. Seven questions before our next meeting:
- How far do people read our blog posts?
- What should we name the new conversion? (We already have 「資料請求」, “Request materials.”)
- We want bookings through our booking system (a separate domain) credited to ads
- How many people browse on mobile and then sign up on a computer?
- We want to compare with the same period two years ago
- Is the form tag you fixed last week working?
- Which version of the landing page won the A/B test?
XYZ Marketing, account manager
Thank you for your questions. Here are our answers.
①We’ve turned on scroll tracking in GA4’s enhanced measurement. That shows you how far each post is read. We’ll set up an Exploration so you can view it by post.
②For clarity, let’s name the new conversion in Japanese too, like 「問い合わせ」 (“Inquiry”). Next to the existing 「資料請求」, everyone in your team will understand it right away.
③The booking system has the same measurement ID tag, so bookings are tracked all the way to completion. We’ll report ad-driven bookings at the next meeting.
④We’ll send a user ID to GA4 at login. Then the whole journey of people who browse on mobile and sign up on a computer connects in GA4.
Also, exporting to BigQuery is an option, but it’s pay-as-you-go and gets expensive, so we recommend skipping it for now.
⑤Explorations only go back 14 months, so a comparison with two years ago isn’t possible. We’ll compare with last year going forward.
⑥We confirmed it fires in GTM preview mode, so it’s working fine.
⑦After two days, version B had 8 conversions and A had 5, so let’s switch to B.
Let us know if you have any questions.
*A fictional conversation created by K.Works
01
Does this sound familiar?
Are these
numbers right?
“It’s just how the tool works.”
Really?
Their analyst seems
out of their depth
But it might be rude… so I don’t ask
02
Inside your agency’s analytics team
In many cases, the web analytics team is small
It’s not anyone’s fault. It comes from how the team is set up, and from a downside of new technology.
That’s exactly where an outside eye helps.
03
We already have AI review it, but…
report
“Conversions are up, great.”
Notices only a little
“Broadly reasonable.
This part may be questionable.”
Can make general points
“This part is missing.
Let’s check this next.”
Catches what was missed, and the next step
*Bar length is an illustration of how much each can notice
What AI is missing
Doesn't know your business
Depends on the skill of the person using it
Tends to give safe, noncommittal answers
Misses lesser-known pitfalls
States unfounded guesses as fact
Often suggests things that can’t actually be done
AI’s answers depend on the user’s perspective.
That’s why reviews should be done by experienced people.
04
Sorting out “the tool can’t do that”
Quite often, it’s actually B: “available with setup.”
05
How it typically works
I never change any settings
- Free consultationBy email
- Get accessSame day+
- Investigate1–2 weeks
- Report
- Written debriefIf needed
*The process depends on the scope. Some audits end with the written report and no debrief.
06
The owner handles your audit personally
10+ years
300+ companies
Senior Web Analyst exam passed
Still hands-on
Independent, so lower cost
4 commitments
Facts and evidence only
No blame documents
No settings changed
As a rule, no switch pitch
However, since I handle everything myself, the number of companies I can take on at once is limited.
Plan
Example plans
*Varies with scope and volume
SPOT
Spot audit
For a one-time, clear answer
¥50,000+ / audit
- Audit from tracking setup to reports
- Audit report (verdicts, evidence, questions)
- Written debrief, with Q&A by email
- Answers to your questions for 2 weeks after the report
MONTHLY LITE
Monthly Lite
A second pair of eyes on your monthly reports
¥30,000–50,000 / month
- Monthly report review (once a month)
- Notes on open questions, with questions to send
- Ad hoc questions via chat
- Help interpreting new proposals
MONTHLY STANDARD
Monthly Standard
A monthly second look at your agency's work
¥80,000–150,000 / month
- Everything in Lite
- Review of your agency's meeting materials and notes (monthly)
- Review of specs for new tracking and initiatives
- Quarterly tracking health check
If you're new, start with a spot audit. Decide whether to go monthly after you've seen the results. Pricing depends on site size, number of agencies and volume of reports. I'll send a quote after the free consultation.
FAQ
FAQ
Do I need specialist knowledge?
No. I avoid jargon as much as possible and explain in terms of “what the problem is, how big the impact is, and who to ask for what.”
Do I need to tell my agency before we start?
Not necessarily. If you can grant access within your company, we can start before involving your agency. When and how to share the results with them is your call, and I'll help you do it in a constructive way.
Won't this hurt my relationship with my agency?
I prepare questions worded tactfully. We'll decide together, including whether to tell them a third party reviewed their work.
What if the audit finds no problems?
Being able to say “there are no problems,” with evidence, is itself useful when explaining internally. You can keep working with your current agency with peace of mind.
Can you look at tools other than GA4?
Yes. I also cover Adobe Analytics, BigQuery, Tableau, Power BI, Looker Studio (now Data Studio) and major ad platforms.
Can I use this for in-house reports?
Yes. Even without an agency, you can use it as an audit of your in-house reports and tracking.
Why not have those numbers checked by an independent eye, just once?
Your first consultation is free, by email.
Document
Everything here, in one document
An 18-page PDF covers this page in more detail, for internal discussions and approval. It is currently available in Japanese only. I'm creating English versions one by one, so tell me which document you need and I'll make it first.
This month’s comments
Conversions grew strongly, up 38% month over month. The ad campaigns are clearly working. The late-month decline is seasonal, and we expect a recovery in October.