By Hiran de Silva
I have been looking again at one of the most interesting attacks on Excel that I have encountered over the years.
It comes from Workday Adaptive Planning and is called The Nine Circles of Excel Hell.
The proposition is familiar.
If organisations rely on Excel for financial planning, budgeting, forecasting and reporting, they encounter errors, consolidation difficulties, version-control problems, disconnected spreadsheets, manual processes, collaboration problems and all the other horrors that have been associated with spreadsheets for decades.
And, of course, Workday Adaptive Planning offers a way out of this hell.
Now, there is nothing inherently wrong with a software company identifying problems that its software is designed to solve.
But I think there is a very important Trading Standards question that needs to be asked.
Are these really Nine Circles of Excel Hell?
Or are they Nine Circles created by one particular way of using Excel?
Because those are two very different things.
What exactly are we comparing?
There is something I have noticed repeatedly in comparisons between Excel and enterprise planning platforms.
On one side we have the planning platform.
It has centrally managed data, shared access, controlled storage, security, workflow and an enterprise architecture.
On the other side we have Excel.
But Excel is represented by hundreds of disconnected files, containing their own copies of data, being emailed from person to person, copied, renamed, uploaded, downloaded, linked together and eventually consolidated.
Then we compare the two.
Unsurprisingly, the enterprise planning platform wins.
But there is something rather strange about this comparison.
One contestant has arrived with its entire supporting architecture.
The other has effectively been sent into the ring carrying an email attachment.
That isn’t a like-for-like comparison.
What do we actually mean by Excel?
This is the fundamental question.
Suppose an organisation has hundreds of people using hundreds of spreadsheets.
Each spreadsheet contains its own data.
Files are passed around.
They are copied, renamed and emailed.
They are linked together.
Eventually somebody has to consolidate them.
Versions get confused.
People overwrite things.
Links break.
And sometimes files disappear altogether.
I have absolutely no argument with anybody who calls that Excel Hell.
I would call it Excel Hell myself.
But now consider a different architecture.
Keep the spreadsheets.
Keep all the things that make spreadsheets useful.
Keep the calculations, the models, the presentation, the flexibility and the familiarity.
But take the data that would otherwise be trapped inside those hundreds of individual spreadsheets and store it in one structured central location.
Every spreadsheet participating in the process can reach that central data.
Every spreadsheet can retrieve the information it needs.
And authorised spreadsheets can submit information back.
Suddenly those spreadsheets are no longer isolated islands.
You have a hub-and-spoke architecture.
One hub.
Hundreds of spreadsheets.
And this completely changes the discussion.
The Digital Librarian
Database professionals have been connecting front ends to databases forever.
So there is nothing revolutionary about the underlying technology.
What I think is different is the perspective from which we look at it.
A traditional database professional might describe Excel as a front end to a database.
Technically, that is perfectly reasonable.
But notice where that description puts the emphasis.
The important thing becomes the database.
The infrastructure.
The centrally managed data.
Excel becomes merely the front end sitting on top.
But turn the telescope around and look at exactly the same architecture from the spreadsheet user’s perspective.
Imagine a business process involving 400 spreadsheets.
The interesting proposition is no longer:
We can put an Excel front end onto our database.
It becomes:
We can take the data currently imprisoned inside 400 disconnected spreadsheets and put it in one structured place that every one of those spreadsheets can reach.
That is a very different proposition.
And that is why I use the term Digital Librarian.
Think about what a librarian does.
Hundreds of people can ask a librarian for different things.
One person asks for France.
Another asks for Department 27.
Another submits this month’s forecast.
Another asks for all transactions belonging to a particular cost centre.
The individual spreadsheets do not need to know anything about one another.
They simply talk to the librarian.
The librarian knows where the information is.
The librarian stores it.
The librarian retrieves it.
And the librarian can accept new information and store that too.
But surely Excel can’t do that?
This is where the story becomes particularly interesting.
Because we don’t need to invent some exotic new capability to make this possible.
Excel already has the plumbing.
Excel can communicate with relational databases.
It can retrieve data.
It can submit data.
It can execute queries.
It can participate in a client-server architecture.
And this capability has been available for decades.
That is what makes the whole discussion about disconnected spreadsheets so fascinating.
The spreadsheet world has spent decades being warned about the dangers of hundreds of disconnected spreadsheets while Excel has had the plumbing to connect those spreadsheets to centrally managed data all along.
Now let’s go back to Excel Hell
Consider one of the standard spreadsheet criticisms.
Someone deletes a formula.
Someone overwrites a number.
Someone damages a workbook.
Excel is therefore declared unreliable.
But there is an architectural question hidden inside that argument.
Where does the authoritative data live?
If the organisation’s authoritative data lives inside that workbook, damaging the workbook may indeed be disastrous.
But suppose the workbook is simply participating in a hub-and-spoke architecture.
Destroy the spreadsheet if you like.
The central data hasn’t disappeared.
Replace the workbook.
Retrieve the data again.
The spreadsheet error is still real.
But the consequences are completely different.
The architecture has changed the nature of the risk.
What happened to consolidation?
Now consider consolidation.
Imagine 400 departments submitting budgets.
Traditional spreadsheet thinking says:
400 departments.
400 spreadsheets.
Therefore we must consolidate 400 spreadsheets.
