Affection for Excel spans generations, from Boomers to Zoomers Despite its advancing years, Microsoft Excel is proving a hit with young finance professionals, many of whom reckon the aging number-cruncher has a bright future. According to a Datarails report, more than half (54 percent) of 22 to 32-year-old finance professionals say they outright "love" Excel, up from 39 percent among the …
I think that conclusion is spot on. Excel is sort of the "this is why we can't change" tent pole of the ms stack and users actually like it. Companies will learn windows administration because excel is an accept no substitutes market leader
Messing with the formula (ha) could be a real cliff edge
Some of them do have to use Excel, whereas others could use any spreadsheet but are so used to the Excel UI*1 that they don't want to change. In both cases, they could use it through a VM*2 if ditching Windows was desirable.
*1 Sometimes, Microsoft has changed the Excel UI and angered users. I knew many people who were very big users of Excel who got very unhappy with the introduction of the ribbon. But now that they've finally gotten used to that, they will react equally unhappily to any other big UI change including going to LibreOffice.
*2 I don't know if you can use Wine or related software with the full Office. I've never bothered to attempt it and my initial thinking is that something will go wrong. The web version of Excel is technically capable of the same things and looks sort of like the desktop version, but I have also never tried having someone who uses Excel a lot go that way and imagine that the decrease in speed would be as noticeable there as it has been whenever I've used that or Google Sheets.
It's called "market advantage".
The problem here isn't, fundamentally, Micro$haft itself.
It's corporate investors who demand ever-increasing profit from companies.
Rather than being satisfied with a steady, rock-solid income, they demand year-on-year increased yields.
Hence the current AI bubble.
I await the inevitable burst/crash eagerly.
You can run Office 2003 in Wine without any problems (other than Access and maybe Outlook). Anything later, not so much. But LibreOffice is able to match Office 2003 (other than Outlook and maybe Access), even if it can't match more recent versions. Outlook 2003 probably isn't going to work with modern email servers at this point, so could be discounted for that reason.
... I suppose custom programs are more common in finance?...
It's not that much about Excel itself as it is about obscure, uber-expensive add-ins that are irreplaceable and only exist for Excel.
And not only obscure ones. Bloomberg add-in exists for Excel, not for Libre Office. That seals it for many.
Then, Bloomberg API would allow you to do plenty of stuff, including piping data to Libre Office or some boutique tool. Probably way better in a flexible environment. Not so for others.
We'll always boil down to best car vs best fleet car on that one.
Do they need Excel or just a spreadsheet.
Some need plugins or add-ons are Excel specific and difficult to replace.
It also depends what heathenry has been committed in the past. A friend worked at an amusement park as a summer job years ago. Their boss had this insane spreadsheet that did some sort of rota/scheduling process. Huge document, multiple sheets. Apparently when they changed an input value you could watch the ripple of locked cells updating. The thing was so ginormous that the update wasn't "instantaneous" (to a casual observer).
I have this strong suspicion that if you opened it in LibreOffice, it would break in all sorts of interesting ways. Although hopefully it's been replaced by something sensible by now (lol, yeah right).
Of course, that's a wild abuse of Excel and there are better ways of doing rotas and time tracking. But it's also a pretty hard blocker if you said "right, you lot only actually use email and productivity, so you're going onto Ubuntu and LibreOffice next month in the IT refresh". Cautious handraise of "will my undocumented business-critical spreadsheet work in this LibreOffice you speak of?".
if you don't consider things like auto formatting data it reads in, forgetting the format of columns in a filtered table read from an external source, or the whole thing being abused for things it was not meant for. Plus the shitty graphs, the inability to do statistics with it (despite people trying and messing things up), the way it translates inbuilt functions so switching between system languages messes with you (thank you very much, I do work in different environments), the bloody stupid way in which frames / cell borders are coloured, its complete inability to read in correct CSV files, the import functions they removed, and the AI they will inevitably add. But yeah, it's a great product. /s To be fair: you can do a lot with it. It's just that so many things are more difficult than they should be. At least the bug that it would only use the first 63 cells for some stats functions no longer exists (I hope, that was a couple of decades ago).
