Why does everyone use Excel?

[Shoseki]

I know this sounds like an obvious question, but bear with me.

I was having a chat with a senior administrator at the university where I work the other day. He is a project manager of a large international and multi-million euro project, he has to manage many different aspects of this project, including conferences, so it includes finances, personnel, and all sorts of other things that I (as a techy) don't really understand.

I was curious as to the advanced tools that such a complex system must use, and I was shocked to realise that 99% of it was done "with Excel", not only as a storage "format" for a lot of data, but the entire processes and everything, all through excel.

Now I am sure that during the 70s or whenever spreadsheets started to be used heavily in business and transferred from the old pen and paper booklets for accounting etc it probably was an enormous step to move to excel, but as someone who writes software for a living, excel doesn't seem a particularly complex piece. Its useful, its practical, it is easy to use, but I don't see it as the be all and end all of management software.

Of course, I have heard of, though not had any experience of, complex management systems that companies like Oracle and SAP produce, though obviously not being from a MBA or similar background, don't really have a clear idea of how they work.

The question that I have, which is more of a question towards lay-men, which is why I put it here rather than in a technical forum, is, why is Excel (or spreadsheets in general, obviously I would push form Open Office equivalent) so widespread for use in management? Why is it, that as soon as someone thinks "ah here is something I need to manage e.g. a list of peoples' email addressses" they immediately think of opening up a spreadsheet?

My opinions of Excel are :

1) Easy to use, but also dangerously easy to use. Its easy to open up the incorrect document, click on the wrong part of a document, edit a macro, rename something, damage some part of your "process", without realising it. And of course, the types of people using Excel in this way don't use something like Subversion to track and backup their data most of the time.

2) Poor at modelling the data structures / process being managed. Because of this, its not obvious if there is something wrong, other than some numbers not being correct at the end of a column/row.

Now, I am not saying that a million bespoke systems per company/organisation are the way forward, processes/systems change all of the time, so they would never be stable. But is the use of Excel as a "catch-all" simply due to a lack of technical expertise? Or is it drummed into certain business courses that the best way to manage something is with spreadsheets?

I myself had to stop someone when they asked me to maintain a user list of email addresses... they were suggesting excel. But it was obvious, if I was going to manage 4.5k email addresses by hand, it would be a pain to get people to email me if they wanted to be removed, track invalid email addresses etc. In the end I decided to use http://en.wikipedia.org/wiki/Phplist phplist, which is open source.

/sigh

 

[Shoseki]

VBA. Simple as that. You could write your program to check the email addresses within minutes as you'd create an Outlook object from within the VBA aspects of Excel and let the Outlook calls handle the rest.

It's a VERY underrated system.

You mean visual basic? But

a) I don't know visual basic, I generally program c++/java, I don't necessarily want to go learn vb just to manage some email addresses.

b) The typical manager I have described can't program, or doesn't have experience programming, and I wouldn't expect them to. Math formulae are one thing, programming is another.

c) most of the stuff I said above, about it being dodgy to manage by hand, easy to corrupt, not great to check. At least the phplist, users can remove themselves from the list as they like, or sign up even more. Also, its a database that can be accessed anywhere, unlike excel which is on my hard disk, is easily lost, damaged by others etc etc

 

[Shoseki]

the beauty of it is that you don't need to know how to program to do what J is describing - you can just go and pinch the code and a set of easy step by step instructions to set it up from a forum you can find by googling in a few minutes.

I guess most people use it as that's whats already installed on 90%+ of office computers, it can do the job, is useable by total tech noobs to perform complicated tasks.

Most managers don't have any say in what apps are installed on the company machines, and IT are invariably a pain in the hole to get to change anything.

Its just a case of doing the job with the tools you have at hand, really.

TG

It is easy, I am not denying it, however, normal administrators are *not* going to go pinch code online and use it. Most of them can't program, or are fearful of how that part works. And even so, that is like saying that a software engineer decides to use a piece of code he used online and just plug it in, without really knowing how it works, what it will do etc. Its just not feasible. I am not talking about open apis etc, I am talking about business processes, and the suggestion is that excel is a software development environment, but useable by non-professionals. That makes it pretty unique.

If, for example, you modelled your business processes, knew exactly what type of information you were going to receive, you could set up databases and queries to produce the types of data necessary and manage the business processes. However, they choose *not* to do that, and just use excel. Access can do this - so why isn't it used?

 

[Shoseki]

I am very aware of the fact that Excel is a powerful tool. However, it isn't exactly billed as a "stable software platform", "powerful database solution", or anything else.

Its just interesting to see that managers become programmers, but even if they might know the processes they are trying to manage, they don't have anything other than Microsoft Office to work with.

I have been discussing this with a guy who works for my other job, I do web development for a smaller independant wholesaler. They wouldn't dream of processing hundreds of orders / packaging/tracking etc using something like office, they have a dedicated accounts package that tracks and balances everything. Of course this is fairly specific to the type of work that they do (selling large amounts of stuff). Doing something like this would be an *absolute* pain with excel.

