Hacker News Reader

Comment by foolsgold | Hacker News Reader

I worked for one of the largest financials services companies in the world and they STILL use excel to drive their trading activities. They are so large and so complex that it's impossible to convert them because no one completely understands how it all works. They were developed by traders/market experts with no help from IT, for years IT didn't even knew they existed.

These spreadsheets are truly amazing, I've only seen glimpses of these spreadsheets but they were massive, nothing like needed to have servers in the local network cabinet and running cables for keyboard, video and mouse, because you needed 1/2 a Tb of ram in order to open the spreadsheet.

View on HN

Also worked for some large financial firms, and can confirm that there is a disconnect between Excel models created by traders and IT. One interesting development was the acquisition of ClearFactr by Goldman [1]. Seems like they hope it will centralize some of these models into one sytem, making them accessible across the firm (Excel sheets can even be imported).

GS has had a centralized risk/pricing system for over 20 years now that is pretty powerful [2]. Fun fact: GS was able to calculate their total exposure to the Lehman collapse 12 hours after it happened using this.

1. https://www.businessinsider.com/goldman-sachs-buy-financial-... 2. https://news.efinancialcareers.com/uk-en/276170/secdb-quartz...

The SecDB article [2], is very interesting!

I worked at a financial services company that created portfolio analytics software. My job pretty much consisted of turning our analysts' excel models into high performance Scala code.

The company wasn't very successful, we just couldn't convince portfolio managers to pay for our software. Even the analysts at hedge funds didn't like to use our software. My theory is that a bunch of analysts need to justify their jobs with incredibly inefficient and error-prone excel workflows. Our software would have eliminated a lot of what they do in a day, and therefore, would eliminate their jobs as well.

It is terrifying that trillions of dollars of assets are run on excel, but the inertia is too strong.

Sounds about right to me! One of the companies I consulted for had a billing department that insisted people pay their utility bills by check. They said "this is the way most of our customers like to pay".

But what they really wanted to say is "we will have to repurpose 200 staff who process these checks and mail out statements, and they're all 30 year veterans, and I don't have the heart to tell them they're fired even though they are the sole reason we lose money"

Excel also gives people the comfort that they can track what goes in and out and adjust easily.

Giving up that control is hard.

> My theory is that a bunch of analysts need to justify their jobs with incredibly inefficient and error-prone excel workflows.

Even if you're right, pretty much any company whose customers aren't 100%, unimpeachably good is going to fail.

Another way of thinking of it is, as soon as they became even slightly an antagonist (think Disney rivals, less Autobots), it was doomed.

As an aside, an alternative interpretation would be that Excel workflows are more intellectually stimulating, so as a side effect they come up with better models. So even if your software had a superior UX for the task at hand, the objective of making money is better served by people using their brains harder generally.

After all, a lot of people practice violin and play chess because it feels good, and then those people make a bajillion dollars on the stock market. Consider that if Excel even remotely reenacts the patterns of those activities, which is almost certainly does, it makes money in ways besides calculating numbers in a user-accessible way.

I think your last paragraph is really spot on. Being an expert at excel makes you feel like a wizard, just like how emacs or vi or the bloomberg terminal makes you feel.

Our software didn’t have that feeling, it didn’t make you feel powerful in the same way that excel does (I lobbied hard for us to make a desktop app, no one likes how web apps feel. But nobody listened)

I have a humorous thought. Maybe it's that excel models require humans to run. The code and data is way messier and it's partially locked up in the heads of the people running the sheets. It's harder to copy, so it's harder to steal.

One cyber-breach and all that really nice clean code is readily understandable to your competitors.

No, it would be terrifying that trillions of dollars are run by programmers. If the hedge fund guy says that Excel is fine, then Excel is fine.

Principal number one; don’t overcomplicate yourself. Keep it simple, stupid

Well considering their terrible returns net of fees and the massive outflows into passive, I think it's about time the industry did some soul-searching.

I don't think it's the fear of losing jobs, it's more because of the flexibility it offers. There's a lot of adhoc adjustments / corrections to data / formulas that require a lot of flexibility and a coded program won't be able to manage it that well. I used to do a lot of my stuff with vba but in the end I would still use excel formulas when I could

People outside of finance really underestimate how important excel is in the finance world. It really is the lingua franca ( or the universal software ) of the industry. It is so entrenched that even google hasn't been able to move the dial. One of the major selling points of google spreadsheet is that it is free. What google didn't realize is that finance companies are wealthy enough to afford to buy or license software. As poor high school or college students, we all used google spreadsheet, but in the industry, we all use excel.

There are so much sunk costs ( decades of excel worksheets and decades of technical know-how ) that most companies aren't going to switch to another spreadsheet and migrate all that data and retrain their employees.

As a long-time finance developer, I can attest to this. I think the only company that will be able to kill Excel's dominance in finance is Microsoft, and sometimes I think that they're trying. It seems the past few releases (basically 2013 on) have alienated a lot of power users by needlessly shuffling the menus/toolbars, adding lots of graphics that slow down sheets and add nothing for people who already know how to use the app, and not keeping a laser eye on memory consumption/performance issues.

In our company we also have a lot of excel files. Most things should done different, but then we would have to work WITH the IT guys, but they have even to solve a simple ticket more than a week, some tickets are open for more than a half year. That's the problem why we do so much as possible without the IT guys. I know that's bad in so many ways ..

Excel just added similar functionality built in: https://support.office.com/en-us/article/get-a-stock-quote-e...

I can see that. Traders in Wall Street are too practical to build something maintainable. Now G Sheet brings the same power but to software developers this time.

> Traders in Wall Street are too practical to build something maintainable. Now G Sheet brings the same power but to software developers this time.

When did software devs lack the power to write trading tools? The issue is that the traders had the knowledge and wrote their ideas into a spreadsheet which was not only maintainable for them but easily extensible. I cannot overstate that second part. They become spreadsheet behemoths because it's trivial to make them do that one extra thing (at least for the first fifty or so features).

Reply on news.ycombinator.com