Big data offers an insight into behaviour – whether that is the behaviour of individuals, a process, group, machine or an organisation. Research into understanding its impact can be a vital tool for businesses to gain insight on its customers. As such, the results can have a large effect on company decision-making.

But there is a fine line to tread between detailed customer insight and intrusion. Big data can be a powerful tool, but it can’t yet predict large-scale trends.

Sander Klous, managing director of Big Data Analytics at KPMG Netherlands and professor at the University of Amsterdam, talks about how this data should be used – and the pitfalls of overreliance on predictive analytics. Analysing big data can be key, but you must ask the right questions. 

In the age of shared platforms such as Uber and Airbnb, how can a business best adapt? Companies need to consider:

• Pooling insights and analysing data patterns can be of mutual benefit for a variety of businesses. The amount of data available for online retailers puts them at the forefront.
• Despite the insight this data offers, there is an ‘invisible barbed wire’ around customers – businesses must walk the fine line between understanding the customer and getting too close.

• There are limits as to how accurately this data can predict the future. Rather than relying on data to work out which technology is liable to succeed, businesses should work to be more flexible and able to adapt to changes.
• Big data is more successful within confined spaces. For example, technology to determine the tipping point for panic in a stadium (using CCTV and Wi-Fi) should be possible in five years.
• Klous’s team has developed the KPMG Affinity Index to work out which combination of stores in a mall attracts the most visitors.

• One obstacle to reaching the next level of predictive analytics is the reluctance of businesses to share data. Why would online retailers want to share information with Amazon?
• Technology is moving towards a platform business model; and those platform models work better with combined insights. 
• Improvements in efficiency are driven partly by big data analysis. The limitations of big data must be understood and smarter predictive strategies implemented.

• On your reading list: We are Big Data: the future of the information society by Sander Klous and Nart Wielaard, Atlantis Press.
  

• On your board agenda: How do we limit the scope of our decisions to confined systems in order to apply big data predictive analytics more usefully?
  

• Anticipate tomorrow…: When artificial intelligence like cognitive systems such as IBM Watson is a commodity service available to all businesses, how will you secure competitive advantage?
  

• …deliver today: How do we get our existing data sets clean and useful for analytics; and ensure new systems are capturing all the data, even from unstructured sources?