How the democratisation of technology changed the Excel conversation — and perhaps caused us to forget what Excel could already do
By Hiran de Silva
I came across a post recently making a familiar argument about Excel.
The argument, essentially, was that when you import data into Excel, what you have is a snapshot. By contrast, connect to something like Microsoft Business Central and suddenly your data can be live.
And I found myself asking a very simple question:
Couldn’t we always do this with Excel?
Not last year.
Not since Microsoft 365.
Not since Power Query.
For something approaching three decades.
That question led me somewhere rather more interesting.
Because I think Microsoft has a problem.
And, strangely enough, the best way I can explain Microsoft’s problem is with the Flat Earth.
Imagine the Flat Earth Became Mainstream
Most people accept that the Earth is a globe.
But imagine, for a moment, that something changes.
Flat Earth believers become much better at communicating their ideas. Social media gives them a voice. Algorithms amplify them. Their audience grows.
People who previously had no particular opinion on the shape of the Earth are exposed repeatedly to Flat Earth arguments.
Gradually, Flat Earth becomes a significant demographic.
At that point, something else happens.
The market notices.
Marketing departments don’t necessarily exist to decide whether the Earth is flat.
They exist to understand what people believe, what people want, what people fear and what people might buy.
So if enough people become worried about falling off the edge of the Earth, somebody will eventually start selling them parachutes.
Then somebody else will sell bungee cords.
Consultants will appear.
Training courses will appear.
Books will appear.
Influencers will appear.
Perhaps there will be conferences discussing the safest methods of approaching the edge.
Eventually, an entire industry can emerge around solving problems that exist only because of the original mental model.
And once that industry becomes sufficiently large, something extraordinary has happened.
The market is no longer merely responding to the belief.
It is reinforcing it.
That is the analogy I want to use to talk about Excel.
The Excel World I Remember
Go back approximately 30 years.
The professional computing world was very different.
A laptop was an extraordinary investment.
In the early 1990s, I had a Toshiba portable computer costing somewhere around £4,000.
A few years later, when I was considering becoming more mobile as an independent Excel professional, I looked seriously at buying an IBM ThinkPad costing around £7,000.
That was in 1996.
Think about what those numbers mean in today’s money.
Computers were not yet ubiquitous consumer devices.
And the people doing serious work with Microsoft Office included a relatively small population of professional developers, consultants and power users.
The conversation was correspondingly different.
I remember attending Microsoft’s quarterly briefings.
I remember Microsoft’s newsgroups.
I remember what we were excited about.
One of the enormous developments was connectivity.
Excel could talk to databases.
ADO opened an extraordinary door.
You could have data held centrally in a relational database and Excel acting as a client.
Excel could GET data.
Excel could PUT data.
Multiple Excel users could interact with the same central source of truth.
In other words, you could build a client-server architecture with Excel.
What I now describe as Hub and Spoke.
Or, in the terminology I use today, the Digital Librarian.
The spreadsheets did not have to contain the enterprise’s data.
They could simply be windows onto it.
That distinction is enormously important.
And we were doing it decades ago.
Then the Demography Changed
During the 2000s, something profound happened.
Computers became cheap.
Laptops became cheap.
Microsoft Office became ubiquitous.
Then came widespread internet access.
Mobile devices.
Cloud computing.
And, critically, social media.
The technology industry was democratised.
Millions upon millions of people who would never previously have regarded themselves as professional computer users suddenly had access to extraordinarily powerful technology.
That was overwhelmingly a good thing.
But it also changed the demography of the conversation.
And when the demography changes, the market changes with it.
This is where my Flat Earth analogy matters.
The underlying technology did not suddenly stop existing.
But the audience changed.
And increasingly, products had to be explained, packaged and marketed to a vastly larger audience that had never experienced the professional computing world from which technologies such as Excel had emerged.
At the same time, another powerful force appeared.
Everything had to be new.
The Tyranny of “New”
Technology has a peculiar marketing problem.
If something is 30 years old, people assume it must be obsolete.
We even have a wonderfully dismissive word for it:
Legacy.
But many business problems are not new.
Consolidation isn’t new.
Budgeting isn’t new.
Forecasting isn’t new.
Collecting information from hundreds of people isn’t new.
Management reporting isn’t new.
Controlling access to information isn’t new.
Audit isn’t new.
The requirement to get information from one place to another isn’t new.
So why should we assume that every solution to those problems must have been invented recently?
Nobody says:
“Dentists are legacy technology. We’ve had them for centuries.”
The problem continues to exist, so we continue to use the solution that works.
Yet technology doesn’t work like that.
Technology must continually produce something that looks new.