I wish I had a job where it was explicitly forbidden and people were forced to store and analyse data in a real database...
Germany (and any country that uses "dot" as thousand separator and "comma" for fraction): specific...
IP: 1.102.100.100 = silently autocorrected to the number 1102100100. Same for IP 234.123.323.023 = 234123323023. Only if your have something like 123.10.213.222 it does NOT autocorrect, since the ".10." does not match the thousand separator logic.
LO at least gives me an assistand upon opening a CSV, so I can set the format as needed.
If only the formular on a German Excel installation could be in English....
I wish I had a job where it was explicitly forbidden and people were forced to store and analyse data in a real database...
Absolutely.
Unfortunately, like many others, all we get are lousy (customizable) reports on lousy data in a lousy tool with very limited connection to financial data. Hence, I'm really glad that I can fiddle around with Excel and its native array functions.
Whatever you think about Micro$haft as a company, Excel (and Word etc) do the job pretty damn well.
I personally tend to use Libre because I prefer simpler and functional. Without fancy bells and whistles. But I'm admittedly old-school.
I just wish M$ would stop trying to shove "AI" and other "features" into everything.
To reverse the old analogy... When you've got yourself out of the hole - don't start digging again! ;-)
Then you need a spreadsheet and nothing else will substitute. That's why Visicalc was such an instant killer app and why people bought an Apple to run it.
Excel, as crappy as it is, is about it for options, unless you're lucky enough to be able to use Libreoffice. Visicalc, SuperCalc, Lotus 1-2-3 etc have all fallen by the wayside
Google Sheets sucks even worse than Excel, which is pretty impressive.
I remember telling someone "ok you put a number in this cell, and a number in this cell, and this formula in this cell and IT AUTOMATICALLY COMPUTES IT" 40 years ago and watching his head explode.
OTOH, I recently showed someone a paper spreadsheet and his head exploded that people did all that manual labor.
Sheets requires you to be logged into Google and working in a (slower) web browser with Google slurping your data. Sheets will never compete with Excel. I run a standalone lifetime Office 2021 installation on a PC where I am logged in on a local account (and not logged into anything MS or Google) with the PC having O&O ShutUp10++ set to turn off all telemetry & MS AI crapola. In addition to Excel being a hell of a lot faster than Sheets for complex spreadsheets, I have a reasonable expectation my data stays private on my PC without me having to air gap it. With Sheets, no way.
I still miss true xyz plot, especially for science things it should be there... https://bugs.documentfoundation.org/show_bug.cgi?id=51670. So either Excel (since v4, 1992), other plot programs, or powershell "draw whatever I want" way.
This!
It boggles my mind, that the plotting of XYZ data is not possible, and that nothing has been done about it for decades.
Twiddle around with the UI? ..every 4 years.. ..but fixing this glaring omission in plotting capability? ..Never!
...Oh, and since we are at it, do not only add xyz point clouds, but also add xyz surfaces ( preferably with surface coloring from a 4th table ).
And since we are working on improving the capabilities of the spreadsheet (Excel or Calc, does not matter):
Let us have named cells and equation formatting in the cell entry.
As it is, my eyes start bleeding when I have to create or modify more complex cell entries, and falt finding is very difficult.
There is no reason why ((A2+C$3)/(D2-E2))^2 can not be displayed as:
| (A2+C$3) |^2
| ------------- |
| (D2-E2^2) |
(apologies for the ascii art)
Subscribe to that feature request / bug. Problem is: LO is for free, so there is no enforcement or way to push specific features. Though the ways suggested were among paying companies to make it for LO, usually lower five digit. So currently I either install an old office, the last version which did not need activation (for its volume license key) or something older. Or use something else like Freeoffice (aka the limited softmaker office).