So why are other managers in other positions using excel? Are their processes really noddy? Is it not necessary to be able to prove that their systems work, are able to be rigourously tested, that the formulaes used can be easily checked and managed?

I worked for a company as a statistician a couple of years ago, and within a couple of days immediately noticed and flagged a big obvious mistake in one of their macro heavy excel spreadsheets - they were doing averages of averages, which doesn't work. However, none of the people using the macros had any idea how to fix it, were too embaressed to flag it up (especially as it was really obvious). Their entire system was very manual, difficult to check... and difficult to explain to someone else! They couldn't even diagram exactly how their system worked, who developed it or anything, it just kind of "grew together" piecemeal without any real structure / validation.

 

[Shoseki]

I am confused J. You are saying that "I could be the best programmer in the world, but I could (and do) introduce bugs into the code", which is true, and the basis of quality control, which is surely difficult in something like Excel.

If I use an accounts package to do accounting, and it has been certified and tested to certain standards, then managers using anything other than using certified accounting packages would be nuts. They are creating the code, as you have said - and therefore Excel is doing exactly as you have said, you are creating your own code (even on a spreadsheet basis) that is open to problems.

The entire point of having a really good system is that you create it and manipulate it from a top down approach, so that you can control the quality and efficiency of the system. This is not part of excel. All that excel can do, is provide basic mathematical processing. And even if they aren't programmers, what they are doing amounts to programming, and surely the processes surrounding programming therefore applies. Either it does or it doesn't.

Either managers are independant of the systems that they use, and use them only, or they have to understand how it works in order to design/implement/test their own systems.

I can't help thinking that, if this approach were taken properly, 99% of the administrative ****-ups that occur in real life wouldn't have happened.

And the comment about diagrams, well - a diagram can illustrate something that would take a great deal of time to describe. It doesn't matter what it is applied to. If the processes surrounding a business need to be modelled, a suitable diagram should suffice.

EDIT : most of the other replies, specifically the "power" of excel are, as I have said, well grounded. I have not fought against the fact that you could do a full implementation of whatever those folding guys are doing with the current distributed projects are, or SETI at home type analysis, in Excel. The practical reasons that we don't, however, such as performance and reliability, are enough to make them *not* implemented in Excel.

The replacements for Excel? I am not sure off the top of my head. Managing different types of resources requires different types of software. Its a management process in general. I am not talking about one-off calculations or scientific evaluations, such as high-level industry modelling (I guess aerodynamics of planes or some such, although I suspect you'd be nuts to implement it in Excel), the problem is that you would test hundreds of times for each and every single person, the spreadsheets they set up, the calculations they used, whether the same versions of the process/implementations were being used, compatibility etc...

Simple example : I make a spreadsheet with a series of numbers, and put a total in column E6. I then continue to develop other spreadsheets calling on E6 to produce that total. Other people do the same, calling on E6 as a total. Then later on, I decide I need to rework the original template, but I need to change E6 to be something else. How do I manage the change? How do I manage the changes needed to correct all the templates? What about the ones which I don't know about but are still essential to the business? How do I document the changes properly?

In proper software development, you would develop an API which would be published (probably online) and then subsequent calls / changes would be tracked, you would be able to see if people made calls to something deprecated or at least provide backwards compatibility, none of which you can (as far as I understand) do with Excel.

 

[Shoseki]

The results of many many computational engineering processes (CFD, Stress analysis) are tabulated and then used in Excel sheets to give you your final values.

Can you name me any that have been done in a standard way? Could I take that spreadsheet, and say, apply it to the same process in another company?

How would you teach someone who is technically able to do the work, but unused to the system of that particular company, where and how the spreadsheet macros work? Does your company have standards and policies set up to guide how these should be managed? How can you check those hundreds of files for complicity with those guidelines?

What I am trying to get at, is that although data/process management can be done in either excel or programmed into a system, at least with a system (using a database, maths/physics libraries, standard APIs and storage etc, possible functions for distributed processing, easy to debug write once libraries, possible open source equivalents / alternatives for various functions) could be extended properly, modelled properly etc as opposed to Excel which wasn't designed to be able to handle anything but numbers and figures, not be able to communicate with anything (other than word or other microsoft products).

Ask yourself the question, if Excel wasn't there, wouldn't all those projects foster innovation and standards? If managers actually *had* to get a proper system developed and commit to a particular model of process/management, instead of switching around all the time, they take more seriously their process and modelling requirements?

And finally, if Excel is so widely used, and so Microsoft so widely accepted, why is it that every system has a backup to "CSV" files as the only "ascertained technique" of communicating between non-microsoft and microsoft products?

 

[Shoseki]

Well, its hardly cheap at the licence fee, but I guess when every business auto-turns to microsoft to solve all of their problems, I guess its not really such an issue.

Heres a fun example of someone who is using Excel in a business environment :

http://en.allexperts.com/q/Excel-1059/macro-convert-excel-columns.htm

Now, tell me that is "damn good, cheap and easy for people with next to no knowledge to use. It really is that simple!".

That is programming, involves datastructures etc but without the support of a good engineering environment, quality checking etc.