The world of analytics technologies is proliferating – business intelligence, predictive analytics, prescriptive analytics, machine learning, data mining, statistical analysis, visual analytics. On top of this we have a similar proliferation of data technologies – in-memory columnar database, relational databases, Hadoop, Spark, noSQL, and so on. It is easy to lose sight of the one reason we might entertain this jungle of technologies and techniques. At the end of the day all we want to do is make better decisions. This whole array of technological wizardry has no other purpose than this, and yet we easily lose sight of this anchoring fact.
All decisions are made with imperfect information. We reduce the risk associated with a decision by gaining access to more relevant information, and the primary role of all business managers and leaders is to reduce risk. This applies just as much to the millions of operational decisions that might be made each day, through to the large strategic decisions, such as whether to enter a particular market or make an acquisition.
In an attempt to give some shape to the current orgy of technology, the concept of Decision Management has been developed. This says that the decision comes first, and the technology second. It is always going to be about decisions, no matter how elaborate and sophisticated the technology, so let’s use a methodology that focuses on decisions instead of getting lost in algorithms and charts. To this end various standards are developing, one of the most prominent being Decision Model and Notation (DMN). This provides the conceptual and diagrammatic tools to develop, maintain, document and implement decisions within the context of business processes. Suppliers such as FICO and IBM have been involved in its development, and slowly but surely the notion of Decision Management is gaining some ground. Decision Management Solutions, lead by James Taylor is probably the best known consulting firm in this field. Once an organization starts thinking about decisions, it automatically assumes a top down strategy. Acquiring technology and finding out what it can do, is a bottom up strategy that is often wasteful of resources – money, time and effort.
There is room for both top down and bottom up approaches to decision making. In a top down scenario we are implicitly saying that we know what we want to know, in order to make better decisions. The discipline of working in this way clears the fog of technology overload quite considerably. However, sometimes we want to know what we do not know, and this is where an exploratory approach to analytics can pay dividends. The two approaches are quite different. Top down is production oriented, whereas discovery is essentially R&D. Right now there is an overemphasis on the R&D role of analytics, and Decision Management is needed to correct the balance.
Looking forward, the tsunami that is rushing toward us is AI. The role of AI in business is to find solutions to problems (how to allocate resources for example) and make better decisions (optimize marketing spend say). If we do not acquire the skills associated with decision management, we can look forward to a business environment were intelligent agents run riot, and no one really understands whether the most pertinent decisions and problems are being addressed. It is quite common today for many organizations to have little or no documentation on their decision making processes – manual and automated. As a result businesses are deploying predictive models, optimized solutions, and now AI agents, that get lost in operational infrastructure. The net result is that managers lose knowledge of how the business is actually working. The solution is straightforward. Focus on decisions, and document the processes and technologies that are used to execute them. It’s less glamorous than the latest big data technology or whizz algorithm, but it is much more important.