For about 10 years I had a consultancy that in the latter half focused on B2B. One of the biggest qualifier pitfalls was thinking we could replace Excel with anything else. There are a myriad of reasons this fails to work. I swear some of them are near voodoo.

Google Sheets is a fit in some situations but the real facts are that Excel is amazing and when someone has been running core biz logic on it for a long time; that operator is someone you need to get to know before feeling too certain.

There's money in replacing that person and fucking the whole thing up so they are stuck with a long tail of consultants. I never ran that sort of business. However on a couple painful occasions I overzealously pitched what python and psql could do to solve a problem and later had to eat crow.

I know what that voodoo may be. Of the many failed "replace Excel" projects I've seen in the investment banking \ hedge fund world over 20 years, the usual mistake is that the focus was on the specific Excel model itself and not the generic building capabilities that Excel provides. An Excel model, i.e., the specific map from inputs to outputs, is typically fairly easy to replace with another tool. But what you've done is replace just the pants, not the sewing machine. Excel is the pants AND the sewing machine. And that sewing machine part is very hard to replace. So you can end up with the same model in C++, the very same map from inputs to outputs, but the user's ability to make changes and understand the model is greatly diminished. So they say "no thanks" and stick with the Excel model.

You summed it up perfectly. Jobs I successfully finished after a few bites of tragically not understanding involved working closely with the Excel department and making sure they could function as usual but extracting data to go into the systems and providing a net gain for those people.

At the end of the day, the people that do the magic that keeps everyones job paying aren't always (or nearly never are) the people that make the contract hire decisions. If you can give those people a tool that lets them 2x their work and be excited about it there's a big net win. If you want them to move to a statically modeled web interface with dynamic tables and they lose all their ability to do everything they know in Excel, everyone loses.

Portability and timing is key if you have to take on that sort of gig. Team buy in might be worth more than that. Never take somebodies sewing machine if you still want that biz to make enough money to pay you.

Agree. If you're selective about which bits and pieces you move away from Excel, and not religious about getting rid of Excel entirely, you can often have a great result.

What trips people up is that Excel models can be deceptively inviting of replacement ideas. They may have many workflows within them that are obviously much easier to accomplish with a different tool. E.g., grouping and applying aggregate functions; filtering; joining; filling missing data, especially with constraints such as filling forward EPS estimates but only into the same fiscal quarter; etc. You could look at an ocean of INDEX(,MATCH(...)) functions in Excel and discover that the same manipulation in Pandas would require just two lines of code. But it is easy to overlook the importance of seemingly simple calculation flows, in the same model, that are uniquely suited for being expressed in Excel, the universal language for describing calculations.

For example, I have seen data analysts move from Excel to Python, replacing their models as they go -- after being wowed by Pandas' data manipulation capabilities -- and then get bogged down for months trying to recreate what Excel's =RTD("BLOOMBERG.RTD","",...) already did for them. And when the portfolio manager tells them, "this number looks wrong", which happens often, they spend half a day dumping data into Excel and building a sheet that's illustrative for the PM.

>Excel models can be deceptively inviting of replacement ideas.

In my mind that's actually what is the root at any talk of "Why isn't there an open source alternative for X amazing program?"

As a wiser person than me once said, "It's easy to imagine so I imagine it's easy." Deceptively simple operations in things that have a long legacy of supremacy can just be near impossible to clone even if you can churn out a POC that looks reasonable in 20 minutes. That trap is a big hill a lot of us die on. At one point in my life I swore off ever doing ETL, and now I do it for a living. Seems like that's actually the world now. Sheesh. With MS embracing Python for big data, maybe the next big thing is just what it's always been... Excel.

Excel is the code and its interpreter. You don't want to lose the REPL flow for development, but maybe you can add an AOT compiler for production?

I worked six years in the IT department of a small domestic manufacturer, and deciding when a model had outgrown Excel and needed to be created as a tool (sometimes just a pretty interface for our ERP system) was one of our main tasks. I hate the hell that users can bring on a company when they misuse Excel but it will never go away.

We converted our excel wiz to do R. She will never return to excel for advanced stuff. The simple stuff could be done with lobte office (except for excel sheets that come from outside).

I'm not an Excel user. Genuinely curious: Is it really that the alternatives are worse, lacking features, unstable, or does it come down to "they're not Excel", in that they're not a drop in replacement that can perfectly load and run every Excel sheet ever created.

