Disclosure: I work for one of the biggest US investment banks.
They fed us this line during training, and often senior leaders who left our company would say the same thing when interviewed on CNBC. This is naïve at best though. Banks are only recently starting use "Big Data" (how I love this buzzword), because they have unmaintainable mainframe systems running on LPARs. The reason it's "Big Data" is that as time goes on, the retention of financial data will increase as well. Banks are definitely NOT technology companies, at least as far as the compensation structure is concerned ;)
Don't believe me? Compare any bank CTO to an equivalent position in front office.
I think you're sort of right. Banks don't think of themselves as tech companies. Their culture, priorities and upper managemet aren't those of a tech company.
But they are now tech businesses. Here's an example: Last year RBS made a mess of deploying some code[1]. It shut down their online and telephone banking services as well as stopping money transfers for a few days. Their ATM and branches were still giving out money but this pretty much shut the bank down because people rely on transfers to get paid and pay their bills. The backlog of transactions took days to work through and the whole thing cost the bank 100s of millions in compensation and damaged reputation.
Their branches remained open, but it didn't matter. What people need banks for now is the storage and processing of electronic money transfers. They are a SaaS business with a branch network bolted on. They just haven't realised it yet.
Right, but being a "tech business" wouldn't have prevented the root cause of that, which was laying off their experienced workers and offshoring to a cheap location. Tech businesses make that kind of blunder all the time.
I completely agree with you that currently most banks only pay lip-service to modern technologies, even if they throw buzz words around like crazy. Mainframes, COBOL, etc, are still the standards in many, if not most, large financial institutions.
My point is that when you actually boil down what a lot of commercial and retail banking is to it's core (I would exclude capital markets here, mostly for simplicity) then it's a technology and data problems. It's a technology problem that is surrounded by legacy systems, culture, and regulations.
Ultimately, though, I believe that those banks that can most quickly infuse themselves with modern technology and -- most importantly -- actually run themselves like technology-first institutions, will be the ones that have the most success in the future. I'm not sure any bank is currently doing that, exactly to your point about compensation.
I agree with your point on commercial and retail banking. I also agree that if banks pivot in the direction of "technology-first", they will succeed in the future. I'm a bit skeptical (perhaps I'm biased in this regard since I work for one) that they will pivot in the near future though.
I've worked with several hundred regional banks and credit unions.
One thing people forget is that there are over 14,000 banks and credit unions across the country, not including the top 10. The challenges of regional FIs is dramatically different than that of large organizations. (Note, most comments on HN are from large banking backgrounds, very few from the other 14,000).
FWIW, the two themes I have seen emerge are:
1) Regional banks outsource a lot of their technology, often times leaving them with the inability to quickly adapt.
2) The regulatory environment drives technology and has created a gun shy approach to tech.
3) Most CEOs of banks and credit unions tend to have some type of finance background as opposed to technology.
I also work in tech for one of the big investment banks and agree with the author's thesis. Essentially, the banks are going to continue to replace their processes with technology solutions in all areas of the institution, and it's already happened in many parts.
It's just a matter of them moving slowly over time to make sure nothing breaks. Security, not negatively affecting customers, and not losing any data are the primary priorities. Those priorities are what keep adoption of technology slow.
This is a really naive view of banks. Big banks, especially investment banks are mostly politics-driven institutes. Technology is important, data is important, but those two are not nearly enough to run a successful institute.
I think everybody knows that who has been reasonably close to people actually leading these institutes. Look at GS, MS or the rest - the CEOs and other top managers are many times coming from the government, and in many cases, ending up in government.
Small hedge funds, prop trading shops are big data/tech companies, that might be true. Even in those cases knowing the market (which means significantly more that knowing the hard data) is of paramount importance.
'Beating the market' is not how these institutions make most of their money. Sales, politics and other kinds of human interaction is a make or break factor.
> I think everybody knows that who has been reasonably close to people actually leading these institutes. Look at GS, MS or the rest - the CEOs and other top managers are many times coming from the government, and in many cases, ending up in government.
