Posted by: Richard Watson
(Mike Rollings and I are guest posting to each other’s blog today to give our readers another perspective – check out Mike's post on role reversal at the Application Platform Strategies blog here.)
Robert Hayes from Harvard Business School posted a provocative blog entry last week, “Outsourcing Is High Tech's Subprime-Mortgage Fiasco”. This post helpfully provided me with some new language and clarity to add my take on what Burton Group is describing as “Externalization of IT”. The desire to use external providers to replace or augment IT functions presents a menu of options including cloud computing (at its multiple tiers), and outsourcing (including off-shoring).
Here’s my Homer Simpson one line summary of causes of the financial crisis:
Institutions did not price risk correctly because they were making too much money.
What was the lesson we learned from IT business process outsourcing (BPO) efforts, (primarily off-shoring software development) in the mid-2000s?
Enterprises did not price risk correctly because they (believed they) were saving too much money.
Many of these efforts were me-too strategies, driven entirely by short-term cost saving outcomes. For IT organizations that couldn’t adapt, an expensive repatriation exercise for some IT functions has cost who knows what. The costs were both hard: penalties for extracting themselves from BPO contracts, rehiring local development expertise; and soft: lack of business and project continuity, and the loss of trust with the business customer.
Robert Hayes again:
The supposed savings they expect to generate from such activities are based on costs that often do not properly reflect the damage they are causing.
Hayes is primarily discussing high-tech manufacturing, but the same arguments are valid considering software delivery. Shifting those developer FTE numbers from the New York or London rate to the Bangalore rate on a budget spreadsheet got the CFO excited, but hid the damage caused by dismantling a (possibly distributed) team knitted closely together by culture and shared goals. It also massively underestimated the complexity of rebuilding “industrial commons” as a hybrid of internal/external capabilities. Executive strategist and ex-CIO Jack Santos remarks “When outsourcing first started, everyone knew it was human arbitrage, and like all forms of arbitrage, the numbers eventually even out.”
When Robert Hayes says:
A company's competitive advantage is rooted in things it can do (e.g. design, make, distribute, or market) that its competitors cannot do as well, if at all. As the number of these core capabilities decreases, the company's competitive vulnerability to those that are able to master the same capabilities goes up.
The key discipline he highlights the need for is portfolio management. Not all IT functions provide us with the same value. We need to treat them accordingly.
I’m with Nick Carr on the impossibility for sustained IT-derived competitive advantage. With notable exceptions, running an IT shop does not give you that advantage. Sure, grinding down the cost per claim in an insurance company is supported by IT automation, but really it’s having a more optimal business process that does it. Data transmission arbitrage that once afforded the bulge-bracket investment banks an advantage in high-frequency trading is now available to all willing to pay the exchanges for co-location space. As Carr points out, an advantage based on infrastructure rapidly accrues to all.
Please do not misinterpret my message here. I’m no IT protectionist. And I disagree with Robert Hayes fundamental thesis that by outsourcing enterprises are “simply cashing out their intellectual assets”. With cloud computing, we have a chance to take an amazing step towards the industrialization of our profession, a transformation even, but only if we approach it with the right mindset.
1. Do we even know the cost of our existing infrastructure commons? For example, do you know your internal cost per processor hour, storage size, or network bandwidth? What is the lifespan of an infrastructure investment? Or what does maintenance per function cost for your existing software?
2. Do we have an architecture that is flexible enough to withstand the elimination of boundaries? Mike Rollings has just published (Burton Group clients-only) “Planning Considerations for Externalization in Cloud Computing” and discusses the implications of removing the location, perimeter and application container boundaries that have served us well in the pre-cloud era, but can no longer be assumed.
3. Do we have a mature enough solution delivery process that values transparency, and has the right blend of formal and informal governance?
4. Can you segment those core competencies from your commodity IT processes and infrastructure? Focusing internal talent and investment on those competencies will help maintain or regain competitiveness.
5. Are we correctly managing the expectations of what’s possible from current offerings? Chris Howard on keeping our feet on the ground:
“Externalizing IT can be a means of driving down cost, but there is no guarantee. The same risks organizations experienced with past outsourcing, for example, still exist. Organizations that view externalization as a panacea will be sorely disappointed. Cloud and SaaS options are simply not mature enough (yet) to handle the scale and complexity of enterprise computing. Immature IT shops will not gain the benefits of externalization, but are likely to incur greater costs and complexity.”
Let’s do some planning first this time, to avoid another off-shoring fiasco.