Nicholas Carr, author of the Harvard Business Review article titled “IT doesn’t matter” in 2003 followed up last month with an article called “The end of corporate computing". He projects a move to “utility computing”, similar to how companies moved away from proprietary power generation to buying it from large, more efficient utilities. It fits with his first argument that IT doesn’t matter - now go outsource most or all of it.
I am a big fan of outsourcing – done selectively, done right and done at the right price/service levels. But reading Carr’s assumptions around utility computing and his justification of IT outsourcing I ended up with 6 questions below. While Carr says his vision is “long term”, I am concerned (as you read question f) about the consultant “ecosystem” which will give his ideas short term momentum before they are scrutinized.
a) WHEN WILL THE ARCHITECTURAL UNDERPINNINGS BEHIND HIS ASSUMPTIONS MATURE?
Two of Carr’s underpinnings for utility computing are the maturing of grid (and related on-demand) computing, and Web services. IBM has been pushing its “On-demand” model for years now. Forbes magazine has a sobering account of its progress in its article “"On-demand" Letdown"
Services oriented architectures are still evolving in very divergent ways - see my blog titled SOUP – Services Oriented User Procrastination
Neither is mature enough as a platform for the ambitious IT outsourcing Carr recommends. Not sure they will be 3 years from now. If anything, it makes the case for more selective outsourcing in mature areas.
b) ARE BUSINESS EXECUTIVES READY TO BUY AND MANAGE TECHNOLOGY EFFECTIVELY?
Carr is cynical of the CIO and of corporate IT in general. After Y2K, ERP and ebusiness project overruns a number of business executives are also justifiably cynical. But to jump from there to assume a VP of Sales or a CFO could buy or manage technology any better is huge leap of faith. There is a reason why the Deloittes and SAPs try to sell direct to the business execs and try to bypass IT – because the margins are better. Show me any contract where IT was not involved in procurement and I can show you 25% to 50% fat that could be cut out.
There is also an assumption that business execs will “buy” more than “build” technology than CIOs do. Most CFOs I know are far more loath to outsource even low value stuff in their areas than their CIOs are. How many CFOs continue to cut checks, handle collections, process time and expenses? I was trained as a CPA – trust me I can come up with all kinds of internal control excuses to not outsource them. Here’s another example - I know a CIO who wanted to discontinue pricey annual maintenance from their HR software vendor and move to a lower priced, utility model, third party support firm. The project was killed by the company’s VP of HR who did not want to displease her “strategic vendor” - the HR software vendor.
I would like to see the conversation not just focus on outsourcing IT – let’s look in to outsourcing everything tactical.
c) WHAT ABOUT LABOR, NOT CAPITAL, INTENSIVE ELEMENTS OF IT?
Over 70% of IT budgets are in labor intensive components (software/content, external services, internal IT staff). The rest – hardware and telecommunications - are capital intensive and much more aligned with electric utility scale and economics Carr talks about. IT labor has been stubborn to deliver economies of scale/repetition around. EDS has over 110,000 employees. The average Fortune 500 CIO has 500 IT employees. IBM has 40,000 sales people – more sales talent than most Fortune companies have employees in total. Yet, in spite of this scale, IBM and EDS often cost more per employee than do internally burdened staff costs.
There is actually considerable excess capacity around IT labor – in offshore markets and “at home” markets. For all the noise around offshoring, all the offshore vendors together have less than 5 % market share of the global IT services market. But there are significant structural issues around using this capacity – political issues around offshoring, and cynicism around using “at-home” labor.
Let’s open up the labor debate, not just outsource it..and not let people like Lou Dobbs keep that conversation in back-room, whisper mode.
d) WHERE WOULD YOU BUY "INNOVATION" ON A UTILITY BASIS?
