I recently watched a Charlie Rose interview with Léo Apotheker, co-CEO and a member of the Executive Board of SAP AG and Andrew Mcafee of the Technology and Operations Management Unit at Harvard Business School. The discussion was how technology can be a catalyst to companies staying competitive and emerging through a downturned economy. The reality is that even as they discussed technology in terms of process and enterprise solutions, the underlying value of technology was information – business intelligence. So if information is so important, why is our business intelligence practice held back?
Charlie Rose: A conversation with Léo Apotheker and Andrew Mcafee
There were four key technology areas discussed that get at the heart of why it is so difficult to forge a business intelligence practice.
1. Business process optimization in silos – leveraging disparate applications creating siloed databases
2. Cloud computing – applications utilized and managed outside the corporate environment
3. Open source – unlicensed technology created outside of traditional process management and plugged in either as an entire solution or component
4. Mash-ups – combining disparate applications to create added value
What is interesting to me across these four drivers is that as business intelligence professionals we are hampered by them, but we can also learn from them.
Pitfall 1: Data for Insight Was an Afterthought
Over the past 20-30 years, we have focused our efforts on implementing technology to improve processes and reduce overhead. Data was only as important as getting from one process step to the next. Data was collected in the process according to a department’s business needs and it was leveraged to facilitate that department’s processes. Applications came with pre-packaged databases to house information. Developers installed, customized the user interface, verified the database still captured the data, then released. Add in the adoption of hosted solutions for sales force automation or partner management systems and it is even more complex. Data and its potential was an afterthought.
Pitfall 2: Building a Practice is a Road Map
While business processes optimization has been the holy grail, mapping a process, finding the pain point and bottle necks, and applying solutions in those areas has created enormous efficiencies that have translated into huge profits for our companies. GE pioneered Six Sigma, you have TQM, ISO, SDLC, and every other BPO acronym. These operational management processes are there to ensure projects are run on-time, on-budget, and this should drive success. Forging a business intelligence practice has also come under this umbrella with people looking for a 9-step plan for success.
Pitfall 3: Database Mash-ups Turn Out Mashed Potatoes
Our initial attempts at business intelligence were to link disparate databases through a simplified ID system. The Enterprise Data Warehouse was really just a backup repository for all our application data minus the business processes that encoded the data for use and insight. Analysts became database modelers creating linkage tables on their desktops to pull together cross departmental information. This caused inconsistent analysis, inaccurate analysis, and opened the company up to risk as data sat on relatively unsecured laptops. As we now begin more mature business intelligence practices, we have analytic solutions that manipulate warehouse data and provide a window. However, the underlying warehouse is still a mash of un-modeled data and has only achieved control over information rather than providing insight.
Even as we are challenged by these pitfalls, as said above, I think we can learn from the four drivers. My biggest take away is possibly the notion to break from past thinking.
“Open Source” Your Business Intelligence Practice
Why do I like the idea of a free form approach to business intelligence vs. a step by step? The concept of open source is that anyone can build applications and components and bring them to the masses. There is no traditional software development life cycle. Step by step processes are thrown out in favor of free form innovation and creativity to meet a need. Predicate the project on information needs rather than a roadmap. Each need drives an area for development. Only draw from your traditional development and project processes to prioritize needs and focus on smaller tasks.
The other reason why I am in favor of this approach is that everyone is an expert and everyone can contribute. This is not to say you don’t have a leader of the effort, but leadership may shift or be shared depending on who has the knowledge and expertise rather than seniority or title. For instance, analysts have, over time, developed a deep understanding of the data in your warehouse. I’ve seen cases where the analysts know it better than IT. When it comes time to consider how to best model the data for analysis, this aspect of the project could be lead by the analyst or jointly with a modeler. On the other hand, how can you leverage an open environment for contribution? There may be core players on the project, but allowing ideas come from all directions inside and outside the group may improve your deliverables.
Dovetails over Mashed Potatoes
Mash-ups are really what business intelligence is all about. Andrew Mcafee uses the example of connecting rental listings and Google maps for a better perspective of the rental property. The term is really coming from the mashing up of applications and the difficulty and effort it takes to do so. This is a great example of how a business intelligence practice can shift the focus of value to the business away from process toward data driving and fulfilling business need.
The difficulty for analysts and data warehouse professionals has been putting data into terms that the business understands. If you talked about process, this was easy to relate to. As business intelligence leaders, we need to take charge of building our projects from the data level up rather than having applications drive our data. In the example of rental properties and Google maps, conversations about how the information will be used and how to best connect disparate databases is the tipping point to combining the applications, not the other way around. Dovetail the data for clean seamless insight and applications will follow more naturally.
To-Do
My big take-aways from this interview are:
Tip 1: Free form approach
Tip 2: Everyone is an expert
Tip 3: Seamless data integration optimizes applications and processes
A business intelligence practice has been held back by a lack of forethought on putting insight first, data management best practices, and rigid processes that don’t meet the needs for fresh thinking and approaches. I challenge you to try a free form approach and allowing for open contribution to build a better data foundation. You’ll then achieve not only better insight to drive your business but will contribute to streamlining your business process.
