Bill Inmon recently shared a conversation with a consultant who has been involved in many big data proofs of concept. The consultant reported that 80% to 90% of the big data projects had stalled out after the proof of concept. A key reason cited was that no clear cut business value had been identified at the start of the Big Data project. In other words, there was no business strategy for big data. As a business intelligence and data warehousing consultant to Fortune 500 companies for over a decade, I’ve seen several cycles where the “next big thing” prompts unrealistic enthusiasm about “silver bullet” data strategies that will purportedly provide a competitive advantage. When cooler heads eventually prevail, companies do the necessary strategy work to align data and business processes in a way that creates business value. Here are three key tasks that contribute to a solid business strategy for big data:
1. Reach a common understanding of how the term “big data” is being used. There is no accepted definition of what constitutes “big data.” Once approach talks about data “volume, variety, and velocity” – the implication being that big data can be differentiated from regular data by those three factors. Another approach talks about structured data from standard enterprise IT applications like ERP versus unstructured data from mobile devices, web logs, and other newer sources of data. BI strategies for leveraging structured data have been in place for over a decade now. For example, see “The Profit Impact of Business Intelligence.” Strategies for leveraging unstructured data are emerging – and in many cases it looks like big data is a “solution” looking for a problem to solve.
2. Develop a big data value proposition up front. Let’s assume that by “big data” we mean “unstructured data.” The key question that needs to be answered amounts to this: how can our company convert the massive amounts of social media “content” into something that either increases our revenue, reduces our costs, or both. While the technical methods for making the conversion of social media content into useful content are somewhat different, the method for linking big data to business value are the same as for any business intelligence (BI) strategy. Our BI Pathway Method is one such approach, and it can be used to make the business case for big data in the same way it has been successfully used on numerous BI strategy engagements with large, complex companies in a range of industries.
3. Assess organizational readiness for business process change driven by big data. Once a link between big data and one or more business processes has been established, companies need to assess the breadth and depth of business process change that will need to happen in order to leverage big data and achieve an ROI.
If your company can accomplish these three tasks, the groundwork will have been laid for a pragmatic business strategy for big data. Traditional business intelligence and data warehousing strategies are well-understood and relatively lower-risk. Big data strategy is essentially research and development strategy, where a hypothesis about how big data can improve business results has to be established and then tested through a proof of concept. Without a business-driven hypothesis, investments in big data become even more risky that there already are