Hard problems in banking

I’ve spent the last two days at the Xerox research centre in Grenoble, France, looking at some of the work their scientists are doing. And last month, I visited the Microsoft Research Labs in Cambridge. There was a significant difference between the two, and not just because they are working in different fields. Xerox, whilst it drives forward with pure research, seems to have its focus well and truly on commercialisation and partnerships with industry. Microsoft, by contrast, seemed way more academic, and in fact, made it quite clear that they (as in the research arm) didn’t do industry partnerships to any great degree.

From an innovation perspective, the track record of Xerox is very, very clear. They dreamed up the first windows based interface, the first mouse, created Ethernet, built the first laser printer, and have spun off a range of companies the likes of which include Adobe and Documentum. Microsoft’s innovations have also been startling, if less visible. But Microsoft has constrained its technology transfer so that you don’t necessarily see all the hard research that is under the hood. For example, the face tracking based on who is speaking in RoundTable, or the automatic image correction features one sees in some of the imaging products.

From a computer science perspective, the minds at Microsoft are doing very, very significant research, and I can’t tell you how impressed we were when we visited. But all that effort hasn’t, to date, translated into anything that is industry-game changing. At least, not compared with the long history of Xerox.

Banking is a game where everything is undifferentiated to a surprising degree. You can’t really do much that really makes you stand out, and even when you do, it is copied quickly. That is why getting our hands on technologies (or processes, or cultural changes) which haven’t yet seen the light of day are one of the opportunities we are pursuing. You can imagine the market advantage if we find something that is truly unique, even if it requires some effort to make it all work in the real world.

My own view is that corporate research facilities, staffed with all those bright people with doctorates, don’t necessarily see how their fundamental research can translate into good business for the customers of their company. And why should they? They are scientists, not bankers. Their interest, and skill, is in solving hard problems.

Bankers, on the other hand, do not go out of their way to solve hard problems. Hard problems are the ones that we’d prefer to avoid. They lead us to “bleeding-edge” pain, and require skills that are probably beyond what we have in-house. When you get a hard problem, we tend to say its too hard, and focus instead on incrementalism.

Perhaps this one of the reasons that innovation in most of the financial services sector moves forward at the slow, almost glacial pace it presently does.

My point with all of this is that corporate research facilities are a storehouse of solutions to hard problems, and the hard problems are usually going to be a source of differentiation, and probably one with a quite high barrier to entry. If we were to help Xerox or Microsoft, (or anyone else for that matter) commercialise something that they might not have focussed on otherwise, we would have an advantage that we might be able to reinforce with intellectual property and other protections.

Typical hard problems? Recognising faces in a crowd. Mining data for which there aren’t any existing relationships across systems. Getting to the posterior risk density of a portfolio of trades more regularly than overnight. Determining the concentration risk of a basket of loans in an unrelated corporate portfolio… it is a list that seems to be never ending. And one that grows ever time some researcher finds the answer to another hard problem.

Hard problems in banking can be addressed by some of the most current research thinking, if only the researchers are in touch with banks. It is an untapped opportunity that we are just itching to exploit.


2 Responses to“Hard problems in banking”

  1. September 20, 2007 at 2:40 pm #

    Data Mining — beer and nappies.
    A couple of years ago Teradata did a press mailing of nappies rolled into a beer six-pack carton. Obviously would have preferred it the other way around, with beer wrapped in nappies for protection in transit. It was to illustrate the data mining story about young men buying nappies at end of day and then buying a six-pack. Turns out to be something of an urban myth (http://www.regdeveloper.co.uk/2006/08/15/beer_diapers/) but it does suggest that one duty of IT users is to think of the questions and then ask computers to check. If LTCM had asked what happens to highly leveraged positions when liquidity disappears it might have avoided its squeeze.
    Good thing that lesson has been learned so well.
    Of course, the next step beyond BI is action. I was recently on a panel with a Wells Fargo banker — they are running around 5 relationships per customer and aiming for 8, which they say still gives them plenty of opportunity since the average person has something like 16 financial relationships across checking, investments and insurance.
    This potential has been a staple of banking conferences for years — but only a few banks, and perhaps Tesco, seem to have done anything about it. All those investments in CRM, and for what?
    IT can do great things, especially when users can develop imaginative questions and take useful action.

  2. April 1, 2008 at 9:04 pm #

    I think this recent announcement from Bank of America gets at what you are saying in your post.

Leave a Reply

Your email address will not be published. Required fields are marked *


Proudly powered by WordPress   Premium Style Theme by www.gopiplus.com