Moving Big Data to big value

The Institute of Advanced Studies held the first of its new series “Conversations on Big Data” on Monday 25th May. The Mega Data Cluster was a co-sponsor for the event, which booked out quickly and was attended by just over 60 people.

Coert du Plessis, Advanced Analytics Practice Lead Partner at Deloitte, gave an excellent presentation Moving Big Data to Big Value: crossing the last mile.

“Big data” is commonly sold as a valuable tool to derive insights. Yet often when people discover new facts and patterns the reponse is “that’s interesting”. And what happens? Nothing! So, as Coert challenged, the real question is how do we get to impact?

Creating insight is probably about a third of the work. The chasm, and it’s more of a chasm than a gap, is to get from insight to value. And it takes most of the effort.

Insight-and-ImpactOnce the analytics is done, there’s a need for human effort: understanding what needs to change and bringing people on side. Once the pathway to value is understood, we’re likely to get decisions leading to action. We should not underestimate the effort required to get from the insight to the point where a decision is ready to be made. Things in which an organisation already has a high investment may need to change, such as culture, business processes, and technology. Context and timing are also important.

Coert noted the tendency to lump “Big Data Analytics” under one umbrella. In reality, there are different types of analytics such as:

  •  Descriptive analytics: what happened?
  •  Diagnostic analytics: why did it happen?
  •  Predictive analytics: what will happen next?
  •  Prescriptive analytics: what should I do?

We also need to think about data in a time setting. Up to now there has been a focus on storing and managing data for analytics, where the data is historical and represents a snapshot of the past. There is also a predictive and planning part of data analytics which looks at the future, by building models and looking at the future. Such models can be validated against historical data, and produce forward-looking data which can guide decisions.  The newest trend is towards operational analytics, which uses a near real-time stream of data to direct operations. There is not sufficient time to pool the data or to generate analytic insight; decisions are made on the stream of input. This trend is being accelerated by equipment and systems which make decisions near the point of data generation.

Coert suggested one solution for organisations is to build insight centres, which create services to support front-line staff searching for insights, and link to decision-makers. An insight centre brings together all the data science, visualisation, and communication skills into one place. The aim is not for the central group to be the heroes that solve the organisation’s problems, but rather to edge execute: to have the people at the front line drive the system and empower them to do the analysis and derive insights, and to assist navigating pathways to impact.

Coert outlined the need for organisations to seek out and use their purple people. One could think of the technical geniuses as blue people, and commercially astute people as red people. The people who can link the two and understand both aspects are the purple people.

Coert suggested one of the hardest challenges is to change the current middle management mindset, from “gut feel” to insight driven. This requires several things:

  • Democratisation of access and action on data.
  • The culture of crunchy questions: making sure we ask the difficult ones that get to the point.
  • Making sure we ask, do you think or do you know?


Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s