Enhance corporate IQ by incorporating big data into predictive analytics

The world is currently undergoing an information explosion, with huge volumes and varieties of data being generated at a rapid velocity.

December 4, 2012

By Domenico Griessel, Consultant at Cortell Corporate Performance Management

The world is currently undergoing an information explosion, with huge volumes and varieties of data being generated at a rapid velocity. This ‘big data’ has changed the information game for businesses, providing more information for analysis but also in many instances leading to ‘information overload’ and providing a host of challenges. Along with the rapid and increasing volumes of data comes a need to make faster, more agile and smarter business decisions, and key to achieving this is obtaining a complete, holistic view of information. As such, it is critical to incorporate big data into Business Intelligence (BI) and predictive analytics tools, increasing business IQ by ensuring that decisions can be made based on the entire picture and not just a segment of the truth.

Big data is currently a huge buzzword in the business space and is often misunderstood by organisations. Understanding the concept of big data is, however, critical to harness the potential benefits of the information to derive business benefit. Big data does not simply refer to the volume of data, although this is one component, but also to the variety of data and the velocity at which it is generated. To put this into perspective, every day an estimated 2.5 quintillion bytes of data is generated and 80% of the data that exists today was created in the last two years. Of the new data that is generated, 80% of it is unstructured content, including emails, images, audio, video, click streams and more.

While big data poses a challenge, it also provides many opportunities. Not only has the velocity of data generation increased, so too has the velocity of required decision-making. The rapidly changing business climate, along with business imperatives to accelerate innovation, improve business optimisation and agility and reduce time to market, means that organisations need to move ahead of the curve by predicting issues and events and addressing them in real time. This requires predictive analytics, which can be greatly enhanced and new insights gained when structured and unstructured data are analysed together.

The challenge with analysing structured and unstructured data in combination is that traditional BI tools are simply not up to the job. Even if tools are available to aggregate data from a variety of unstructured sources, including email and social media, the key to gaining value from big data is the speed at which this data is able to be collected, sorted and analysed. For example, a company may be analysing its big data, but when the volumes of this information run into the petabytes and queries take weeks to run, this information is practically useless. Any insights gained weeks after the fact will fail to deliver the agility required in today’s business environment. Analysis needs to be completed in real-time or near real-time otherwise the accuracy of insights gained will be compromised.

Getting value out of big data is not simply about analysing all available information, but about increasing the speed at which this data is accessed and analysed. The time period for analysis needs to be optimised in order to gain the greatest value and insight from the data. The key to big data is to analyse a greater volume of information in a shorter space of time. This will deliver more accurate insight, greater numbers of insights into a variety of different business aspects and a more complete picture of the business, which then enables more accurate predictive analytical capabilities.

A 360 degree view of the problem, including interaction data (email, chat and call centre data), attitudinal data (opinions, preferences, needs and desires, descriptive data like characteristics, self-declared information and geo-demographics) and behavioural data (orders, transactions, payment history and usage history), is an important source of competitive differentiation. This requires the incorporation of unstructured data sources to achieve, and in turn, lead to a more agile decision-making ability.

Business analytics is key to gaining a competitive edge, from the insights gained by Business Intelligence, in other words what is happening, through to the realms of predictive analytics, or why this is happening and what will happen next. Only then is it possible for businesses to understand what should be done about these insights. Predictive analytics uses current and historical data to discover hidden patterns and relationships for better and more informed decision-making in real time. However, running predictive analytics without incorporating big data and data from unstructured sources leads to an incomplete picture of future events, which in turn compromises the accuracy of decision-making.

While big data is not currently an urgent consideration in the South African market, the reality is that data continues to grow, in greater volumes and more variety, at a higher velocity. Ignoring this data will affect the accuracy of predictive modelling tools and solutions, which in turn will compromise business insights gained from analytics tools. When it comes to dealing with the pace of today’s business world, it is critical to be able to make predictions using all data streams, faster and more accurately than ever before.