Silicon Valley entrepreneurs, investors, IT vendors, and corporate executives gathered at Stanford University on March 25, 2011 for the SDForum Next Wave Analytics Conference to explore innovations driving the next wave of analytics.
New Sources of Data and Usage
John Fraser, a leader in Accenture’s Analytic Organization, moderated a panel that highlighted the key features of analytics’ next wave: real-time, predictive, immediately actionable, and encompassing new sources of data. “We are now looking out the front windshield,” Fraser said, “rather than the rear view mirror.”
ShotSpotter sensors collect one of the most startling new data sources discussed during the conference. Placed high on city buildings, the sensors detect gun shots in real-time, pinpoint the location to within ten meters, notify police dispatchers, direct surveillance cameras toward the scene, and send officers directly to the crime. According to Ralph A. Clark, president and CEO of ShotSpotter, the analytics enabled by the sensors identify short term and long term patterns that empower cities to optimize scarce policing resources.
Less dramatic but remarkably practical, Streetline sensors make cities smarter by gathering real-time data on parking, directing consumers to free spaces and police to expired meters. Holding up a tablet computer displaying free parking spaces in Los Angeles, Zia Yusuf, CEO of Streetline, hopes that the Streetline mesh network will reduce the thirty five percent of city traffic caused by people driving around looking for spaces. The analytics generated enable dynamic pricing and better congestion management.
In promotion of clean energy, Enphase places sensors into inverters, devices within solar panels that covert direct to alternating current. According to Ilen Zazueta-Hall, Software Product Manager at Enphase Energy, the sensors send data real-time over electrical wires to a gateway in the home, which relays the data to Enphase servers. Enphase analytics include rich data visualizations that direct homeowners to the specific location of a problem, assuring efficient energy collection.
Morning Fireside Chat: Tom Peck and Sanjay Poonen
Switching focus to the needs of large enterprises, Sanjay J. Poonen, President, Solutions Go-to-Market for SAP, and Thomas R. Peck, Jr., Senior Vice President and Chief Information Officer, Levi Strauss & Co., shared the stage for a fireside chat. Poonen held up a tablet computer to show the audience SAP’s balanced scorecards, illustrating how data and analytics form the nervous system and brain of an organization. Speaking of recent trends, Poonen stated that the rapidly expanding volume of data and need for real-time analytics require that organizations shift away from traditional data warehouses and nightly uploads toward in-memory systems fed by real-time data streams.
Peck told how Levi Straus, traditionally a regionally-based organization, needs to globalize IT to support the organization’s current strategy of global brands. As part of this transformation, Peck explained, Levi Strauss must make analytics real-time and immediately actionable by placing it directly on the mobile devices of store managers and consumers. Rather than backward looking reports, Peck wants to develop predictive analytics by mixing transactional data with survey data on consumer sentiments.
New Deployment and Data Architecture Models
Keeping the focus on large enterprises, Eileen Boerger, President, Agilis Solutions, led a panel discussion of big data experts: Scott Burke, Senior Vice President, Yahoo!; Dr. Anant Jhingran, IBM Fellow, VP and CTO, Information Management Division, IBM; and Oliver Ratzesberger, Senior Director Analytics Platform, eBay.
Jhingran took issue with Poonen’s comments regarding data warehouses, arguing that data warehousing and the grunge work of ETL (extraction, transform, and load) remain key to analytics because of the need to incorporate data from disparate applications. The challenge now for analytics, according to Jhingran, is to achieve speed and agility while reaching out to encompass less structured data sources. Achieving needed speeds, Jhingran stated, will require top to bottom software and hardware optimization.
Ratzesberger spoke to the challenges of meeting the demand for analytics at eBay, where analytics drives every aspect of the business from search optimization to the surveillance of infrastructure. Business units demand solutions fast. In response, eBay IT pioneered analytics as a service—a private cloud system for the self-provisioning of virtualized data marts pointing to a single underlying MPP (massively parallel processing) platform. This approach, according to Ratzesberger, solves the problems of duplicate, out-of-sync data and wasted storage. To head off potential bottlenecks on this single platform, Ratzesberger explained that the system utilizes a rules engine to evaluate user queries and sophisticated workflow management to spread out the load throughout the day.
Also facing the challenge of big data, Yahoo started out with home grown solutions, according to Burke, but not all of these scaled. Yahoo now deploys multiple proprietary solutions from Oracle, Microsoft, and MicroStrategy. And as an enthusiastic believer in open source, Yahoo pioneered Hadoop, an Apache project that enables data-intensive distributed applications and scales to thousands of servers.
Asked whether analytics had reached the limits of what it could contribute to enterprises, the panelists answered with a resounding no. Scott explained that the optimization of online advertising through analytics remains an unbounded opportunity. Ratzesberger said that eBay had barely scratched the surface of the many use cases enabled by ever more sophisticated analytics. Jhingran emphasized that analytics represents one of the three key growth areas for IBM and the foundation of its Smarter Planet campaign.