I remember SuperCalc and longed for GUI based Excel which came out on the Mac first. Then it with, crucially, VBA became core to everythng I did in Engineering. The ease of VBA automation meant you could create fontends for file I/Os, data parsing/manipulation, conversions and presentation/reports and not have to wade through all the rows and columns of data.
Google Sheets isn't the competition and, despite their efforts, neither are Libre/OpenOffice or MS Office clones such as Softmaker and OnlyOffice. Spreadsheets are increasingly becoming relegated to the report format for work that's actually done using Jupyter Notebooks and the like. With a little training most users are quickly more productive and their work more accurate than anything you can do in Excel.
I have never needed a spreadsheet in the six decades-plus of my life.
I have created and used them for some of the same reasons many other people do so:
* they're fast to create vs "real" programming;
* they have a built-in data presentation model which most people can get their heads around;
* they work well-enough for small amounts of data; and,
* using them is faster than doing things with a paper spreadsheet and a calculator.
Those advantages do not in themselves make spreadsheets "good", in the same way that eight-sided cart wheels are better than square cart wheels, but still are not "good".
The reason so many people think they need spreadsheets is that they've never been taught any other way.
Using the traditional tools?
Obviously, I use a chequebook, although since CoViD it’s become home to moths …
The programmer/blogger Joel Spolsky claimed that Microsoft had done a survey and discovered that most Excel users used it to make and save lists. Those users one could win away to something else, and Google Sheets would be a candidate. But I'd hate to try to take Excel away from an accounting department.
My real complaint about Excel is that some people will use it as if it were a database (lists, yes, but really big lists), which leads to problems.
I think that does account for a lot of users. In case people are interested in the blog we're talking about, it is here and it's from 25 years ago.
The problem with the complaint is that we're entirely correct that Excel has a lot of problems when used as a database but, from the perspective of most users, there isn't anything else for them to use. Following the same design advice from that post, we haven't considered what the people who don't have experience with databases need and therefore haven't built something that solves their problems. If we tried, we'd likely end up with something that looked a lot like a spreadsheet but just didn't do as many of the spreadsheet things. It's likely not very hard for us to write an SQL table schema to describe the data we're going to store and some input validation restrictions, either running at the SQL level or slightly above it. Both are foreign to a lot of people who need to create and store records with specific fields. Therefore, it's often quite difficult to figure out what to tell people to use instead of Excel. The stereotypical comment doesn't even try, just saying "a database" which doesn't specify what database engine or how they'd configure it and certainly doesn't cover the entry and display parts of the problem.
Assuming that alongside a knowledge of SQL the user had the authorization to buy and the permission to install SQL server
Then if you are in a government agency, this involves calling Crapita/Cap-Gemini/Fujitsu and telling them that you need a database to do XYZ and wait for a quote which is equal to the price of small aircraft
carrier
There are plenty of free databases using SQL or SQL-like syntaxes, notably Postgresql (you still need to install it somewhere) and SQLite (works entirely portably). Licensing isn't really the problem.
The problem is that you can put a spreadsheet in front of a user and they'll figure out how to find an intersection between a row and column to enter data, but put them in front of an SQL CLI and they'll be immediately lost. Most systems that use a database have programmed UIs built specifically for it, and if you have those, they are generally quite helpful with two major drawbacks. It took someone with more programming knowledge to write them, so if you don't have one of those, you don't have them. And, if you need anything that's not already in the UI, now you've got an annoying process of trying to find someone to add it and the addition won't be a simple formula or view, a problem that gets worse if your change needs to change the database's structure by, for example, adding a column which some things will handle, most things will ignore, and some things will break on. That's why spreadsheets have and will continue to win for this kind of task. We have not succeeded at making anything better from the users' perspective.
