Practical Big Data series, Part 2
In my first blog about Practical Big Data here, we explored some of the common pitfalls of big data projects. We identified some common mistakes such as:
- Big data initiative does not have a clear business ownership
- The initiative does not work with a business case to drive value
- No upfront plan of how to put the data into practical use
- Focus on less mature parts of big data too early
Today we will take the lessons learned and turn our focus on how to make big data practical from the business perspective.
Definition of Practical Big Data
Let’s start by defining Practical Big Data: in simple terms it is about connecting enterprise big data with real-time business.
Real-time business means interaction with your customers without any latency in the key customer engaging processes. When using big data in customer interaction, the window of opportunity is often very narrow. Usually, you do not have more than 2 seconds to deliver a new product proposal, marketing message, price indication, or other key piece of information to the customer when she/he is still engaged – otherwise the opportunity may be lost forever. Real-time business may also refer to other, non-customer facing business process (more on this further below).
Another key word is enterprise big data. You need to work with big data tools that are enterprise-ready and can be adopted widely in your organization. This means ease-of-use, stability, and security. Some big data tools fail to meet these criteria as well as the applications that they drive.
The acid test of Practical Big Data
From a project standpoint, for Practical Big Data to succeed, the following criteria need to be met:
Strong business ownership. Ideally, top management participates in the in the Practical Big Data project from scratch to finish. The project starts with the formulation of the use case that drives value. C-level ownership would be assigned to the project and an iterative approach would be taken to create an end product that would serve the business best. Finally, company-wide adaptation would be achieved by a careful roll-out of the end product into daily business use with support from the management at all levels.
Ties with business processes. At the outset, the whole innovation process is turned towards a specific end product. Rather than focusing on the data itself, the goal is to design a data-driven end product: a highly accelerated business process, a digital engagement tool, or a digital product/service. The typical value is new or improved sales, better margin, increased customer satisfaction and loyalty, or cost avoidance.
Provides real-time engagement. Engagement would most typically mean customer interaction, but could also mean other key stakeholders such as your service organization, decision makers, or the dealer channel. In business processes time is money, and eliminating latency may enable totally new business models that would not be possible without real-time computing.
Uses enterprise-ready big data tools. In-memory and predictive analytics are more mature parts of big data. Real-time business usually requires both, and building these capabilities first will more likely lead to faster and more certain value with big data.
Three levels of Practical Big Data
In practice, the end result of a Practical Big Data project is a significantly accelerated business process, a digital tool that frontline workers use in their daily work, or a fully digital product/service. The business case, tool set, and skills required for each of these scenarios is slightly different.
Therefore Practical Big Data can be described to have three levels:
- Power up your business processes with big data and real-time analytics;
- Engage your customers and other key stakeholders in real-time;
- Transform your industry with new data-driven products, services, and business models.
(click to enlarge)
Does this all still sound too academic for you? Look ahead to my next blog in this series for some practical examples on Practical Big Data!