Microsoft itself was advancing enormously.
The internet infrastructure developed.
Cloud computing developed.
Data centres developed.
The physical location of computing changed.
The scale changed.
The interfaces changed.
And Microsoft quite properly developed new products to exploit all of this.
But meanwhile, something curious happened to the technologies we already had.
We began forgetting what they could do.
The Cloud Is Not Magic
Consider the cloud.
When somebody says:
“Our data is in the cloud.”
Where is the data?
It hasn’t evaporated into the atmosphere.
It is on a physical computer somewhere.
Probably in a data centre.
Probably in a database.
What makes it appear to us as “the cloud” is an extraordinary infrastructure of networks, protocols, servers and software that enables physical computers to communicate.
“The cloud” is a useful mental model.
So is “the internet”.
But behind those abstractions are computers communicating with computers.
And that is precisely why an idea that was powerful 30 years ago remains powerful today:
client-server architecture.
The geography changed.
The principle didn’t.
Then Came Power
Microsoft has always been rather fond of branding generations of technology.
At one time, everything seemed to be Visual something.
Then Active something.
Then came Power.
Power Pivot.
Power BI.
Power Automate.
Power Apps.
And, of course:
Power Query.
Power Query is an extraordinarily capable technology.
But conceptually, what is it doing?
It is getting data.
Transforming data.
And loading data.
ETL.
Extract.
Transform.
Load.
Before Power Query, we had Microsoft Query.
The capabilities changed enormously, of course. Power Query is vastly more sophisticated.
But the fundamental requirement — connecting to external data and bringing it somewhere useful — was hardly invented yesterday.
The interesting part is what happened next.
Social media discovered Power Query.
And Power Query became whatever people could make it become.
It is so flexible that it is almost like Play-Doh.
You can make an elephant.
You can make an aeroplane.
You can make a house.
And every new creation becomes another tutorial.
Another video.
Another course.
Another clever solution.
Another piece of content.
That is excellent for engagement.
But engagement and business architecture are not necessarily the same thing.
Extract. Transform. Load. Where?
There is a question I think we don’t ask nearly enough.
If you are using an ETL tool, where are you loading the data?
Suppose you extract information from several sources.
You clean it.
Transform it.
Combine it beautifully.
And then you load the result into an Excel workbook.
Fine.
If one person needs that workbook, perhaps that is exactly the right solution.
But now suppose 100 people need the information.
Or 400.
Or 10,000.
What happens next?
Do we email the workbook?
Put copies in folders?
Create multiple versions?
Build elaborate processes for synchronising them?
At that point, I would ask a much more fundamental question.
Why did we put the shared data inside a file in the first place?
If the purpose is collaboration, shouldn’t the ETL process load the data somewhere centrally accessible?
Then every spreadsheet that needs the information can GET it from there.
And when appropriate, every authorised spreadsheet can PUT information back.
Now the spreadsheets don’t have to travel.
The data travels.
That isn’t a new idea.
That is client-server architecture.
And Excel has been capable of participating in that architecture for decades.
How Did We Forget This?
This, I think, is where the populism argument becomes interesting.
The Excel conversation increasingly came to be defined not by enterprise architecture, but by what could be demonstrated conveniently to enormous audiences.
And those are two very different things.
A YouTube demonstration needs to work for somebody sitting at home with a copy of Excel.
A LinkedIn post needs to make sense in seconds.
A training course needs to be reproducible by thousands of people who don’t have access to corporate infrastructure.
Naturally, therefore, the workbook becomes the universe.
The data is in the workbook.
The transformation happens in the workbook.
The reporting happens in the workbook.
The solution is demonstrated in the workbook.
And when collaboration becomes necessary, we start moving the workbook around.
Eventually, that becomes our mental model of Excel itself.
The social-media world has, in effect, reinvented Excel in its own image.
And then something even stranger happens.
We begin blaming Excel for the consequences.
The Spreadsheet Problem
We’ve all heard the criticisms.
Too many files.
Too many versions.
Email attachments everywhere.
Broken links.
Consolidation nightmares.
Data trapped in spreadsheets.
Nobody knows which version is correct.
And the conclusion is invariably:
Spreadsheets are the problem.
But wait.
What if Excel never required us to architect the process that way?
What if the spreadsheet didn’t need to carry the data from Alice to Bob?
What if Alice and Bob could always have been clients of the same centrally accessible data source?
What if hundreds of spreadsheets could all GET from and PUT to a common relational database?
Then perhaps the problem isn’t Excel.
Perhaps the problem is our mental model of Excel.
And perhaps that mental model has been shaped by the demographic and commercial forces surrounding Excel far more than we realise.