Licensing is ok, but permission to run it in production is a problem in many cases. I'm really glad we are finally moving away from Oracle and MSSQL, but that took a concerted effort to do, and retrain some manglement people who were enthralled by the copious amounts of money free lunches fed to them by insultants. Support is (always was) a big thing, as are guarantees (though for software that usually is "provided as-is" and "not guaranteed to work as described" in the EULA).
So while we that do the work would gladly use Debian servers for (almost) everything, and k8s, and postgreSQL and whatever, The Powers That Be actually force us to use DeadHat (or maybe SuSE if they are having a good day) or Windows Server.
The problem with the complaint is that we're entirely correct that Excel has a lot of problems when used as a database but, from the perspective of most users, there isn't anything else for them to use. Following the same design advice from that post, we haven't considered what the people who don't have experience with databases need and therefore haven't built something that solves their problems. If we tried, we'd likely end up with something that looked a lot like a spreadsheet but just didn't do as many of the spreadsheet things. It's likely not very hard for us to write an SQL table schema to describe the data we're going to store and some input validation restrictions, either running at the SQL level or slightly above it.
Yeah, as you say, it's a "good enough" solution for a lot of things. I help with a small sports org. 70 members. Yes, I store the membership details in a spreadsheet. That's how it was provided to me, that's how I've kept it.
Could I write an app over SQLite? Eh, probably. I'm not really a software developer by trade - through I could undoubtedly create the db and have some python scripts to perform functions. Good luck handing that off to my successor though!
MS Office used to (maybe still does?) come with Access, which was also a long way from perfect but did at least try to offer users the ability to graphically build a bit of a low-code application as you could build views and forms on top of the database layer. But it was always a poor cousin and people weren't going to take the time to learn it when they could just make a list in Excel.
LibreOffice has Base but I struggled to get into it the last time I looked (some years ago!) and I haven't tried again since. Looks like it still needs Java as well since it's based on HSQL if you're running standalone and not connecting to a MySQL/PostGres server. Which is a shame because a portable, intuitive, non-technical front end for SQLite would be amazing for that no-code group.
For what it's worth, I know an organisation maintaining a couple of Windows 95 laptops to keep a DataEase application going that they use once a year... Kind of insane, but they've never found the time or resource to port to anything newer... (also nothing I have anything to do with nor the time to unfortunately, as much as I'd like to!).
In my work with election data processing, we were forced to move from a spreadsheet-y system where you could freely scroll through data, to a database-y system where you could only see one elector's data at a time. The productivity destruction from moving from a "phone directory" format where you could see an entire street ALL AT ONCE to a "index card" format where all you could see what one person at a time was immense.
Excel, or how an ERP is written by the accountants themselves. Except without the security, scalability, reliability, compliance, supportability, auditability and upgradability of a modern ERP.
Costs of that abomination are hidden in years of over serviced, fat Accounting/Finance departments. All Op-Ex, so every one is happy.
Move one file, even migrate to a new file share. Change a drive mapping to a UNC, and the whole lot falls over faster than an Italian football player.
However, most modern ERPs don't have those attributes either, so, really, no loss.
The main problem with Excel is that except in a few select use cases, at best it's the second-best tool for the job.
Its ubiquity, familiarity and its ability to be bastardised with VBA functions though mean that despite not being the best tool for the job, it's the best tool that users have at their disposal without having to fight the finance department for a licence for something more suitable.
I've seen some unholy behemoths in corporates over the years that have grown fungus-like out of Excel. A special mention goes to the UK-based laptop manufacturer I worked at 20 years ago where their entire robotised build line was controlled using a truly spectacular amount of VBA on top of Excel.
One of my first software jobs was to write a program to replace a Lotus-1-2-3 spreadsheet which was used to generate estimates and quotes. Only problem was that there were at least 3 different versions of the spreadsheet and no one could tell me which was the most current one. The project was doomed from the start.
I use Excel on a daily basis, but I’m very wary of using it for anything important.
