The pressure on IT to contain costs is well understood. Every vendor has a presentation about how their products can be central to lowering costs. Research companies are publishing cost containment frameworks and consultants have lots of white papers.
Over the past several months we have worked with a number of clients to help them in their cost containment efforts. At 50,000 feet, it appears that everyone starts with the same playbook – renegotiate with vendors (starting with your telecomm vendor), cut consultants and contracts, virtualize everything, defer some maintenance, cut travel and training…and hope that is enough (so you don’t have to cut permanent staff). Frankly, this approach gets many CIOs down to an expense level that is “acceptable” to corporate leadership. Many CIOs also “accept” that this approach will likely increase long-term IT costs as well as inhibit long-term business strategy.
On the other hand, those clients who seem to make the best tactical and strategic decisions, when executing a cost containment exercise, appear to have two consistent characteristics.
First, they are very disciplined in their approach to ensure a comprehensive review of all areas of IT costs (i.e. IT Procurement & Asset Management, Product & Technology Rationalization, Application Portfolio Rationalization, Server Virtualization, Open Source, PC Costs, IT G&A etc.). Even ‘un-thinkables’ are reviewed to ensure completeness, for example:
- “Are there staff who can use OpenOffice instead of Microsoft?”
- “Should we re-examine our commitment to clusters and look at horizontal scaling?”
- “What are the implications of replacing EMC with NetApp?”
This discipline includes doing an objective and ‘truly loaded’ cost analysis. ‘Truly loaded’ refers to being sure that all costs are fully accounted for. For example, in looking at the cost of internally managing Microsoft Exchange and BlackBerry Enterprise Server versus a specialized outsourcing option (e.g. MailStreet) they make sure they include the cost of internal staff’s travel, training, PCs, cell bills, office space, etc. as well as the full loaded data center/support costs (e.g. Help Desk, space, storage, electricity). It is very clear that internal IT staff significantly underestimate the ‘truly loaded’ costs of providing specific services.
Second, they make sure they have adequate knowledge to make their evaluation. They don’t rely on speculation or anecdotes and, instead, look for hard facts and case studies. And when they see an example they really work to understand what are the dynamics in play at companies like:
- Wal-Mart – which uses SaaS (Red Prairie) to optimize employee scheduling at their stores versus running a similar application in house
- Sabre – which executes 60% of all travel reservations in North America on a server farm using very inexpensive HP Intel servers running Linux and MySQL, which saves Sabre approximately $20 million a year in operating expenses in support of AA.com, Travelocity and several other travel-related sites
These clients recognize that a meaningful evaluation of Cloud Computing requires an in-depth knowledge of the types of offerings available, and comparative strengths and weaknesses of each category of offering as well as the different vendor offerings in the category. These clients make sure that if they are evaluating an open source alternative to Cognos (e.g. Pentaho) that it can meet their needs.
The bottom line is that ‘discipline’ and ‘knowledge’ are key characteristics of cost containment assessments that yield the best tactical and strategic decisions. (Based on our experience working with clients, we have developed a Cost Containment Workshop that helps ensure these are achieved.
For more information on this topic and to read a description of a new IT Cost Containment Workshop check out the entry on the Education and Seminars page.
Highlights of IASA World Summit 2010 from EAdirections
October 4, 2010The IASA World Summit was another interesting event from Paul Preiss and his crew. Seems like the speakers get more interesting and relevant to EA practitioners at each IASA event. The ongoing debate on “Is EA a profession, a role, a specialty or something else entirely?” continued in full force.
Scott Ambler, IBM’s Chief Methodologist for Agile and Lean, spoke of his impressions of the roles and responsibilities of architects in an Agile environment. As always, when listening to Scott, I learned a lot and agreed with most of what he had to say. Since Agile is a development approach, most of what he talked about was pertinent to software architects, solution architects, project architects and application architects, but not enterprise architects. Not to speak for Scott, but I believe that is a distinction that he did not make, but would agree with. Scott?
George and I participated in a lively panel discussion with Scott Ambler, Paul Preiss, and Alan Hakimi, a senior architect in Microsoft’s Cloud Architecture group. A special thanks to Andy Ruth, who quite ably moderated our panel with a great set of questions and a little bit of an edge. We had a great discussion on topics like CEO elevator speech, best traits of an architect, the value of an architect spending time in the trenches, metrics for architecture, and a variety of other topics that the audience seemed very interested in hearing about.
We concluded our participation with a presentation on the roles and responsibilities of various architects across the different levels of enterprise architecture, including a discussion on our position on the separation of business architecture from IT architecture. The presentation was received very well. For a link to the whole presentation, see my post entitled “Architect Roles and Relationship with EA.”
An interesting, although not unexpected, bunch of side comments and discussions throughout the event held a common theme – the impact of cloud computing on EA, IT and business models. George and I will be sharing more of our thoughts on this over-hyped, but definitely relevant theme over the next few months.