Informatica Perspectives

Just about everyone has heard about Nicholas Carr’s Harvard Business Review article titled IT Doesn’t Matter. I’ve heard many reactions to the paper presented at conferences and in the blog-o-sphere, so I figured it was time to go to the source and purchased IT Doesn't Matter (HBR OnPoint Enhanced Edition) from Amazon for $6.50 as a downloadable pdf. The paper is excellent in that it presents not only the original paper first written in 2003, but also features no fewer that 14 reactionary letters plus a final response from Carr to the lettersInterestingly, not a single one of the 14 letters agreed with Carr’s fundamental premise (although most of them were polite enough to acknowledge some value in the debate). I don’t know whether it was the editor’s choice to select only negative letters out of the hundreds that I’m sure were submitted, but I suspect these 14 are representative of the masses. So by way of this commentary let me add the 15th letter.

The reason there are so many negative responses to the article is for one simple reason – Nicholas Carr is wrong. I imagine Carr was never an IT practitioner, because if he was he would have known that it’s not the “technology” that is scarce in the modern IT shop, it is the ability to leverage information effectively in a complex enterprise environment. The technology components are indeed commodities as he suggests, but a computer may as well be a boat anchor if it wasn’t for the information it contained and how it is connected to other systems.

Several of the letters from readers reinforce this point. Vijay Gurbaxani states “The situation is subtler than he [Carr] would like us to believe. The scarce resource never was technology, as Carr assumes; it was always the set of managerial capabilities needed to create value with that technology.”

Marianne Broadbent, Mark McDonald and Richard Hunter in their letter write “Sustainable advantage comes from consistently delivering greater value to customers. This comes from the “information” in information technology – that is, it comes from better understanding the customer, applying that understanding to your products, services, and processes, and integrating these to deliver on an improved value proposition.”

I am particularly impressed with Paul Stassmann’s detailed response. Stassmann also has been somewhat critical of IT investments over the years as evidenced by his books on the IT Productivity Paradox – so I was surprised, and pleased, to see his objective rebuttal of Carr’s assertions. The difference between them is that Stassmann is driven by data and facts rather than Carr who makes his arguments based on analogies and metaphors (which are rather weak as it turns out).

It is interesting to note that in Carr’s response letter to the criticism that he seems to be back-peddling. While he still defends, and reiterates, his main points, he seems in fact to be saying that “IT does matter”. It’s about time that Carr, and the editors, admit that the title of the paper is overly provocative and misleading. If the only purpose was to generate debate – then after five years that purpose has been served. Mr. Carr could go a long way to correcting the negative effects of his paper by writing a new one – maybe with a title like “IT matters and here’s why”. What do you think?

You may need to reorganize your company to take advantage of modern IT architectures such as SOA or MDM. Some people would say this is heresy – that this is a case of the tail wagging the dog and that the IT strategy should echo the business strategy and operating model. True enough – but IT is also “part” of the business and also has an obligation to play a role in leading the enterprise into the future.The fact is that much of the current IT “mess” is because it has directly mimicked the business. Any organization that builds systems will naturally build them in a way that supports the organizational structure and the lines of communication that are inherent in the functional silos. The legacy problem that most organizations face (high cost of maintenance, lack of agility to change, poor data quality, unpredictable operations) is a direct result of building systems piecemeal in support of silos.

There is no shortage of technology silver bullets to solve the problem. The current crop of solutions includes SOA, MDM and ERP to name just a few.

There may another reason why the Gartner hype cycle exists – you know the one; peak of inflated expectations, trough of disillusionment, plateau of productivity. The reality is that most of these technologies are not technologies at all – they are enterprise strategies. And strategies require a C-level commitment that literally starts with top the organization – the CEO.

Lets take a look at an SOA example. A common business service for a bank is a “Credit Decision” service. Credit decisions should be done consistently across the bank in a way that reflects the bank policies, market strategy, risk management controls and regularity constraints in all the different legal jurisdictions that it operates in. But rather than implement one credit decisioning engine, the Credit Card business implements one, as does the Mortgage Business, the Deposit Business, the Line of Credit business, the Personal Loan business, etc. And if that isn’t enough, since the internet bank is run by a different executive from the retail branch bank – we also have separate decisioning engines for different customer channels. Each of the business groups as a result needs to hire staff to constantly maintain and update the business rules in each of the redundant systems to comply with corporate directives.

If you go and ask each of the LOB managers if they need their own decisioning engine, along with all the overhead that goes along with maintaining it, they will likely say no. It doesn’t make sense. But if you try to implement an SOA strategy in a large bank with 10 different decisioning engines, you will have an uphill battle. The first challenge is one of timing. While all 10 groups may in fact be totally on-board with the concept, they all have different priorities in any given year. You may be waiting forever for all the stars to align such that a critical mass of them are prepared to all focus on implementing a generic solution at the same time.

It’s clear from this example that reorganizing the business to create a “credit decisioning function” that serves the enterprise would dramatically simplify the IT solution. The typical pattern however is to avoid the hard work of addressing the organizational, people and political issues and instead try to solve the problem “bottom up” by incrementally and thereby iterate towards an enterprise solution. The bottom-up approach can in fact work – but don’t make the mistake to think that it’s necessarily easier more successful. The reason so many organizations are in the SOA Trough of Disillusionment is exactly because they are building it bottom-up and have not addressed the organizational issues.