In my experience, Excel is riddled with data-corrupting bugs that are nearly impossible to mitigate from within Excel. This includes silent data loss due to static values in cells being “autocorrected” to nonsense values, even when autocorrection is turned off. This also includes several mathematical functions not doing the calculation they claim to. The values returned seem superficially plausible, but in fact are wrong.
I’m also wary of using Excel for spreadsheets with a substantial number of data cells. Excel becomes unstable and crash-prone.
My advice: Learn Python. Much better at anything that can be done in Excel. Or for simple stuff use an open-source spreadsheet program such as Gnumeric, where fatal bugs get fixed rapidly. In Excel, major bugs are left unfixed for years, possibly decades.
On the other hand, I know an engineer who is an internationally-recognized leader in his field, who uses Excel to design major projects. When gave a presented a talk on his work, he noticed me rolling my eyes, and got a bit offended.
I recommend powershell over python here. For once, if you have excel powershell is usually there too (MacOS? not often seen here...). And you have access to the whole dotnet stuff, 128 bit floating point when needed etc. And it is SPACE-tolerant :D. And you can clearly distinguish between a variable and non-variable.
Which is probably why the Pandas library was developed in a bank… You can have 128-bit floats in Python using Numpy, which is what most people will be using. Though I don't understand why you would want to do that and work with Excel with it's 15-digit precision…
I can sort of understand if programmers complain about Python (syntax, whitespace, dynamic typing…), but in situations like this, such criticism is entirely out of place because the users do not consider themselves as programmers. For longer than I care to admit, I've watched Python spread into non-traditional computer science domains because of the combination of a rich ecosystem of tools, and a syntax that non-programmers are comfortable with and that helps them get stuff done. And now you threaten them with Powershell?
I get your point, you need to get python AND install additionally the right libraries, you mention Pandas an Numpy. My point: Why get something you have to download and install in first place, if powershell is, like, literally on every Windows machine since Vista? Microsoft really does not know how to promote good things...
Two reasons.
1) Powershell started out decent, but quickly became terrible and is getting rapidly worse.
2) It's not intended for data analysis, it's for scripting Windows administration tasks.
For us the killer was that the security model means that most normal users can't run ps scripts at all. That's probably because of (2)
The final nail in the coffin was that python is trivially available on all platforms, while powershell isn't. Again, that's probably because of (2)
Python is basically a glue language intended for munging the data output by one tool into the format desired by another. That's pretty much the definition of a data analysis package, and it does it quite well.
To be honest, pretty much the only thing I don't like about Python is the significant whitespace, which introduces errors and makes diffs harder to read. Other than that, it's pretty good.
> 1)
> 2)
Depends on what you use it for, and be aware, the default array type is slow. But of course, you can hit limits when you need something else.
> most normal users can't run ps scripts at all.
If you allow other languages, like python, then you can allow powershell as well. No difference regarding security here. You can use PS-ISE to work around the limitations though.
> python is trivially available on all platforms, while powershell isn't.
You have a point there. Powershell is not available on as many platforms as python. But where Excel is, you have powershell, which is somewhat the point of the article... PS is a first unloved child of MS, and the non-integration in ecxel is one o the things they missed. I hate vbscript, it is such a <got no word> language.
Which platform that has Python but doesn't have PowerShell was important to you? I prefer Python, so I'm fine with this, but if this was because PowerShell isn't on Linux or Mac OS, it easily can be installed and you would have to install Python manually too. I doubt you're referring to embedded systems running MicroPython. This doesn't look like a valid reason against PowerShell.
And the security model isn't because it's used for admin, it's because IT departments like to lock down user machines and sometimes for good reason, so when Microsoft designed another tool, they decided to give admins more control of what users are allowed to do with it. If you change the policy to unrestricted, you have exactly as much protection against user actions as you do when they're running Python scripts which do not have limits on what actions, under the users' privileges, the scripts can use.
