LAS VEGAS – With the promise of “fast, used and smart” systems that are intuitive and powerful enough to drive user adoption, IBM unveiled its new vision for analytics at the Information On Demand (IOD) conference this week. For IBM, the business value in analytics is clear and considerable, and the company worked hard at the event to demonstrate this to 13,000 attendees through compelling client testimonials from companies such as Honda, HSN, and the Pew Charitable Trust. But the company is looking to drive more mainstream adoption, and took opportunity at the conference to showcase a refreshed line of solutions aimed at helping customers of all shapes and sizes translate analytics vision into reality. Cognitive computing, predictive analytics and cloud delivery represent IBM’s current keys to unlocking data insight, and IBM is hoping that updated capabilities in these areas will also facilitate the consumption of analytics by new groups and new market segments.
”Data is the new natural resource,” proclaimed the SVP of IBM’s Software Solutions Group, Mike Rhodin in his keynote on day two of the IOD event. But rather than focus on the actionable potential in data insight, IBM has found in a study of analytics usage released this October that many organizations continue to see more challenge than potential in Big Data: while researchers noted that close to 40 percent of companies see a return on investment within the first six months of analytics adoption, they also discovered gaps in executive sponsorship, trust and the skills needed for companies to execute effective analytics strategy. The “Analytics: A Blueprint for Value” report concluded that only a quarter of C-suite leaders are strong advocates for the use of Big Data and analytics, and a lack of trust in the interpretation and veracity of data is endemic in many organizations, as is a lack of skill – and the combination of analytics skill and business knowledge in particular, the primary attribute of the much sought-after ‘data scientist’.
So how specifically is IBM working to resolve these barriers to adoption? According to Rhodin, analytics serve as a ”universal translator” that can help businesses negotiate the information “tsunami” resulting from the explosion of mobile, social and machine data. Analytics-based products, he added, “play a central role in IBM’s quest to reach new markets, and new buyers” and the company’s “current task is to democratize analytics,” to reach users with varied analytics abilities, and to facilitate access to analytics to accelerate business transformation within client organizations. Specifically, IBM announced updated process-driven solutions in the following different categories, with rapid access via cloud:
IBM launched expanded DB2 with BLU Acceleration, which leverages in-memory RAM and columnar storage, analysis of data in compressed format, multi-core data parallelism and data skipping for more rapid analytics speed in database and data warehousing. According to IBM, expanded capabilities in the BLU Acceleration portfolio have produced “8 to 25 times faster reporting and analytics, and cases of more than 1,000 times faster answers to queries,” and “24 times faster query performance“ when used to complement in-memory Dynamic Cubes in the Cognos BI solution. For the customer, these performance enhancements provide what Les Rechan, GM, IBM Business Analytics, described as “analytics at the speed of thought.”
Analytics for everyone
To address the issue of gaps in analytics trust and skills identified in its research, IBM unveiled a new line of solutions designed to make analytics tools accessible to the business user. Today, most available software requires technical skill and facility with modelling; however, the non-technical business user is increasingly reliant on data-driven decision making. To provide this user with quick access to meaningful insight incorporating built-in predictive analytics, IBM has introduced a series of new or updated products.
For information specialists, IBM announced new capabilities based on the March acquisition of Star Analytics, Inc., for the automated integration of financial information, report applications and analytics tools, through the Cognos Command Centre. For the non-analyst community, the company also announced SPSS Analytics Catalyst, which finds the key connections in data by automating portions of data cleansing and modeling, interpreting results through natural language explanations, and presenting analyses with interactive visuals. Rechan called the product a “statistician in a box.”
And to enable users to explore the power of data visualization, IBM highlighted the Cognos Visualization Customizer, used for editing fonts, colors or icons, and the Visualization Marketplace on IBM Analytics Zone, which offers more than 30 visualization options.
IBM also introduced new products to expand its analytics portfolio. The first is “Project Neo,” an analytics platform featuring natural language capabilities and automated analytics to help business users better engage with data, which will be available in beta in early 2014. In response to a natural language query, Neo finds the appropriate data set, models the data, and presents results to the user without any iterative scripting. The solution features IBM’s Rapidly Adaptive Visualization Engine (RAVE), which automatically provides flexible and interactive visualizations to enable line-of-business users to quickly grasp the meaning in graphic representations of data. According to IBM visualization expert, Noah Iliinsky, users’ gravitation towards visual form is not coincidence: the human brain is wired to more quickly and easily detect patterns and mapping in optics than in text. Another key Neo feature is interactivity: in a product demo at the event, Neo presenters layered successive natural language questions onto data from an SPSS Catalyst repository, using different criteria to refine the answer. The result is a deeper dive into the data than is possible in static presentations such as PowerPoint.
To offer the contextual insight that can enable better decision making, IBM introduced Concert on Cloud, a mobile-ready, social analytics platform which offers collaborative tools allowing remote workers to view, discuss and add input to specific performance insights. An integrated analytics and project management or performance management tool, Concert’s single pane of glass management and collaboration capability is designed to provide quick response to data insight, and support for real time planning and forecasting.