This doesn't disprove the author's point. The author is pointing out that a lot of technological innovation (eg - ' such as application programming interfaces') is churning inside of banks with no notice to the general public. This point is completely removed from politics and CEOs. The article just points out that there is a lot of software development within banks.
> This point is completely removed from politics and CEOs
No, it's not. I happened to see the biggest investment banks' technology departments from the inside (i.e. I worked on very core technologies), and this statements is just simply false. Yes, there's a lot of software development going on, but it's quite a slow-moving animal. They _used to have_ a lot of cutting-edge stuff, but nowadays most of these are superceded by better engineered implementations available to the public. And this is a big struggle for them, the more they've invested in proprietary IT decades ago, the more money they made with it around the late 90's, and the more difficult it is now to move towards industry standards.
I've been working at banks and financial institutions for years, and am currently working at a large investment bank and do not agree with you.
Yes, these large institutions have to support older technologies, but they are still creating innovations in the way the author suggests. There are several products out there available to the public that are better than what some banks are putting out, but the products banks are making for their core clients that aren't the general public as very good and relevant to those niche users.
I agree, banks and financial institutions are data driven businesses.
I have seen several banks from inside, and most of them think that they are driven by financial analysts, figuring out clever equations for managing money (e.g. lending rules, mortgage policies, etc), while IT is just a cost center for tracking the transactions. It's old school, and they change very slowly.
Sooner or later they shall realize that the transaction data they are tracking is a good source of information, and they could use "hard" data (through some machine learning and data mining), instead of or alongside of the "soft" intuition of their analysts.
This is not true. Transaction volume for consumer banking, even including credit card processing, is minuscule - a heavy customer might generate a half-dozen transactions over the course of a day, and many of those (particularly credit card transactions) are able to be asynchronously processed over the course of days or weeks.
Ultimately banks are in the business of money management, not money tracking and have been for the 700 or so years they've been around in recognizable form.
Whether they could do this with less overhead than they presently do with new technology is a separate issue.
I beg to differ. In a world of fraction-reserve banking, banks are in the business of capital aggregation and debt issuance - in order to generate all important "spread". But, that world - the world for the past ~700 years - is disappearing.
We live in a world with collapsing credit spreads and Basel III will tighten them even further. In the US, Dodd-Frank's prohibition on lots of fee based income (which was a huge profit center for banks) and the disappearance of debit inter-change due to Durbin makes banks very concerned about the future of their business.
Banks are increasingly turning to transactional systems to make money.
I have a friend inside Capital One who tells me they're very concerned about the future of the industry. In a nutshell, C1 makes money three ways: (1) debit interchange, (2) net interest margin (NIM) and (3) fees.
Dodd-Frank basically killed debit interchange and fees (esp. with the overdraft changes). If you think spread will keep collapsing, I'm not sure what's left revenue-wise.
Eh, consumer banking sure, but most focus in banking is rarely around the small consumer, as most people with just checking accounts cost the bank money because of how little they interact with the bank financially -- They only become profitable with 1) a mortgage or lots of debt 2) a lot of money
I worked at one of the large banks on a commercial operations team (terrible job btw) -- the sheer amount of data surrounding transactions that came through our office alone was ridiculous and barely manageable.
Commercial banks have substantial trading activities on their balance sheet as well. I'm not sure I understand your distinction between management vs. tracking, but systems that forecast the balance sheet to assess interest/market/credit risk need to be fairly sophisticated. Running a bank effectively goes well beyond correctly tracking deposits and other transactions.
Good point Zac, It's interesting to see how banks are shifting away from traditional sources of revenue based on trading, balance sheet leverage, advisory, etc into potential new sources that involve the monetization of data more like a traditional technology company. I'm curious if there will be a dramatic cultural shift within banks as traditionally technology has always been viewed as "back-office"...
This reminds me of a VMWare acquisition last year of a cloud automation firm DynamicOps. It turns out that DynamicOps was actually a spinoff from one of bulge bracket investment bank Credit Suisse's IT groups. The bank appears to have been quite supportive, complete with venture funding. Quite an interesting story of how tech innovation can come from some unexpected places.