I agree with Carr IT does not often bring strategic advantage. It can, however, be a huge driver of innovation – even if in tactical areas. CIOs I have worked with in the last couple of years have a) rolled out Blackberries and mobile applications through their entire sales force b) rolled out WIFI to entire corporate campuses c) enabled RFID and GPS tracking throughout their logistics chain d) implemented sophisticated data mining around customer data. In each case the business productivity payoff was phenomenal. In each case, when they started these projects 2-3 years ago, even if they wanted to they could not find a systems integrator to pull the project together. Their internal teams had to pull together early, often beta technology from multiple vendors.
Not sure “utility vendors” will support future innovation needs – or that the VP of Sales above could have implemented the Blackberry program or the VP of Logistics the RFID initiative without solid internal IT support.
e) IS THE IT VENDOR COMMUNITY DISCIPLINED ENOUGH?
Carr extrapolates price/performance around hardware to all of IT. See my posting around “What if Intel was a services vendor?” Systems integrators have done thousands of SAP projects – in level of effort their work plans even today show hardly any productivity improvement since the early 90s. Accenture has had a Philippines unit for two decades. It essentially used it as a regional delivery firm. Till Indian firms showed that the global delivery model could work around software, Accenture did not much leverage its Philippines cost arbitrage for its Western client base.
Financial, MRP and other core functionality in SAP, Oracle and other ERP software was developed in the 80s. They recouped their investment a long time ago and still expect to charge similar pricing even today for even that functionality (you can understand the higher pricing for more recent supply chain or CRM functionality). A CFO at a fashion company recently marveled at ERP pricing “I wish I could sell even 6 month old merchandise at today’s prices”
Swapping out one electric utility for another is child’s play compared to swapping out a major enterprise IT vendor. Till most IT vendors show consistent productivity improvements and pass along savings of scale, we should be wary of the “lock-in” the “utility computing" model would give them.
Even around hardware which has shown huge price/performance gains over the years, there are substantial maintenance and other “add-ons” which negate the gains. Ask many CIOs and they will tell you EMC, the storage vendor has been one of the most difficult to manage costs around. Ask CIOs who are experimenting with Linux how much more they are saving compared to their Sun gear. In the utility model, would they have this pricing transparency?
Electric utilities are regulated. Not sure IT vendors, with their entrepreneurial roots, can ever accept being regulated. While we occasionally slap Microsoft or Oracle on their wrists, not sure our bureaucrats would even know how to regulate them.
f) HOW MANY STRATEGY CONSULTANTS WILL CAPITALIZE ON THIS "PARADIGM SHIFT”?
I have been in a number of clients where a McKinsey or consultant of similar ilk was brought in to do a “technology strategy” project. There was one project where one of these firms developed a technology sourcing strategy – there was no customization for buying PCs v/s buying software. No different thread for planning a software maintenance renewal or negotiating a hosting contract. My instructions were “Do not criticize the work this firm did – they have friends at the board level. But flesh it out 5 levels deeper”. Such firms happily offer to develop “offshore strategy” when they could not tell you the difference between firms like Aztec or Zensar, or the difference between EAI and PLM skillsets.
What Carr is talking about at his high level perspective will be read in a number of board rooms and airline magazines – and will bring such “technology strategy” consultants out in droves. Many probably went to school at Harvard and know Carr. Really smart consultants – just not very technology knowledgeable. But, they cost companies millions and their report plus six quarters may buy you a cup of coffee at Starbucks – may be.
MY VIEW OF THE WORLD
I am all for outsourcing - I constantly put up GE’s 70% IT outsourced model as a benchmark. But it has to be done well, selectively, and cannot just focus on IT. And to do it without the CIO is foolhardy. He/she is far more sophisticated than his/her other executive peers around process benchmarking, price/performance analysis and importantly in setting up service level agreements with vendors. He is also used to a much faster rate of change.
Utility computing, web services and open source, VoIP, blade servers, global delivery models are positive developments every CIO should take advantage of – but on his terms.
The CIO also needs to learn to swim through the sea of reports “technology strategy” consultants I mention above are about to generate quoting Carr early and often.
Comments