Ad platform
Attributed conversionWhich attribution window?Data Health Check
Trust starts with a check
Trace where each number comes from, what it means, who owns it and how long it will last. Four signals tell you what to trust, what to fix and where to start.
See our approachDOES THIS SOUND FAMILIAR?
Does this sound familiar?
Traffic, conversions and orders each tell a different story. Before fixing anything, trace where the number came from, what it means, who owns it and how long it will last.
ONE WORD. DIFFERENT MEANINGS.
All called “conversion”. Not all the same thing
GA4
Recorded eventWhen does it fire? Any duplicates?CRM
Qualified leadWhat qualifies it?Understand the difference before choosing what to fix.
Illustrative concepts, not a diagnosis of your accounts.
FOUR SIGNALS
When numbers disagree, ask four questions first
There is no overall score. Each of the four areas gets its own signal, plus a written interpretation: what to trust, what to fix, what comes first. Every signal comes with its evidence, so a different reader can see why it is that colour. Signals reflect the scope and state at the time of the check; items with insufficient data or not yet verified are marked separately.
01 / SOURCE
Can this number be traced back?From the report back to where it was first recorded, and how many hands in between.
02 / MEANING
Is everyone counting the same thing?How many definitions and calculations one word has inside the company.
03 / OWNER
Who finds out first when a number is wrong?Not only who notices, but who confirms it, who fixes it and who tells the people affected.
04 / SHELF LIFE
Can this number still be trusted next month?When process, setup or definitions change, when it gets re-checked and by whom.
OUR APPROACH
A clear question. A practical approach
- 01
Trace how many hands each number passes through and where it forks
- 02
Find how many definitions one word has inside the company
- 03
Check who looks after each field, and how long since anyone did
WORKING TOGETHER
Useful beyond the handover
These are starting points for scoping. Deliverables and timelines are agreed for each engagement.
01
A health report with four signals
02
What each risk means for the business
03
Which fix comes first
SCOPE
What the check covers
Tag installation
- Whether tags and events fire correctly
- Duplicate or missing events
Configuration
- Settings sorted into healthy, needs work and broken
- Quick fixes separated from scheduled work
What you receive
- A GA account health report
- Official references for every item, so owners can fix them
This service at a glance
| Item | Details |
|---|---|
| Typical situation | Traffic, conversions and orders each tell a different story. Before fixing anything, trace where the number came from, what it means, who owns it and how long it will last. |
| How we work | Trace how many hands each number passes through and where it forks; Find how many definitions one word has inside the company; Check who looks after each field, and how long since anyone did |
| What you receive | A health report with four signals; What each risk means for the business; Which fix comes first |
| Before we start | The numbers that disagree, their sources and any recent site or tracking changes. |
OFFICIAL REFERENCES
References: Google documentation
Our approach follows Google’s official documentation. When the rules change, the latest official guidance applies.
- Monitor events in DebugViewsupport.google.com
- Consent mode on websites and mobile appssupport.google.com
- GA4 data freshnesssupport.google.com
BEFORE WE BEGIN
A useful place to start
The numbers that disagree, their sources and any recent site or tracking changes.
Discuss your situationCan we start without knowing the root cause?
Yes. Describe the confusing numbers or workflow, and we can discuss whether diagnostics, analysis or another service is the right starting point.
Do we need to change tools first?
We first review the current setup and constraints. A new tool should follow a clear need.
How are scope and timing agreed?
We scope the data, systems, participants and deliverables together before proposing the engagement.
COMMON QUESTIONS
When the numbers don't match
Our back end shows the orders, but GA4 reports fewer conversions. What usually causes this?
Five causes come up most often:
- the same order is recorded twice, for example when the thank-you page is reloaded;
- visitors returning from a payment page are counted as a new referral;
- checkout runs on another domain without cross-domain measurement, so the order loses the visit that led to it;
- visitors decline tracking or block ads, so the event is never sent;
- the two systems define an order differently, for example the back end includes cancellations and returns while GA4 counts the moment of purchase.
The size of the gap is not the answer. Find out which of these it is first.
Conversions suddenly dropped. What should we check first?
First work out whether the business changed or the recording broke. Check whether the site, the checkout flow, the payment provider or GTM changed on the day of the drop, then compare GA4 purchase events with back-end orders day by day to see when they start to diverge. If only GA4 dropped and the back end looks normal, tracking usually broke; sales did not.
Does GA4 remove duplicate orders on its own?
Yes, with conditions. On web data streams, if two purchases carry the same transaction ID, Google Analytics removes the duplicate. If no transaction ID is sent, or a different one is sent each time, nothing is removed. App data streams do not use this mechanism.
How we check
- Agree on definitions: how orders, payments, cancellations and returns are counted.
- Compare daily back-end orders with purchase events and find the date they start to diverge.
- Check transaction IDs, duplicate triggers and payment-provider referrals.
- Check cross-domain and consent settings.
- Write a fix list: who changes what, and how each fix is verified.
Sources for these answers
- Google Analytics Help: Minimize duplicate key events with transaction IDs
- Google Analytics Help: Identify unwanted referrals
- Google Analytics Help: Set up cross-domain measurement
- Google Tag Platform: Consent mode overview