It really depends on what team you're on. When I worked at a bank, I worked on a pretty competent team. We didn't do anything that important, but it did work pretty well... we had automated tests, continuous builds, and all that good stuff. Most people could open up someone else's code and work on it. (This was largely due to the efforts of one engineer on the team, but you have to give credit to the management for letting him do it :)
At the same time, I would work with other teams, and it just amazed me the kinds of problems they caused. One team worked on streaming data from spreadsheets to other spreadsheets. (Don't ask me why all market data was calculated using spreadsheets.) This was implemented as a set of servers that read from and wrote to a proprietary multicast network that was usually down. It didn't matter, though, because the servers themselves "health checked" each other, and if it detected one was responding too slowly, it would kill and restart it. Except sometimes the killing and restarting caused that server to fail to respond to health checks, so the system basically consisted of servers killing each other all the time. (No, it didn't actually work despite all this. It was a constant source of problems, and was over a million lines of C++! It couldn't even quit without core-dumping because of memory corruption!)
Another team I worked with wrote their application on Java. The company had a policy of only allowing developers to develop on Windows machines, but this team needed to deploy to Solaris. Their solution was to develop and test it on Windows, copy it to the Solaris machines on launch day, and turn it on. But a Windows-only library crept into the core. If this were rocket science, we would not have gone to space that day.
But, there were some pretty good initiatives. One was for a single team to rewrite all software at the bank, in their own C++/C#/Python hybrid, with their own IDE, making all other software engineers obsolete. And despite how crazy that sounds, they actually did a good job! They gave every obscure system at the company a unified API, so a normal user could stitch it all together with a tiny Python program. And write tests for it. It was rough around the edges while I was there, but obviously the way forward. (The big advantage: middle managers had no control over technology, only making a product. So there were no debates ... or rather, horrifyingly bad decisions ... over what servers to buy, what language to use, who would pay for the sysadmins... all they needed to do was design and implement. It was brilliant, and I wish them the best of luck.)
When I worked at GS a few years ago, they often talked about the "high tech" and "high touch" parts of banking.
You're right that transactional, lower-value activities like basic commercial banking, credit settlement, mortgage origination, etc. are at their core, technology problems, but a lot of banking, especially on the higher, relationship-driven end (private banking, investment banking) is more about customization and individual client service than technology.
Mentioning it because I find that model (tech vs. touch) useful.
I've worked at one of the largest Bank's in the US for the last fifteen years, and I pretty much agree with the author. I first discussed this with a co-worker ten years ago and at that time I took the contrarian position but looking back now I think he was right even then. This is a huge shift in culture and focus but its recognized pretty much across the board, even I think by the majority of people who will not be able to adapt and will be left behind.
Except for the largest banks, all the ones I've seen don't write much original software for their core business - they buy packages and integrate them. They then rely on the vendor for support.
This article really rings true for me. I am a software engineer at a then software startup that is now acquired by FIS. When I first started at this small startup it has the makings to be technology based. Then we hired individuals who thought of the business purely from a financial services point of view and slowly killed the culture. It muted it and dulled it's appeal. I wish everyone at standard treasury luck and disrupting this space in a big way.
I'm very much not convinced by the idea that banks have technology at their core with a mere wrapping of business functions around it. Having worked for a bank for many years, I'd say it's exactly the other way round.
Management of capital, liquidity, risk, etc. are the true core banking elements.
"There, far from the public eye, new technologies – such as application programming interfaces – are allowing financial institutions and businesses of all sizes to interact in new, faster ways."
This "new" technology called API sounds amazing. Everyone should start using it.
Ultimately all companies will be technology companies. Software is eating the physical world pretty quickly. Taxi drivers are going to be on the short end of that stick in the next decade or so.
Banks as information and technology companies is a great insight. I think we'll start seeing more of this creep into the noosphere as p2p public ledger systems (like Bitcoin) becomes more mainstream.
