Been thinking more and more about the relationship between these concepts:
- IT portfolios (application, project)
- Service catalogs
- Enterprise architecture
I've discussed portfolio management here, and have been doing more research into it; spent some time today on the phone with Chris Bell from Prosight who helped educate me on the worldview, and confirmed again that there is a fundamental difference between project portfolio management and application portfolio management.
Portfolio management as a term of course comes from investing; the literature I've reviewed from both Niku and Prosight is full of terminology like "balancing" and "investment." Applications would be seen as existing investments, with a lifecycle (e.g., buy/hold/sell or build/maintain/retire).
The portfolio management discipline involves quasi-financial modeling; one can play out various investment scenarios and optimize them. I am still trying to understand just where the data comes from, and the relationship between quantitative/qualitative analysis; application management, with issues like change turbulence, incident & problem management, and other non-financial metrics would seem to call for a balanced scorecard approach, not simply financial.
The decision to re-balance or refresh an application portfolio results in project portfolio investments; I think the word 'portfolio' is a little questionable with respect to transient things like projects but the precedent seems to be set. Is project portfolio 'management' the equivalent of day-trading? Seems once you've made a decision regarding your application portfolio, the projects are simply the "buy" and "sell" orders, not really a portfolio in and of themselves. But I'm getting philosophical.
A service catalog is an IT Service Management term. Rather than an investment portfolio, the paradigm here is merchandising: we have these offerings (applications, services, processes) for sale (or rent). If you purchase them, we commit to delivering them at this level of service. If we fail, you can hold us contractually liable.
The service catalog seems to be more operational in flavor than the application portfolio. A savvy investor would of course be very concerned with how the offerings of their companies were being received, but the service catalog per se wouldn't necessarily have all the information needed to assess itself; the service catalog is just that, a catalog -- not all the operational information associated with every item in it, which is what is needed for the portfolio management process. However, having established a formal service catalog, it would seem to be a good thing to have the tools and processes supporting detailed record-keeping with respect to those services.
Service level management if so operationally supported would of course provide invaluable information to the portfolio management process, such as how effectively a given service met its obligations.
Finally, enterprise architecture is perhaps the keystone of them all. It means many things to many people; the major concerns seem to be applications, functions/processes, and data. Modern enterprise architecture tools allow the user to model and relate all three, usually from a top-down orientation. The Zachman framework is probably the best known conceptual model in this area. Some enterprise architecture vendors, like Popkin, are starting to shade into the Business Process Management space. Enterprise architecture has not historically done much with financial or service management information, which is a notable weakness.
So, what's the relationship? Enterprise architecture, as probably the oldest concept, would seem to have things to learn from portfolio & service management. However, I would argue at the end of the day it should be one team owning all three things: the IT portfolio, the service catalog, and the enterprise architecture, with as integrated a toolset as possible. In fact, enterprise architecture teams are among the leaders purchasing modern portfolio management tools (and probably having headaches with building interfaces between these new tools and the EA tools, but that's another story).
One final note on comparing/contrasting portfolio management with enterprise architecture: from a portfolio management perspective, process and data don't seem to figure as prominently as applications, although they both have their partisans. Capturing financial and nonfinancial metrics is hard enough for applications; for process and data management it's even harder, if possible at all. (From a service perspective, you don't want to decompose process/data/app at all; what matters is the service, not what makes it up.)
Just one more example of how the world of IT tools and concepts is converging.
Thoughts?
-Charlie