Afternoon Fireside Chat: Simon Khalaf and Sharon Wienbar
During the afternoon fireside chat, Simon Khalaf, President and CEO of Flurry, and Sharon Wienbar, Managing Director of Scale Venture Partners, discussed the history and future of web analytics.
Khalaf recalled how in 1996 the focus was on traffic. Few thought about segmentation, personalization, and monetization. Then Amazon pioneered the use of web analytics to improve the consumer experience.
Wienbar explained a tactic enabled by web analytics known as retargeting whereby retailers track those consumers who have abandoned transactions in an effort to draw them back to complete the purchase.
Khalaf and Wienbar agreed that web analytics offers great promise when applied to mobile devices. The mobile space, however, poses challenges, according to Wienbar, due to its fragmentation and the complex ecosystem of application developers, operating systems, carriers, and manufacturers. In addition to fragmentation, Khalaf pointed to the growing concern over privacy. He noted that, while critics must be careful to differentiate advertising optimization from fraud, the industry must do a better job at disclosure.
The Investor Perspective
Following the fireside chat, Harold Yu, a partner at the law firm of Orrick, Herrington & Sutcliffe, led a panel discussion on the investor perspective with leading venture capitalists.
Ping Li, a partner at Accel Partners, divided the analytics industry into data platforms, analytic tools, and applications. Next generation analytics, Li said, will go beyond the firewall to encompass external data. In the software industry, in particular, Li said, data from customers has become a driving force in product development, enabling developers to test which features offer the greatest value.
Lars Leckie, Principal at Hummer Winblad, stated that next generation capabilities have sparked a renewed investor interest in analytics. Leckie favors companies that put analytics at their core, pursuing business models made possible by analytics. He hopes that some of the start-ups following this strategy will avoid acquisition and grow into large independent companies.
Asheem Chandna, a partner at Greylock Partners who invests in enterprise IT companies, noted that analytics has become a top priority for IT and a driver for cloud computing. Vispi Daver, a partner at Sierra Ventures, believes that analytics has become a key form of intellectual property and an effective barrier to entry. He noted that venture capitalists have already scored home runs in analytics and he expects many more.
Afternoon Keynote: Eric Peterson, Web Analytics Demystified
Eric T. Peterson, the founder of Web Analytics Demystified, told the audience that dashboards are not a strategy, only a component of a strategy. The goal is to analyze web analytics using business acumen to achieve organizational goals. Too often, according to Peterson, dashboards offer little other than eye candy and degenerate into Excel hell. He asked when was the last time a dashboard had prompted an action that increased revenue.
Peterson spoke of an analytics hierarchy starting with data and leading to information, insights, and recommendations. Moving up this hierarchy to achieve business value requires people, processes, and technology, not just technology. Peterson advocates a hub and spoke organization with a core set of expert analysts working with savvy business users throughout the organization.
After Peterson’s keynote, Ted Shelton, CEO, Open-First, led a panel discussion on consumer targeting. Jim Dai, Founder and CEO of CalmSea, argued that many social media managers shy away from using social media for promotion when, ironically, customers are looking for deals. He believes that businesses should use social media to convert viewers into customers.
Expanding on this theme, Dennis Yu, Managing Principal of Facebook Marketing at Webtrends, pointed out that many companies have amassed social media fans and properties. Now they must figure out what to do with them. The value of a fan, Yu said, depends on what you can do with that fan.
Josh McFarland, Founder and CEO, TellApart, stated that organizations need to do a better job at managing data. When clients ask McFarland which data they should load into an analytics platform, he says all of it. In most cases, he argued, more data trumps better algorithms.
Phil Davis, the Managing Director Marketing Services at RapLeaf, noted the importance of looking for actionable tools. “If a dashboard is not actionable in two weeks,” he argued, “get rid of it.” He also explained the importance of distinguishing three forms of data: first party data created within an organization, second party data pulled from partner firms, and third party data extracted from diverse external sources.
Closing Keynote: Bill Schlough, San Francisco Giants
Bill Schlough, CIO of the San Francisco Giants, closed the conference on a celebratory note with highlights of the Giant’s World Series victory and an explanation of how analytics contributed to that victory.
Winning costs money, and analytics boosts ticket revenue. The Giants, Schlough said, are the only club in major league baseball to use dynamic pricing for every seat. The dynamic pricing system, which processes multiple inputs such as game time and opposing team, has increased ticket revenue seven to eight percent.
And analytics can enhance winning performance. The Giants are pioneering the use of analytics for player performance, going beyond measuring pitching speed to capturing every player movement on the field. While it has taken time to develop algorithms that can identify relevant movement (and filter out the pigeons that keep flying in front of the camera), the Giants expect to use the data for player benchmarking, pinpointing weak spots, and monitoring improvements.
Echoing a theme heard often throughout the conference, Schlough expects big wins for the next wave of analytics.