A few years ago, the scientific community updated the names of several genes to avoid Excel autoformatting:
https://pmc.ncbi.nlm.nih.gov/articles/PMC9325790/
The consensus, apparently, was that this was easier than teaching people to disable the feature or use a more appropriate tool for the job.
This wasn’t autoformatting, this was autocorrection. Autoformatting doesn’t destroy data. Autocorrection does. And does so irreversibly.
Excel provides a toggle to switch autocorrection off, but it doesn’t stick – the toggle automatically toggles itself back on again without notifying the user. There was a lot of discussion about how to mitigate this, but no one was able to come up with a fix other than avoiding Excel like the plague.
> teaching people to disable the feature
The “feature” can only be temporarily disabled. And when it silently turns itself back on even for a fraction of a second, you have a huge mess on your hands.
So this isn’t a case of the users being at fault
Unfortunately, for many years a lot of the leading biomedical journals _required_ authors to submit their data as Excel spreadsheets. More recently, other formats such as CSV are commonly allowed.
Users are only at fault in the sense that Microsoft Office has become such a deeply-ingrained part of workplace culture that most workplaces require it.
Also Economists have problems with Excel:
The 28-Year-Old Who Caught the Excel Error Heard Round the World
and so debunked the neoliberal myth that countries with high levels of national debt saw significantly slower growth....
From the article:
"According to a Datarails report, more than half (54 percent) of 22 to 32-year-old finance professionals say they outright "love" Excel, up from 39 percent among the older generation."
I knew someone working in the compliance department of a large financial company, and they described Excel as their absolutey worst nightmare. It was the job of the financial whizzkids to cook up some cunning scheme for make money, providing a service, etc. It was the job of the compliance department to vet the scheme for legal compliance issues. It was then supposed to be passed over to the softies to develop the code to run the scheme, which also had to be reviewed and checked for compliance in its own right.
Where Excel came in was that it allowed the financial whizzkids to run their dreamt up scheme without needing the software to be written by the softies, and without troubling the compliance department at all. They could bypass all those boring checks, balances, processes and just wing it in a spreadsheet on their desktops...
A big part of the compliance department's job was scouring through company laptops actively hunting out such spreadsheets, and finding them to be in plentiful and continuous supply...
The dismal financial history of the last four decades must rank "finance" "professionals" up there with the great oxymorons… Excel getting into act sort of, pardoning the pun, figures.
Spreadsheets generally look like a decent model-view-application whose specification has been masticated by AI which vibe coding has spat out.
The idea of tabular data entry and presentation pre·date spreadsheets and aren't a bad idea. Embedding the applications logic in the table isn't the greatest idea and compounding that felony by doing it in the least abstract and least transparent manner possible must rank with the least best ideas of all time.
How does this article not discuss the biggest issue facing a modern spreadsheet application such as Excel?
For 40 years now, we're plagued with "Sorry, Excel can't open two workbooks with the same name at the same time."
All those wannabe players like LibreOffice (other Office suites are available) have somehow solved this incredibly tricky mesa-level problem that's vexed Excel engineers for decades, yet you still can't open two files in excel from two different folder paths that share the same filename and extension.
No-one yet has mentioned Lotus 123. I preferred it to Excel but eventually had to port everything I used to Excel. At which point (199x-ish) I discovered Excel didn't have a 2-D lookup, you had to nest formulas. M$ took years to add it (some time after Excel 2019).
And before that there was Smartware, a DOS-based suite. Had a lot of stuff in it that more "advanced" suites did't have until much later.
This post has been deleted by its author
A bank I was working at had about 400 people in the UK office. I recall a routine query on the NetApp filer showing there to be 780 spreadsheet files open. There were of course orders of magnitude more not in use at that instant in time. Moving away from Excel would have been unthinkable as the licensing costs were peanuts compared with the migration cost of compatibility checking everything. The risks would have been enormous.
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2025