Let’s take another example – Master Data Management. MDM promises to deliver quality information and one version of the truth for business data. But in order to achieve this, the organization needs an information owner (or steward) who is accountable for data quality, consistency, currency, security, etc. For most companies that are organized around LOBs or products – who owns customer information? In the absence of a business owner, a common pattern is for IT to step and presume a stewardship role. Unfortunately (as is evidenced by a long list of less than ideal CRM implementations) the IT group often does so without an official sanction from the business executives so the results should not be a big surprise.

This is not to say that any of the technologies that are inherent in SOA, MDM or ERP are bad or flawed. In fact, they are the result of the smartest people in the industry creating the technologies and the result of market forces testing the all the innovate ideas and having the best ones bubble to the surface.

At the risk of repeating myself, the point is that successful implementation is much more than a technology. You can't “buy” SOA or MDM – these are strategies that demand organizational top-down commitment. While a bottom-up strategy can make meaningful progress and deliver significant results, IT organizations should not be afraid to tackle the “big” issues of organizational design and politics. IT is part of the business. And as IT has become a larger and larger player in most organizations due to the natural shift to an information economy, it would be wrong to think of this as the tail wagging the dog.

It’s been almost 5 years since Nicolas Carr started a lively (and sometimes hostile) debate with his Harvard Business Review article IT Doesn’t Matter. And now, just this past fall, Cynthia Rettig added fuel to the fire with her MIT Sloan Management Review paper The Trouble with Enterprise Software. There are indeed serious concerns with the perception of IT and many business leaders are fed up. You don’t need to look very hard to find evidence of trouble. The Standish Group has been tracking IT project success rates for years and the findings are dismal. The 2004 Standish CHAOS report found that only 29% of all IT projects succeeded while the remainder were either challenged or failed. Cynthia Rettig in her paper writes:

“As a percentage of total investment, IT rose from 2.6% to 3.5% between 1970 and 1980. By 1990 IT consumed 9%, and by 1999 a whopping 22% of total investment went to IT. Growth in IT spending has fallen off, but it is nonetheless surprising to hear that today’s IT departments spend 70% to 80% of their budgets just trying to keep existing systems running.”

So while IT is gobbling up larger and larger amounts of investment spending, it seems to be doing so in a way that is expensive to maintain and difficult to change.

Consider this, why is it that the Gartner Hype Cycle is such a well-accepted pattern; that technologies go through a cycle that includes the “peak of inflated expectations”, “the trough of disillusionment” and finally the “plateau of productivity”. It seems that the industry has a habit of over-promising and under-delivering so it’s hard not to blame business users for being cynical about any new technology.

By the way, Gartner missed a key phase in their technology hype cycle; the “pit of obscurity”. Some technologies never reach the plateau of productivity and instead fade away until they are just a distant memory of the IT old-timers. Whatever happened to AI (Artificial Intelligence) or 4GL (fourth generation language)? They are now in the pit along with a host of other silver bullets.

I respect both Carr and Rettig for helping to shine a light on the “Moose on the table” (the big ugly thing in the room that no one wants to talk about). If the industry is going to solve the problem, we need to make it visible and tackle it head-on. By the “industry”, I don’t just mean technology suppliers; we need corporate IT groups, academic organizations, standards bodies and government agencies all working together.

Both Carr and Rettig stop short of offering any solutions to the trouble. Carr takes the “necessary evil” point of view and simply suggests spending less on IT and outsourcing as much as possible. This doesn’t really solve the root cause issues and is simply avoiding the problem. Rettig suggests closer communication between business and IT and that executives should educate themselves more about technology. Good advice, but it’s a bit like saying if you want your food cooked right, you need to go to culinary school so that you can cook it yourself.

So what is the solution? At the risk of sounding pretentious, I’d like to suggest that there is a silver bullet solution. The answer is not more technology – we have plenty of technology and quite frankly, most of the core technical challenges were addressed by the 1980’s. The wave after wave of new technologies over the past 20 years are mostly incremental improvements over long-standing practices. Don’t get me wrong, there have been amazing developments in the past 20 years such as the low cost and tremendous capacity of the modern PC and the World Wide Web which is arguably the largest application in the world.

What I am talking about are the fundamental disciplines of managing change in an IT environment. Specifically, Information Architecture, Program Management, Systems Integration and Configuration Management. The silver bullet solution to today’s IT complexity is to simply perform these four practice areas in a disciplined fashion. And the interesting thing is you don’t even need to buy the latest IT Management Best Seller to learn the techniques – if you follow the advice of The Mythical Man-Month, first written by Fred Brooks in 1975, you will have 80% of the Silver Bullet solution. In his book, Brooks talks about his insights for how to effectively build large-scale integrated systems that require many teams consisting of hundreds, or even thousands, of staff working in a cohesive manner. His advice is as applicable today for a large-scale enterprise IT department consisting of hundreds or thousands of staff working to build an integrated enterprise system that meets the business needs.

Note that I didn’t list Software Engineering as one of the 4 pillars of the real Silver Bullet. The reason is that the industry was doing a good job of software engineering 30 years ago and is still doing a good job today. The development languages have changed and the tools have evolved, but fundamentally, the discipline of writing software is not a root cause of the trouble that Cynthia Rettig speaks of, or the frustration that Nicholas Carr writes about. The root cause is lack of discipline (read leadership) in the practice of architecture, integration, program management, and change management. If an organization was to focus on these four practice areas, put their strongest leaders in charge of them, and invest in training and tools for the staff, then the hype cycle will disappear.