That is my Flat Earth.
Microsoft Is Caught in the Middle
And this creates a fascinating problem for Microsoft.
Microsoft built the technology.
Microsoft knows perfectly well what professional architectures look like.
Microsoft has SQL Server.
Azure.
Dataverse.
Business Central.
Power Platform.
Fabric.
An enormous range of enterprise technologies.
But Microsoft also has to sell products into the world as it exists.
And the modern world is driven heavily by social media.
Influencers explain products.
Influencers create excitement.
Influencers create demand.
Influencers teach millions of users.
Microsoft cannot simply ignore that ecosystem.
Indeed, Microsoft benefits enormously from it.
But that creates a feedback loop.
Microsoft creates technology.
The social-media ecosystem interprets it.
Users learn that interpretation.
That interpretation creates demand.
And Microsoft quite rationally responds to that demand.
The market begins influencing the technology just as the technology influences the market.
Which brings us back to the Flat Earth.
Once enough people believe they need parachutes, there is money in selling parachutes.
That doesn’t make the Earth flat.
This Is Not an Attack on the Excel Community
I want to make an important distinction here.
I am not arguing that the Excel community should stop doing what it does.
Quite the opposite.
There is extraordinary educational content available today.
People are learning Excel who would never have had access to this knowledge 30 years ago.
Power Query is brilliant.
Dynamic arrays are brilliant.
LAMBDA is brilliant.
Power Pivot is brilliant.
Modern Excel is astonishing.
The Flat Earth analogy is not intended to say that people who use these technologies are stupid.
It is about something much broader:
the power of a dominant mental model.
When enough people share a particular view of the world, industries naturally evolve around that view.
And alternatives become increasingly difficult even to see.
My argument is simply that there is another Excel world.
And it never disappeared.
Take the Argument to the Boardroom
This is why I am increasingly uninterested in arguing about which Excel feature is best.
I don’t think we need another religious war between VBA and Power Query.
Or Office Scripts and VBA.
Or Excel and Power BI.
Or spreadsheets and planning platforms.
Instead, go upstairs.
Go to the CFO.
Go to the Finance Director.
Go to the operational management team.
Put the actual business problem on the table.
Perhaps 400 departments need to submit a budget.
Perhaps 100 managers need to review it.
Perhaps thousands of employees need access to centrally maintained information.
Perhaps hundreds of locations need to submit operational data.
Now ask:
What architecture best solves this problem?
Then demonstrate the alternatives.
Empirically.
Build them.
Test them.
Time them.
Scale them.
Audit them.
Let management see what actually happens.
And let management choose.
That is a very different conversation from:
“Which Excel feature should we learn next?”
The Opportunity
And this is where I think the story becomes exciting.
Because if I’m right, there is an enormous gap.
On one side are senior managers with real business problems.
On the other are millions of intelligent, enthusiastic Excel users learning increasingly sophisticated techniques.
What is missing is the bridge between them.
Someone needs to show those Excel users that their skills can go beyond building more sophisticated individual workbooks.
They can learn to build systems.
They can learn architecture.
They can separate data from presentation.
They can GET.
They can PUT.
They can build Hub-and-Spoke processes.
They can use Excel as a client to shared relational data.
They can become the people who solve the problems that CFOs have been told spreadsheets cannot solve.
And this is not theoretical for me.
It is how I built my career.
The techniques themselves are often remarkably simple.
The value comes from applying them to the right problem.
Maybe Excel Was Never the Problem
So perhaps the question we should be asking isn’t:
How do we finally replace spreadsheets?
Perhaps it is:
How did we arrive at a world in which so many people believe spreadsheets can only work this way?
Why do we assume collaborative spreadsheet processes require files travelling from person to person?
Why do we assume Excel data must live inside Excel files?
Why do we treat connectivity as a modern innovation when Excel professionals were connecting spreadsheets to relational databases decades ago?
Why do we describe old, proven architectural principles as obsolete simply because the technology industry needs something new to sell?
And why don’t we test these assumptions?
The Flat Earth does not need to be banned.
The Flat Earth conferences can continue.
People can keep buying parachutes.
We don’t even have to argue with them.
We simply need to take the people with real journeys to make and show them a globe.
Then let the evidence speak.
Because in the boardroom, the problems are real.
The money is real.
The inefficiencies are real.
And the solutions have to work.
Perhaps the greatest opportunity in Excel today isn’t another new feature.
Perhaps it is rediscovering something Microsoft gave us nearly 30 years ago:
the ability to stop treating the spreadsheet as the world — and start treating it as a window onto the world.
Add comment