Building on Shifting Sands
Most teams don’t have a data problem.
They have a logic placement problem.
When numbers don’t match between reports, it’s rarely the tools.
It’s where the logic was defined — or where it wasn’t.
Instead of structuring the system, teams fix issues wherever it’s easiest.
And over time, that convenience becomes fragility.
The Question Most Teams Get Wrong
When working with tools like Google Tag Manager, Google Analytics 4, and Looker Studio, teams usually ask:
“Should we solve this in GTM or Looker Studio?”
But that’s the wrong question.
The real question is:
Which layer of your Data Estate should own this logic?
Because that decision determines whether your data becomes:
- a source of truth
- or a collection of interpretations
Your Data Estate Isn’t Flat
Most teams treat their data stack like a collection of tools.
It’s not.
It’s a set of layers — each with a defined role:
- Instrumentation → how data is collected
- Structure → how data is shaped and standardized
- Governance → how meaning is defined and controlled
- Utility → how data is presented and used
When logic is placed in the wrong layer, the system doesn’t fail all at once.
It degrades.
Upstream Logic: Where Data Becomes Truth
Defining logic during collection means the data enters your system already structured.
This is where consistency is created.
- Values align across GA4, BigQuery, and reporting tools
- Logic can be reused across audiences and triggers
- Data is clean before it spreads downstream
But there’s a cost:
- mistakes are permanent
- changes require discipline
This is harder.
It’s also correct.
Downstream Logic: Where Data Gets Reinterpreted
Looker Studio makes it easy to reshape data after the fact.
And this is where systems start to drift.
- Logic exists only within a single report
- Definitions quietly diverge
- Performance degrades over time
It feels fast.
It feels flexible.
But it introduces something far more dangerous:
Multiple versions of the truth.
The Breaking Point Most Teams Hit
At first, downstream fixes work.
You patch issues.
You move quickly.
You get dashboards out the door.
But then:
- Different teams report different numbers
- Definitions stop aligning
- Trust starts to erode
At that point, the problem is no longer visible in any single report.
It exists in the system itself.
This is the moment downstream logic stops being useful.
It becomes a liability.
A Rule That Changes Everything
Trusted data is defined upstream.
Use upstream logic when:
- The definition must persist across tools
- The value represents a business truth
- Consistency matters more than convenience
Use downstream logic when:
- You’re exploring
- Testing
- Or presenting a temporary view
Anything else creates drift.
The Cost of Convenience
Every downstream fix feels small.
But over time, they compound:
- Duplicated logic
- Inconsistent reporting
- Slower dashboards
- Declining confidence
Eventually, you don’t have a reporting problem.
You have a trust problem.
The Principle Most Teams Miss
Looker Studio is not where your data should be defined.
It’s where your data should be revealed.
If your logic lives in your dashboards, your system isn’t structured.
It’s improvised.
The Shift That Actually Matters
Every team goes through this transition:
From:
- Fixing outputs
- Patching reports
- Moving quickly
To:
- Structuring inputs
- Defining logic once
- Building for consistency
That shift is what turns analytics into a system.
Final Thought
You can build impressive dashboards on fragile logic.
They’ll look right.
They’ll feel right.
But underneath, they’ll be sitting on shifting sands.
If This Feels Familiar
If your reports don’t align — or you keep solving the same problem in multiple places — you’re not dealing with a tooling issue.
You’re dealing with structure.
This is the point where most teams realize they need to fix the foundation — not the reports.
That’s the work we do in Elevate — establishing a clean foundation so your data doesn’t need to be constantly reinterpreted.
Because once your system is structured properly, your reporting becomes simpler — and far more trustworthy.
Doug McCaffrey
Designs and maintains analytics systems that remain reliable over time.
Explore how this connects across your data estate: