Silo Culture at UBS

Following my note on Silo-Busting, I wanted to discuss a specific example - UBS.

On its website, UBS offers a full range of jargon-compliant banking services to other banks, including "straight-through processing" and a "modular client interface". It also claims that its approach is unique "because we implement this solution with you – and focus on integrating what have traditionally been functional silos". [UBS: Cash Currency Services]

I also found an advert on the UBS website entitled Global Cash Custody: Thinking Outside The Silos (pdf).

But how is this silo-busting reflected in UBS's own organization and business model(s)?

---

In April 2008, an internal report blamed "silo culture" for poor risk management across the group. In response, UBS merged the Group's Market Risk and Credit Risk functions into a single unit, with responsibility for the monitoring and controlling of the firm's overall risk exposure across credit, market and country risk, as well as performing portfolio analytics on risk evolution at the overall portfolio level. According to Joe Scoby, Group Risk Officer, "These changes are designed to establish a best in class Risk Control team with an overall view of all risks. The creation of an integrated portfolio and concentration risk group will help break down any remaining information silos between the different risk functions." [Reuters, May 2008. See also Risk Australia Spring 2008 and Online Financial News, June 2008]

Or perhaps not. On 12 August 2008, UBS announced that it was stepping away from its integrated model and establishing its investment-banking, wealth-management and asset-management divisions as stand-alone entities [Economist, 14 August 2008]. UBS chairman Peter Kurer said: “Our review has clearly revealed the weaknesses associated with the integrated ‘one firm’ business model. Some of these weaknesses – such as the blurring of the true risk-reward-profile of individual businesses – are the source of substantial risk, as we have seen in the past few months.” In other words, risks result from putting different businesses in a single silo. [Wealth Bulletin, 18 August 2008]

Writing in CIO Magazine, Chris Potts viewed this as raising two important questions for Enterprise Architecture: The Model or the People? [20 August 2008].

"Firstly, to figure out the balance in our own organizations between having the ‘right' business model, and everyone's collectively ability to manage it. But also, much closer to home, to remind us to strike an appropriate personal balance between modelling our organization's EA, versus influencing the people who manage it in practice and invest in changing it. "

I think there are two deeper questions here. Firstly, enterprise architects need to be thinking of a business as a viable sociotechnical system, which means (among other things) that management command-and-control capability must be included. Secondly, this management capability depends on how complexity is distributed across a large organization - the proper balance between differentiation and integration.

It is complete nonsense to imagine a straight choice between differentiation (completely different business model and organizational silo for each business line) and integration (completely uniform one-size-fits-all business model across all business lines). For me, one of the exciting aspects of service-oriented architecture is the degree to which it supports more sophisticated ways of getting the best of both worlds - differentiation AND integration. But this in turn requires a more sophisticated approach to business modelling. Which (sadly) many enterprise architects don't seem to be ready for.

It will be interesting to see how the current UBS experiment with organization structure pans out. This is certainly the sort of initiative that enterprise architecture should be able to contribute to. If anyone has any inside information, on or off the record, I'd be delighted to hear from you.