At an organizational level, IBM pointed to Watson, the cognitive system of Jeopardy fame, which Watson GM Manoj Saxena described as “the next generation of computing.” An advanced analytics platform that has been adapted over the past two years for applications such as healthcare research and call centre management, Watson can learn, reason, sense and predict. Unlike other analytics platforms, Watson may be accessed by users through natural language queries, can “understand” the context in users’ questions, and can improve its own performance by continuously learning from experiences. Tasked with commercializing Watson technology, Saxena’s team is currently developing additional applications and expects to make several announcements in the coming year, including the opening of API’s for developers to access Watson via IBM’s SmartCloud.
To illustrate how new IBM technology can enable smarter decisions and practice, IBM VP of Analytics, Beth Smith walked conference attendees through a hypothetical applications of IBM’s new PMQ (Predictive Maintenance and Quality) solution. According to Smith, PMQ combined a number of IBM predictive analytics and decisioning technologies to “capture data, make predictions based on analysis of that data, and to enable real time action.” In Smith’s connected car example, predictive maintenance starts with the collection of sensor data on components, driver behaviour and outside environment from thousands of cars, which is combined with social data and warranty data. Based on this, and telemetry data streamed from car, the PMQ user (the car manufacturer) can predict car failures in real time, send this information to the car owner who can make decisions in real time about preventative maintenance, and to a garage which is alerted to the need to order in the appropriate parts. This example illustrates the ways in which analytics can deliver benefit throughout the supply chain: the manufacturer is able to avoid costly warranty expense, service delivery is rationalized at the garage, and the driver can escape the cost and inconvenience of vehicle failure.
A critical layer in making these solutions accessible to a broader market is cloud delivery of the applications. Based on its July acquisition of SoftLayer, a leading Dallas-based provider of cloud and IaaS services, IBM is building an impressive, OpenStack compliant portfolio of SaaS offerings that are accessible via a single-pane web portal or open APIs to allow customers to mix and match public and private cloud deployments. Currently, IBM has built over 100 SaaS offerings on the SoftLayer infrastructure architecture (including the products described above) to improve line of business access to analytics, and to facilitate access for smaller businesses, or departments within IBM’s mid-market or enterprise customer segments.
Marc Dietz, director of IBM SmartCloud Solutions, calls this growing SmartCloud platform an “evolution of the way we have been selling software to date, either through channels or direct. On one hand, the ability to find an application – a small analytic vs. a large application in some cases – try it before you buy it, download and use it, whether you are an individual or a department or a major IT shop, is a trend that will keep evolving.”
As important demonstrations of IBM’s intent to ‘democratize analytics’, cloud delivery of solutions such as Project Neo and Concert are in some ways a response to IT’s increasing inability to quickly deliver the services and information that business users have come to expect. But does marketing technology to the business end user introduce additional issues for the organization? Does it exacerbate long standing tensions between the IT department and line of business that have increased with the tendency towards rogue deployment of cloud applications?
IBM would argue no: according to Dave Laverty, the company’s VP marketing, Big Data and Analytics, historic issues between the business user and IT are being swept aside by new requirements. While IT departments are increasingly required to speak in business language – to generate outcomes – IT budgets are too tight for internal delivery of the IT functionality need by business users – when they need it. In Laverty’s view, by providing a tool that can make life easier, i.e. analytics for the business user, some of the burden of delivering IT capability is lifted from IT. The best practice in implementation involves both IT and LOB management, Laverty agreed, and he noted IBM’s recent efforts to drawing together CIOs, CMOs and even CFOs in technology discussions. As Chris O’Connor, VP strategy and development, cloud and smarter infrastructure, explained, IT will always have a role to play, through this role will change from technology “implementer” to “integrator,” with responsibility for brokering, integrating and maintaining services in support of business processes, and with ultimate responsibility for governance and security of corporate data.
Despite the seeming instant-on analytics delivery that IBM’s new products promise, the data component remains a thorny challenge that users will have to address. While Watson’s erudition appears magical, this artificial intelligence is based on the load of many thousands of documents into storage systems. And though Neo comes with an introductory natural language vocabulary preloaded, the solution provides a mechanism for introducing dictionaries and vocabularies and requires users to do their own ‘training’.
Presentation of data is another hurdle: while IBM’s Visualization Marketplace offers business users many visualization options, and the website contains a resource that explains the science behind standard graph types and what they are best used to do, bias can be introduced through inappropriate visual communication. According to Iliinsky, analytic tools now have the ability to identify the most relevant data and to “guess at” the graphic that is most appropriate to demonstrate relationships. This connection between data and presentation is “somewhat of a challenging puzzle, which is why it’s not already ubiquitous,” Iliinsky noted, and ultimately, a consumer of data-based insight must rely on the fact that a user of the visualization tools wants “to tell the right story.” Is this assumption more likely to ring true with the data scientist or the sales manager who is compensated on performance? By focusing on “democratizing” analytics capabilities, IBM is betting that both parties will come to appreciate the value of matching the medium and the message with the underlying data.