I am not looking for sympathy, but being an enterprise CIO is a rotten job. The demands are constantly increasing while the budgets shrink. If I am a “perfect” CIO – which means that my systems are never down, I deliver every project on time, on budget and on target, all while maintaining cost leadership – the best I can ever hope to achieve is to meet expectations. As soon as my systems hiccup or I miss a project goal or the enterprise perceives that I am not wringing every possible dollar out of my operation, I have failed to meet expectations and my credibility falls.
Given the rotten nature of the job, you might wonder why I have such a passion for my role. The truth is that being a CIO is a great way to deliver an incredible amount of value. And business intelligence (BI) and analytics tools are often my best friends in delivering this value.
Let me give you an example. Several years ago, I was hired to transform the IT of a struggling specialty retailer. This retailer had suffered through seven years of losses and was being battered by new competition from big-box retailers. Facing these challenges, the executive team took the too-common approach of blaming its IT for its business and market failings. The team then decided to refresh the entire technology stack and brought me in to effect this IT turnaround. Once on board, I quickly realized that the biggest failing of the legacy IT system was that it provided no decision support. Without good decision support, the leaders of the organization were blind to market needs. Every day, they guessed about what products to sell and how to sell them. Now, some of their guesses were good. But there were just not enough good guesses to keep them afloat.
I took a decision-support approach to the selection and implementation of the new IT systems. This meant that we did not talk about transaction processing, hardware options or software functionality. Rather, we started by asking executives and employees, “What decisions would you like to make?” With those answers, we drilled down one level and asked, “What information do you need in order to make these decisions?” We then selected and developed the transactional systems that would generate the information internally and selected BI tools to surface this information for nontechnical decision makers.
To answer “What decisions would you like to make?” we surveyed our customers on what drove their purchase decisions, what they liked and did not like about doing business with us, and their specific needs. From there, we identified four specific customer segments – each with its own decision driver. For example, one segment was interested in the latest products. Another was interested in purchasing what was on sale. Knowing these four decision drivers, we revised our approach to the market to address at least one, but preferably more, of the drivers. All advertising hit these segment needs. We redesigned physical and Web stores so that new and on-sale products were the first things people saw when they entered. Discovering that one of our segments relied on others for product recommendations, we launched our own version of a social network to connect these people with trusted product advisers.
Within a few months we started to see the results. Revenue climbed as did profits. It turned out that knowing the segment drivers meant we could replace expensive and somewhat ineffective broad-based advertising with less expensive, more targeted direct marketing.
And our decision-support approach to the systems improved operations and customer service. One of the decisions we wanted to make was where a product was in its life cycle. We established several lifecycle phases (launch, growth, maturity, retirement) and success indicators for each phase. If, in the launch phase, a product did not achieve its success indicator, we pursued options for wringing out more value or agreed on an early retirement. If, on the other hand, a product outperformed its launch indicator, we put more resources behind the product to accelerate its growth phase.
Taking this decision-support approach also made my IT transformation job much easier. With the company focused on what it needed to make better decisions, there was less emphasis on the functionality of the new applications. Rather than fighting battles about how to configure the inventory management module of the new enterprise resource planning (ERP) system, we could satisfy people by confirming that a standard, vanilla configuration would allow them to make all of the decisions about inventory they ever wanted to make.
Implementing Decision Support
To ease the transition, I suggest the following:
Iterative methods. Enlist the group that is the most willing to try (and then support) new things or the group that is in the most pain. Work closely to select a subset of the decisions as a test case. Once improvement is confirmed, move on to the next subset or pilot group.
Think about the perfect future. Use the technology and information available now, while encouraging teams to think about the future so that new sources of data (such as from social networking) can be integrated.
Reports are not business intelligence. Not all reports add decision-making value. The reports that deserve our attention are those that answer critical questions. The rest can be eliminated.
Ease of use is critical. I want to get IT out of the report-writing and analytics business and leave that function to the decision makers. So I select BI tools that can be understood and used by someone without a technology background.
In the end, superior decision making might just be the ultimate competitive advantage. And if, in my rotten CIO role, I can deliver the tools and thought process that make better decision making a reality, I become an irreplaceable business leader.
Gain access to industry-defining research for 90 days for free. If you would like to join the IIA, use the code SAS90 when you register to receive a 90-day trial individual membership.
This article originally appeared on the International Institute for Analytics website. Copyright 2010 IIA. All rights reserved. Condensed and reprinted by permission.