The new approach to managing data, big or small
Moving from an application-centric to a data-centric strategy
By Tony Fisher, SAS General Manager
Have you taken a minute recently to catch your breath and notice the pace of change? It can be staggering. The challenges that organizations face today – regulatory compliance, social communities, and a constantly connected network of customers, employees and partners – would astonish businesses just 10 years ago. These new concepts are changing the landscape of IT and business alike, especially in the amount and sophistication of data available.
Yet, through all this change, there is one unfaltering constant: the amount of time available every day. There are only 1,440 minutes in a day. There will only EVER be 1,440 minutes in a day. While there is more data to process and analyze, it's impossible to manufacture time. Regardless of how large or demanding these new challenges are, every organization still only has 1,440 minutes to meet their challenges of that day and prepare for the future.
In the IT world, three primary challenges stand out:
Size. Unfortunately, typical data processing methods of the past can't withstand the new data volumes and speeds. Like a torrent of water being forced into a narrow channel, the onslaught of big data is overwhelming existing IT environments and causing organizations to rethink how they manage information.
Complexity. In recent years, IT infrastructures became even more complex, as some business units begin to pursue cloud-based applications versus on-premises solutions. So, you have data residing outside the enterprise that may need to be aggregated to create an enterprise view of customers, products, finances, etc.
Quality. Data – small, medium or big – has always been a "garbage in, garbage out" proposition, where even the most sophisticated analytics can be undone by poor-quality data.
Simply finding ways to store more data is a reaction, not a strategy. Organizations must now focus on how to do more with the same amount of resources and how to re-engineer existing data management processes for the big data world.
The changing nature of data management
For years, organizations have managed from a business function perspective, and the resulting data from these functions fed business applications dedicated to a business unit or division. For example, organizations built CRM systems for sales and marketing, ERP for finance and operations, and so forth. These business units looked to IT to manage the applications and the underlying data, yet IT was disconnected from the business processes behind that data.
In many organizations, the divide between business and IT groups led to confusion over "who's in charge" of an organization's information assets. The fight for control (or, flight from responsibility, in some cases) led to a data management strategy that could only focus on existing problems without anticipating future demands, including the growing volumes of data.
This application-centric environment also created different pools of data within the enterprise. Different groups maintained distinct views of customers, products and other assets within their various business applications. As a result, it can take most companies a few days to compile a single customer list simply because the information resides in multiple systems. Application-driven organizations now face more data in more places, compounding the problem.
Becoming a data-driven organization
Even though the volume of information is the primary focus for many organizations, the big data phenomenon is not just about size. It's about complexity. It's about the organizational impacts of this data. Organizations are actually facing "big, complex and unknown" data. With applications residing on-site and in the cloud – and data arriving from an array of third-party sources – a sound data management strategy becomes even more vital.
To cope with the changing nature of information, organizations must transition from an application-driven focus to a data-driven approach. Applicationdriven organizations are internally focused. The applications that drive the business functions are the only way to understand the true health of an organization. These organizations make decisions that can affect a line of business or a division, but due to the limited scope of the applications, they struggle with enterprise efforts.
Innovative, data-centric companies view information as a common, shared asset – as valuable as buildings, employees, production equipment and intellectual capital. A data-driven organization can look beyond processes or lines of business to the entire organization. By concentrating on managing the data, organizations can improve decision making on a cross-functional basis because the data improves across the organization.
This data-driven focus is even more important in the world of bigger, more complex, and faster data. Financial institutions need to check the validity of transactions every second. Pharmaceutical and life science organizations must evaluate new substances by running millions of scenarios to test new materials virtually. Manufacturing and retail companies strive for a streamlined supply chain by implementing RFID tags and processing the resulting stream of logistical information. All of these efforts bring in massive amounts of data in new and different ways.
To build a data-driven enterprise, start with a vision of where the company needs to go and what actions are necessary to achieve this vision. Next, examine the business initiatives that will be integral to this vision and how to reorganize the company, if necessary, to achieve these goals. This may involve realigning IT resources around business units or finding data "stewards" in your organization who have a blend of experience working on business processes and the supporting data.
However, unlike an application-driven organization, a data-driven business doesn't stop there. The final step is to understand the data that will drive these strategies. Here, you're looking for data that will support the corporate vision regardless of the originating applications. By creating a framework to manage data across the enterprise, data consistency improves on a cross-functional basis, leading to more uniform operations and analytics across the enterprise.
Bio:Tony Fisher, SAS General Manager, joined SAS subsidiary DataFlux as President and CEO in 2000. In his years at DataFlux, he guided the company through tremendous growth as DataFlux became a market-leading provider of data quality and data integration solutions. Fisher is also a featured speaker and author about emerging trends in data quality, data integration and master data management.
Tony Fisher, SAS General Manager