The Next Wave in BI: Decision Analysis
We’ve come a long way in business intelligence, but there are still plenty of miles to travel. We’ve gone through three distinct eras: Data Warehousing, Business Intelligence, and Performance Management. I think the next era is Decision Analysis.
In the 1990s, we focused on building repositories of integrated, historical data (i.e., the era of Data Warehousing); in the late 1990s and early 2000s, we focused on tools for reporting and analyzing information in our data warehouses (i.e., the era of Business Intelligence); in the late 2000s, we focused on using information to improve performance by monitoring key performance indicators (i.e., the era of Performance Management.)
The next decade will focus on improving the way we make decisions. There is a lot to say here, and I haven’t completely formulated all my thoughts, but this era will take a long time to bear fruit because it involves understanding how the human mind processes information and how people interact in social groups to make decisions. To take BI to the next level, we need better insights into human behavior and perception. In other words, it’s time to recruit psychologists onto our BI teams.
In 2010, you will see the first fruits of the era of Decision Analysis. Specifically, you’ll see more robust collaborative capabilities embedded within BI tools and the first attempts to deliver formalized methods for evaluating the effectiveness of decisions made with those tools.
Most leading BI vendors are applying social media conventions to their toolsets to improve collaboration and decision making. For example, the online-based reporting service, Swivel, lets users rate and comment on charts published online by themselves or others. Following the lead of Facebook, LinkedIn, and other social media sites, some BI vendors will let you “follow” people whose analytical skills you admire and be alerted when they publish a new report. BI vendors will also beef up their guided analytics capabilities, enabling users to review the steps that a trusted analyst took to create a great report using a macro-based replay function. And expect every BI vendor to offer some form of annotation, threaded discussions, and tighter integration with email.
We’ll also see a host of new independent collaboration platforms that could provide the glue to link people, process, and documents in more seamless, transparent ways and improve decision making. For example, SAP is working on an online collaborative environment call 12Sprints that provides templates for specific types of collaboration activities. And Google recently debuted Google Wave, its latest collaborative environment that lets groups engage in seamless instantaneous conversations. Of course, many companies already use Skype, Google Docs, Google Groups, Facebook, and Web conferencing systems, such as GoToMeeting, to foster formal and impromptu collaboration. These incumbents will slowly become more formally integrated with BI tools and decision making processes.
Although many BI teams do a great job monitoring BI usage, most have done little to nothing to monitor and evaluate the effectiveness of what users do with information they give them. We need to begin tracking the decisions that users make with BI tools and measure the effectiveness of those decisions against business goals and plans. We need to start studying the decision making process and apply procedures to increase the probability that users will correctly interpret the data and take appropriate actions. We can only do this by applying the same types of feedback loops we’ve applied to our BI systems themselves.
A terrorist’s attempt to blow up Northwest Airlines flight 253 last month revealed some fatal flaws in our country’s intelligence gathering activities, including a lack of coordination and information sharing among agencies. But another intransigent problem, it turns out, is the faulty assumptions that analysts apply to evidence and the lack of organizational controls for testing and challenging those assumptions.
A recent article in the Boston Globe called, “Think Different, CIA” provides some instructive lessons for companies using BI tools to make decisions. The article describes a phenomenon that psychologists call “premature cognitive closure” to explain how humans in general, and intelligence analysts in particular, can get trapped by false assumptions, which can lead to massive intelligence failures. It turns out that humans over the course of eons have become great at filtering lots of data quickly to make sense of a situation. Unfortunately, those filters often blind us to additional evidence--or its absence--that would disprove our initial judgment or “theory.” In other words, humans rush to judgment and are blinded by biases. Of course, we all know this, but rarely do organizations implement policies and procedures to safeguard against such behaviors and prevent people from making poor decisions.
The Next Wave
To take BI to the next level, we need to provide a collaborative environment to improve decision making and evaluate the effectiveness of decisions on a continuous basis. We need to establish processes and procedures to ensure people and teams properly interpret the data, identify and challenge each other’s assumptions, and keep an open mind about the drivers of business activity. By applying collaboration and governance to decision making, we can help our companies get even more value from their BI dollars. This really is the next wave of BI.
Posted by Wayne Eckerson on January 19, 2010