By Hiran de Silva
Few topics in the Excel world generate as much heat — and as little clarity — as the question of whether Excel is a database.
On one side, IT professionals insist that Excel is not a database.
On the other, Excel influencers argue just as forcefully that Excel is a database.
These positions are presented as mutually exclusive. They are defended passionately. And yet, both fail to explain how real-world, enterprise-scale spreadsheet systems actually work.
This article takes the argument apart and shows why both positions are wrong — not partially wrong, but fundamentally flawed. It then presents a third position that resolves the contradiction entirely.
Why “Excel Is Not a Database” Persists
The claim that Excel is not a database usually comes from IT departments and enterprise architects. Their argument is familiar:
- spreadsheets are messy
- data is duplicated
- files are emailed around
- there is no governance
- no version control
- no single source of truth
They point to real failures in businesses and conclude that Excel is being misused — that what people really need is a database.
On this point, they are largely correct.
But their conclusion is wrong.
What they are actually criticising is Excel being used as a container for data.
When data is stored inside workbooks:
- collaboration breaks down
- consolidation becomes manual
- workflows fragment
- scale collapses
- governance becomes impossible
These problems are real — but they are architectural, not inherent to Excel itself.
Why the Excel Community Pushes Back
Excel professionals push back against this criticism because IT’s proposed alternative is often unusable in practice.
Well-known Excel figures such as Mark Proctor have repeatedly made the same point:
If Excel is removed from the equation, users are often left with:
- long approval cycles
- database administrators
- change-control boards
- governance processes that take months for trivial changes
As a result, Excel users are told:
- “You can’t use Excel”
- “You’ll get a database eventually”
- “This needs a proper system”
The lived reality is that the “proper system” either never arrives or arrives too late to be useful.
This stalemate creates a false binary:
- chaotic spreadsheets
or - heavyweight enterprise systems
That false choice sets the stage for the second mistake.
Why Some Say “Excel Is a Database”
In response, another group argues the opposite: that Excel is a database.
A notable example comes from Christopher T. Fennell, who argued that because Power Query allows users to:
- create tables
- define relationships
- join data
Excel therefore qualifies as a database.
He reinforced this view by pointing to demonstrations from Mynda Treacy, showing how Excel Tables and form-based inputs can be used to maintain structured lists.
At first glance, this sounds persuasive.
But it rests on a fundamental misunderstanding of what a database actually is.
The Wheelbarrow Fallacy
A wheelbarrow has wheels.
A motor car has wheels.
That does not make a wheelbarrow a motor car.
The defining characteristics of a motor car are not its components, but its capabilities, purpose, and architecture.
Likewise:
- tables do not make a database
- joins do not make a database
- relationships do not make a database
Power Query is an ETL tool — Extract, Transform, Load.
Its internal data model exists solely to support transformation design. It is:
- not independently queryable
- not remotely accessible
- not updatable by multiple clients
- not a shared, trusted data store
The data is transient and workbook-bound. It exists to shape data, not to serve it.
That is not a database — even if it looks like one.
A Test That Exposes the Flaw
To move beyond theory, consider this challenge.
A global company’s Chairman wants to create a shared CD catalogue for staff worldwide. The system must:
- allow browsing by title, artist, and genre
- allow staff to submit their own collections
- work globally
- require no administrator
- avoid emailing spreadsheets
- be built using Excel
Most responses are immediate:
“You need a database.”
“Excel can’t do that.”
“This requires a purpose-built system.”
Even those claiming “Excel is a database” have failed to deliver a working solution.
So I built one.
Where the Database Actually Is
The solution uses Excel — exclusively — as the client.
The data itself lives in a relational database, specifically an ACCDB file.
Here is the crucial point that almost nobody realises:
Excel can create, write to, and query an ACCDB relational database directly — without Microsoft Access.
No Access forms.
No Access reports.
No Access licence required.
Excel already contains:
- bi-directional data flow
- table creation
- relational querying
- remote access capabilities
The spreadsheets stop being filing cabinets.
They become interfaces.
The Third Position: The One That Works
This leads us to the position that dissolves the entire argument:
- Excel is not a database
- Excel is not “basically” a database
Excel is a client, designed to leverage a relational database.
This is classic client–server architecture.
In spreadsheet terms, I describe it as:
- hub-and-spoke architecture
- or the Digital Librarian
One trusted data store.
Many Excel clients.
No files passed around.
No ambiguity about “latest version”.
Full scalability and reach.
Change the backend to SQL Server and the same system works globally, beyond corporate networks.
Why This Ends the Debate
Both camps fail because they argue slogans instead of architecture.
- IT criticises Excel because data is trapped in files
- Influencers defend Excel by redefining “database” cosmetically
Neither addresses the system design.
Once Excel is allowed to do what it was always designed to do — act as a client to a relational data store — the contradictions disappear.
Excel does not replace databases.
Databases do not replace Excel.
They work together.
Conclusion
So the verdict is simple:
- “Excel is a database” ❌
- “Excel is not a database” ❌
Excel + a relational database = a real system ✅
That third position is not theoretical.
It exists.
It works.
And it has been quietly available for decades.
If this challenges established beliefs, good. That is exactly what thought leadership is supposed to do.
Add comment