While many organizations have a general sense of the growing importance of data and analytics , the keynote speakers at Gartner’s Business Intelligence and Analytics Summit sought to illustrate just how essential proper data analysis is to driving all decision-making processes.
Titled "Analytics Leadership: Empowerment Without Anarchy," the keynote was delivered by two speakers: Rita Sallam, research VP at Gartner, and Frank Buytendijk, research VP and Gartner Fellow. The pair identified key approaches businesses should take to optimize their data: For one, analytics efforts need to be focused on specific business outcomes; and analytics should now be a core component of all business decisions.
Sallam stressed that data analytics are distributive: "It's cliché now to say that data is everywhere, but what's not obvious is that it's staying there—you just can't collect it in one central location anymore. There's just too much of it, and in too many places." She added that organizations need to leave data where it is and organize around it, as opposed to gathering it together and trying to force it to frame certain problems.
And with the Internet of Things (IoT) set to remake the landscape, companies will need to harness increasingly important IoT data to remain competitive and avoid being put out of business by retail giants such as Amazon, warned Jim Hare, research director at Gartner. In his session, titled "Transform Your Business with IoT Analytics Before You Are 'Amazoned,'" Hare identified five major types of IoT data: measurements data, situational data, transactions data, diagnostics data, and context data. Measurements data is the operational data from sensor readings, such as speed, acceleration, and temperature. Situational data refers to dynamic contextual data such as weather, location, and social media feeds. Transactions data represents the interactions between machines and people. Diagnostics data measures the health of sensors, machines, and systems. Finally, context data is static, slow-changing data such as customer info and product SKUs.
This wealth of data, along with increased complexity and automation, makes IoT analytics more challenging, Hare pointed out. The IoT generates high-volume, continuous data from multiple sensors, requiring companies to store, blend, and manage time-series data. And businesses will need to use multiple analytic techniques to make the most of IoT data, and embrace distributed analytics in particular. Lastly, more automation means that companies will need integration with operational systems and business process management suites (BPMS), and ensure bidirectional communication and control of endpoints. "When you buy devices today, half the time they have sensors on there generating data we just never really use," he noted.
Gaming also illustrates how data and analytics can be utilized in creative ways; an interesting case brought up at Gartner focused on video game developer and publisher Blizzard Entertainment, the company behind popular games such as World of Warcraft, Hearthstone, and StarCraft. Blizzard relies on data collected by Hadoop and Teradata to enhance the gameplay experience. Jon Gleicher, business and gameplay insights manager at Blizzard, identified three key areas in which his company works to continually improve gamers' experience: engagement management, business and gameplay insights, and data science. According to Gleicher, the engagement management team is crucial in providing a productive bridge between the game development and the business intelligence teams: "We want to be able to democratize a lot of our analytics," he said. Gleicher stressed that both business and gameplay insights inform Blizzard's employees, helping them to make better decisions. Finally, the data science enables Blizzard’s computers to be smarter, actively improving gameplay.