I think it will also lead to a reforumlation of the concept of currency. Rather than defining and conceptualizing currency as a store of value, it becomes defined and conceptualized as data.
Arguably, the practice of accounting is already treating exchange of currency as data. However, popular concept of money is still tied more with the notion of "value", and emotionally with how much work and suffering one puts up with to trade for a reward. Maybe this popular notion will shift.
Online financial companies (paypal, stripe, square, kickstarter, etc.) are 0.1% moving bits around, 0.9% ensuring that the moving of the bits is robust, and the rest is a bunch of other stuff. Things like fraud detection and mitigation, relationships with other financial institutions, relationships with sovereign governments around the world, familiarity with relevant laws and regulations, customer service, and so forth.
I can set up a physical bank in 15 seconds. All I need is a shoebox (the vault), a notebook (the ledger), and a pen. And, indeed, in some circumstances this sort of bank is actually appropriate. But at the scale of worldwide commerce it is no more appropriate than a "banking" system that thinks that the software is the most important component.
Most of us who work in banking IT, are lured in either by the proported larger salaries, or the opportunity work on big systems, and interesting problems.
The reality is a little less glamorous.
The salaries can be larger than average in some roles, but then again, as you will be working longer hours, your overall hourly wage will be somewhat less. Also, bonus? Interesting. You can work hard all year, deliver, be on track for a nice share of the pie, then bang! Some tard on the desk does a run of bad trades, or worse, hides losses, or the firm gets hit with regulatory fines, then bang - either a small bonus or more famously, a doughnut (zero bonus), with the promise that next year will be better. Just work as hard or harder, and maybe, just maybe, a bonus will come later. Maybe.
If you have ever worked in a competent professional IT setup, your eyes will bleed, your heart sink, your soul evaporate when you are exposed to the full horror of banking IT.
Sybase ASE - you will see schemas with 1000's of tables and many more stored procs - not because they are all needed, but because each one is a sticky plaster fix on the last sticky plaster fix. Oh, and almost nothing gets retired, it just grows like a slow cancer, consuming all that get near to it.
Offshore and nearshore vendor staff will "own" and manage and support the infrastructure, help desk, request system, application support and maybe even more. It will be hell on earth just to get access to one of your application log files to help troubleshoot why johnny-fucked-the-pooch trader tard tries to blame the application for his poor selection of trades.
You will watch so-called partners and peers travel with ease and get what they need, as they work 'for the business', on a business cost-code. You will be on an IT cost-code, and you will need director sign-off for a trip to the bathroom.
In short - the banks need IT, and they are basically fucked without it. But then treat it worse than the department that fills up the vending machine with coke cans each day.
The reason is because YOUR management, IT management seem destined to fuck you harder than the business ever would. I never knew people could stand or walk without a spine, till I saw IT management.
What you describe isn't restricted to banking. I've never worked in the banking sector, but every software engineering position I've had at a company that wasn't a technology (and specifically a software-driven technology) company was in "the IT department." And "the IT department", even in places where IT has a huge impact on the orderly operation of the business, in my experience has been treated like a necessary evil (at best). I will never work in such a department again because of the shit-tastic experiences I've had in them.
Also your description of IT management in these places as merely spineless is far too charitable, in my opinion. Words that come to my mind are "incompetent, hypocritical, dishonest, creature of corporate politics."
seems like you need to pick be more picky next time you switch roles, as very little of what you said applies to my development role at a bulge bracket bank.
our management (upto MD level) are either developers or ex-developers, the hardware/support teams are excellent, and we can ever look at logfiles directly on production!
They fed us this line during training, and often senior leaders who left our company would say the same thing when interviewed on CNBC. This is naïve at best though. Banks are only recently starting use "Big Data" (how I love this buzzword), because they have unmaintainable mainframe systems running on LPARs. The reason it's "Big Data" is that as time goes on, the retention of financial data will increase as well. Banks are definitely NOT technology companies, at least as far as the compensation structure is concerned ;)
Don't believe me? Compare any bank CTO to an equivalent position in front office.