Then everybody complains that spreadsheet consolidation is difficult.
But the Digital Librarian asks a different question.
Why are we consolidating spreadsheets at all?
Let each budget spreadsheet submit its data into the central structured store.
Now there aren’t 400 documents to consolidate.
There is one body of structured data to query.
Give me the Group.
Give me Europe.
Give me France.
Give me Paris.
Give me one shop.
The spreadsheet asks.
The librarian answers.
The entire problem has been reframed.
And that exposes the real comparison
This is where I think the Nine Circles argument becomes particularly interesting.
Suppose we compare:
Excel
with hundreds of independent files, local data, email attachments, manual consolidation and point-to-point processes,
against:
Workday Adaptive Planning
with centralised data, shared access, controlled processes and cloud infrastructure.
Of course Workday wins that comparison.
But we have changed two things simultaneously.
We have changed the application and the architecture.
Then we have given all the credit to the application.
So let’s make the comparison more interesting.
On one side:
Excel + central structured data + shared access + hub-and-spoke architecture + local or cloud database + spreadsheets retrieving and submitting data.
On the other:
Workday Adaptive Planning + central structured data + shared access + cloud architecture.
Now we have something worth investigating.
Now we are comparing the capabilities of two solutions rather than handicapping one of the contestants before the race begins.
The Excel strawman
There is a name for the argument I am challenging.
A strawman.
Take the weakest version of your opponent’s position.
Demonstrate that it performs badly.
Then present that defeat as evidence against the opponent itself.
The important thing is that the examples used against Excel don’t necessarily have to be false.
Quite the opposite.
They can be completely true.
That is what makes the argument so persuasive.
If you run a major planning process using hundreds of disconnected spreadsheets, passing data from person to person, you can create an appalling mess.
I have no difficulty calling that Excel Hell.
But what you have demonstrated is the failure of that architecture.
You have not necessarily demonstrated the failure of Excel.
And this is also why the phrase Excel versus cloud can be misleading.
Excel is perfectly capable of participating in a process where the central data and services are in the cloud.
The legitimate comparison is not:
Excel versus cloud.
It is:
Workday using centrally managed cloud data versus Excel also using centrally managed data.
Then let’s see what each can do.
And then the story takes a delightful turn
There is another part of this story that I find particularly interesting.
Workday provides Excel connectivity.
So does Anaplan.
So does Planful.
Think about that for a moment.
After years of the planning software industry telling organisations about the dangers of spreadsheets, planning platforms themselves provide ways for customers to work with their centrally managed information through Excel.
What changed?
Excel didn’t suddenly become a different spreadsheet.
What changed was where the data lived and how the architecture worked.
Central data.
Multiple Excel users.
Excel retrieves information from the central system.
Users work with it.
Information can flow back.
Does that architecture sound familiar?
Our Digital Librarian has entered the building.
The irony
This is why I think the irony is difficult to ignore.
The planning software industry has spent years attacking the chaos created by disconnected spreadsheets.
And in many respects, rightly so.
But its own Excel integrations demonstrate something rather important.
Excel itself was never necessarily incompatible with centralised data.
Which raises a rather awkward question.
If Excel connected to centrally managed data is useful today, why wasn’t Excel connected to centrally managed data part of the original comparison?
Why compare:
Workday plus centralised architecture
against:
Excel minus centralised architecture?
Why not compare like with like?
Workday plus its central data architecture.
Excel plus a central data architecture.
Now we can conduct a fair trial.
So let’s investigate the Nine Circles properly
I am not going to dismiss Workday’s Nine Circles.
Quite the opposite.
I want to investigate them.
One by one.
Where Workday identifies a genuine Excel problem, we should say so.
Where Excel genuinely has a weakness, we should say so.
There should be no special pleading for Excel.
And there is certainly no point pretending that spreadsheet problems don’t exist.
But for every one of the Nine Circles, I want to ask three questions.
Is this an inherent limitation of Excel?
Or is this what happens when spreadsheets are deployed as isolated documents carrying their own copies of shared data?
And most importantly:
What happens to this particular Circle of Excel Hell when the same spreadsheets share one structured central source of data?
That, I think, is the fair comparison.
The case remains open
Workday’s proposition is essentially this:
Why suffer the problems of spreadsheet-based planning when you could move to a centrally managed cloud planning platform?
My challenge is different.
Why assume that using Excel means accepting disconnected data in the first place?
Take the data currently scattered through hundreds of spreadsheets.
Store it once.
Store it centrally.
Store it structurally.
Let every spreadsheet that needs that information reach it.
And let every authorised spreadsheet contribute to it.
That is the Digital Librarian.
Not Excel as some insignificant front end bolted onto the supposedly important database.
Look at it from the spreadsheet user’s end.
Hundreds of powerful, familiar and flexible spreadsheets, suddenly liberated from having to carry their own private copies of shared corporate data.
And the expensive new technology required to achieve this remarkable transformation?
That’s the punchline.
Excel already has the plumbing.
Which leaves me with one question for Workday.
If the case against Excel depends upon comparing your centrally connected architecture with Excel’s disconnected architecture, what happens to the Nine Circles when Excel gets a hub too?
That is what I intend to find out.
Nine Circles.
One at a time.
CASE REMAINS OPEN.
Add comment