Reconsidering Legacy BI Architectures in the Age of Big Data

By now, we all know the value of big data analytics. In the 1990s, data mining and business intelligence (BI) efforts were used mostly for after-the-fact reporting and problem solving. The need for these capabilities will always remain, and high levels of confidence and investment in these disciplines will continue. However, the current focus on big data analytics emphasizes predicting and influencing the future, based on real-time analysis of 100% of enterprise data.

The harsh reality is, in this era of big data analytics, organizations that cling to the “old way” of doing things—relying solely on traditional BI and data warehousing—will lag behind, as more agile, data-driven competitors quickly move ahead. Some organizations, particularly smaller companies, may believe that the ability to collect, analyze, and derive insights and value from all pertinent data—both within and beyond the enterprise—is simply out of their reach. These companies may also be reticent about the thought of losing their existing BI investments. But the truth is, traditional BI systems and tools are still relevant in a big data analytics world.

Synergies Among Successful Modernization Initiatives

The prospect of undertaking a big data analytics initiative can be daunting. But, in fact, it’s hesitation that should be feared, not big data analytics—because regardless of what an organization may (or may not) be doing, with each passing second, competitors are leveraging information to increase their agility and competiveness.

Further, new pricing models and open source solutions make it possible to reduce some of the costs related to big data analytics. The key is to start unlocking the collective value of enterprise data as quickly as possible, while minimizing significant upfront investments. While there is no uniform prescription for realizing success with big data analytics —and extending the value of BI investments as part of this effort—there are certain characteristics that seem to favor success:

• Quick data discovery and services: Today, businesses can use the cloud to set up discovery environments incorporating all types of data—traditional ERP, social, streaming, and sensor—in just 1–2 weeks. This means organizations can start quickly, which represents a huge industry advance. When modern data analytics are applied to all information sources in parallel, each of these sources becomes inherently more contextual and rich.

And yes, traditional BI and data warehouse systems can play a key role here. For example, a car manufacturer recently analyzed sensor data and determined that a particular piece of equipment in certain cars needed to be recalled. By aligning this sensor data to its traditional data warehouse/BI system, the company was able to quickly and accurately determine the exact production runs needing to be recalled, the cars in these runs, and the impacted customers. This saved significant time and effort in the recall process while also enhancing customer safety.


The harsh reality in this era of big data analytics is that organizations that rely solely on traditional BI and data warehousing will lag behind as more agile competitors move ahead.


• Empower all people: Rapid set-up of discovery environments—as described above—fosters enterprise-wide data sharing and an enhanced ability for all employees to collaborate. Lately, media headlines have hyped up the shortage in data scientists. While data scientists are important, the goal of big data analytics and BI modernization should be to make all workers more agile, regardless of the type of information they’re working with. This includes “new” types of data such as social, sensor, and streaming, as well as traditional data.

Successful BI modernizers have this mindset and encourage creativity, innovation and experimentation at all levels of the company, through broad availability to data and advanced analytics. These companies also support direct IT and line-of-business (LOB) collaboration, so IT teams can help LOB managers develop repeatable solutions for tapping critical insights from all types of data and gaining competitive edge.

• Strike a balance between agility and infrastructure solidity: As noted, companies often believe that big data analytics initiatives mean scrapping one’s existing BI investments and investing in expensive new infrastructure. This is not true. Hybrid data management has evolved as a means to deliver industrial-scale analytics across traditional BI systems with their structured data, as well as unstructured data such as voicemail recordings and social commentary.

Hybrid data management is also available as a service in a pay-as-you-go model, making it easier for companies to begin their big data analytics journey while avoiding costly, upfront investments and technology obsolescence.

Tackling Big Data Now

Imagine the feeling of finding a $20 bill in the pocket of a pair of jeans you haven’t worn in months. You find yourself overly excited about what you will do with this treasure you never knew existed. Such is the life for the many businesses today applying big data analytics. The data they collect about customer behavior, perception, and even the way their equipment is or isn’t functioning can be translated into dollars for the business.

The good news is that even those businesses which believe they don’t have the means to tackle big data analytics can also have an opportunity to feel this way. These businesses may be hesitant—but they don’t have to be. Service offerings that leverage deep, firsthand experience and knowledge across industries are available to guide organizations through the kickoff process. And they will tell you: It’s OK to start small. It’s OK to experiment to find out what works and what doesn’t. Organizations don’t need to make huge, upfront expenditures, nor do they need to compromise on their existing BI and data warehouse investments. But, these systems will need to be modernized, stretched, and leveraged in a new way as part of an organization’s broader data ecosystem.


Image courtesy of Shutterstock.



Newsletters

Subscribe to Big Data Quarterly E-Edition