Small Data Analytics
Nowadays, big data is a ubiquitous topic in the business media; everywhere you look, the financial newspapers and magazines are filled with articles telling you how big data can be transformed into valuable insights. Over recent months, big data analytics has started to creep into the supply chain media too. The question is, what is the current status of big data in the supply chain?
A recent survey of decision-makers around the world by consultancy firm Accenture reveals that the actual use of big data analytics in the supply chain is still relatively limited. Although 97 percent of the respondents say they understand how their supply chains can benefit from big data analytics, only 17 percent of them indicate that they have already used it in one or more supply chains.
I personally suspect that those two percentages are, in reality, much lower than Accenture’s study suggests, because the problem is that no clear definition of big date is given in the survey (and most media articles). What does big data actually entail? Rules of logic imply that if big data exists then there must also be such a thing as ‘small data’. Small data comprises all the structured internal data in administrative systems such as enterprise resource planning (ERP) and customer relationship management (CRM). So big data must cover unstructured internal data, such as emails, plus external data (whether structured or not).
My own research has revealed that most companies at best apply predictive analytics using internal data, mainly forecasting based on historical data. Even sensor-based monitoring in manufacturing facilities – a form of the industrial Internet of Things – is classed as small data and predictive analytics. In Europe, I have come across only a very small number of practical applications of what is known as ‘prescriptive analytics’ based on external data. Just a handful of consumer goods multinationals, such as Procter & Gamble and Unilever, are actually using external point-of-sale (POS) data for forecasting (so-called ‘demand sensing’). Although highly promising, the analysis of social media for supply chain planning purposes is still very sporadic, such as in the case of a manufacturer of headphones. Most companies first have to get to grips with small data analytics before being ready for the more serious stuff.
Martijn Lofvers, CEO & Chief Trendwatcher