In my last post, I explored the concept of being a “data driven” business, and I looked at three different models:
- Completely data-driven, which includes organizations that compete based on transforming information into monetizable assets (examples: Airbnb, eBay, Facebook).
- Data-infused, which are businesses that manage and sell products/services but use information to drive marketing, sales and business process optimization (examples; Amazon, Netflix, Capital One).
- Data-informed, which are more conventional businesses that are trying to adapt new information and data management technologies to fit their existing business models and thereby improve their overall competitiveness (examples: John Deere, General Electric).
Businesses that are completely data-driven are likely to have engineered their operations around their data management capabilities. They probably operate with a high degree of maturity and agility when they consider modifying their business processes and their applications to support new functionalities, new products or new services. Data-infused companies also exhibit some degree of sophistication when blending the use of information with business processes.
The challenge lies with those companies that are data-informed, and the remaining organizations that aspire to be data-driven. These organizations may rely on computing environments that were originally designed to support acute operational or transaction processing needs. In turn, their systems are not engineered to accommodate data extraction, consumption and analysis as a foundational part of driving revenue, managing expenses or improving the customer experience.
Being data-aware: The first step
Businesses that want to be data-driven first have to be data-aware – all the staff members have to understand that the company values data and that it is everyone’s responsibility to ensure its utility. At a macro level, though, there are some commonalities among those organizations that are, or are becoming, data-driven.
- They understand elasticity and scalability. While it's easy to talk about the need for “big data" management, the key is not the technical platform. It is actually the internalization of the need for systems that can accommodate peak load, data-intensive computation to deliver timely results. Amazon’s website can’t wait minutes, or even seconds, for its recommendation engine to suggest other products to promote to visitors. Its systems need to rapidly deliver analytical results on demand in real time – and that means having systems to accommodate that demand.
- They can accommodate data streaming. With myriad potential sources of incoming data, the data-driven organization has designed its environments to process streaming data in real, or near-real time. More data is coming at us faster and faster, and the intelligent organization can ingest many data streams, rapidly filter what is needed, and forward that data to ongoing operational analytics applications as well as downstream processes. Note that, of course, handling an indeterminate number of high-speed data streams relies on elasticity and scalability.
- They take advantage of modern data integration capabilities. Ingesting lots of data is meaningless unless the data can be integrated within the enterprise information environment. Data integration today means more than just extracting, transforming and loading into a data warehouse. With data fueling multiple simultaneous operational and analytical systems, an organization that does not have a handle on data interoperability – especially between on-premises and cloud systems (as well as SaaS/PaaS environments) – is going to be dead in the water.
- They can rely on their data quality. Finally, no self-respecting organization could ever claim to be data-driven without having a mature and operational data quality program. Being data-driven means having an unquestionable assumption of the reliability of the corporate data asset.
As the age of big data continues to dawn, there is no better time than now to evaluate the steps necessary to become data-driven. But the foundation steps are clear: Establish data quality, develop robust data integration methods and techniques, and design and implement a scalable platform that can easily handle growing volumes of streaming data. Once these tasks are completed, your organization can begin the process of understanding the potential for integrated analytics to drive business value.Download – Data Management in Action: Solving Real-World Challenges