Because that would be my gut feeling: Not interoperable enough, since some things just work differently, but not necessarily much worse. If you can't send your sheets to company b without being sure they can properly use it, it's worthless. Pretty much the same problem with Word.

A PM that worked on Excel in the late 2000s explained it to me like this: Most people use about 20% of the features in Excel - which should make it easy for competitors to copy, right? The kicker is that different people/industries tend to use a “different 20%” of the features, making the barrier to really compete very high.

> The kicker is that different people/industries tend to use a “different 20%” of the features, making the barrier to really compete very high.

The thing is that different job functions that are exposed to overlapping user bases in the same industry (and even office) use a different 20%. If it was just per-industry variation, it would be easy for industry-specific competitors to succeed, but the parts of Excel used actively by Andrew, who makes a tool that is also consumed by Bob and Carol are different than those used actively by Bob and Carol (which also differ from each other, and which each have their own users who rely on the tool and the features it relies on, even though they don't actively use the features.)

It was that combined with Microsoft somehow finding a way to make it so the 80% of the features that you didn't use didn't get in the way. Same thing with Word.

Excel gives you the power of being a programmer without having to admit that you're programming. There are people who do absurdly impressive things with it, the CA-GREET model for estimating CO2 outputs is an excel workbook that took me a while to figure out how to even use.

But the barrier to someone using it is much less because it is gentler than, say, giving someone python code to do the estimate and expecting them to do it correctly. I think the interface familiarity is key.

Unfortunately, while LibreOffice is usually pretty good at all this, it's always just one or two peculiar difference in function calls that breaks it. It's pretty close, but just not interoperable for anything advanced.

My impression is that Microsoft has sweated the details of Excel, the way that Apple sweated the details of the iPhone, so that it is extremely smooth and quick. This in turn makes it literally less physically laborious to use. Things like casually swiping the mouse across a range of cells to select them.

I suspect making it work this way requires an army of developers and some corporate will power. Having an inside track on exactly how the OS works may be an advantage.

> sweated the details

I strongly suspect you are correct. As stale evidence of culture, I offer some writing by a software developer who used to work for Microsoft on Excel:

https://www.joelonsoftware.com/2001/10/24/user-interface-des...

I am an Excel user and nothing beats it in terms of functionality and add-ons but it's the win32 level of usability that scores it for me, nothing comes close to packing so much in to such a practical workspace. Don't even get me started on Project desktop and some key add-ons!

One problem with sheets is that its external. Most companies I've worked for prohibit use of external drives/email/apps.

The MS Office suite is far more user friendly than the alternatives.

I still use Libre Office on Ubuntu and I wonder why they chose to be so different. It would be great if they could completely copy MS Office in UI/UX, but there's probably some kind of copyright there right?

The thing is, OpenOffice (and by descent LibreOffice) did copy MS Office's UI/UX. Then Microsoft switched to the "Ribbon" interface, throwing all that shared knowledge out the window. Recreating the Ribbon in LibreOffice is an active development task, last I checked.

Is there any limitations on how much they can copy? Surely Microsoft have some of the design and specs locked down with patents

I don't recall the outcome of that "Apple claims ownership of rounded corners" lawsuit, but I suspect (being absolutely not a lawyer) that'd be the relevant case law.

This really depends on the Google v Oracle lawsuit currently. If Oracle wins then the remains of VisCalc will have a copyright claim against everyone...

Microsoft successfully sued Corel for design patent infringement related to the ribbon interface. Corel had a “Word mode” in WordPerfect X7. The damages were about a quarter million dollars. YMMV.

Oracle v Google is copyright as applied to API’s. Different laws. Patents expire more quickly.

It's a very effective form of distributed / peer to peer lock in.

Even if you individually (yourself or your company) would be willing to move to some decent alternative (they do exist), you will still be exchanging files with people who are more locked in than you (e.g. banks/finance), because of custom plug-ins, custom macros, etc.

If I wanted to compete with Excel I would start with a plugin / add-in or complement tool and work my way from there. Some companies are probably doing it. They're not really replacing Excel itself but capturing some of the value away from Microsoft.

Honestly the suite of Excel + Powerpoint is incredibly hard to replace.

The entry level to power user path of Powerpoint and Excel specifically make it easy to jump in and then get very powerful usage from very quickly.

Once you see Apple Keynote i never went back.

But true for Excel

Powerpoint's for sending information to and between C-level folks. It's basically the only file format you can send them that they'll maybe actually open. Its use in actual presentations is secondary.