Perceptive Information Strategies

As it zoomed toward the international space station orbiting about 200 nautical miles above us, the shuttle Discovery had to dodge clouds of space debris. Once there, the threat to human and technology treasure did not go away; apparently, pieces of debris as small as a grain of sand could cause serious damage. Perhaps I’m reaching a bit, but this seems like an apt analogy for organizations that are beginning to journey into the realm of social media to “discover” data patterns, relationships and other insights that have eluded them thus far in their analysis of data sources closer to home. All that unstructured stuff hurtling by on Facebook, MySpace, Twitter and other social media could be important. But, getting a closer look at it could also be dangerous; companies could run afoul of privacy rules and online social etiquette in ways that might damage their reputations.

At Predictive Analytics World, I noticed that when keynoters spoke about social media, the audience of data miners, software developers, business and statistical analysts put down their Blackberries and listened intently. You could feel it in the room: the next big data gold rush, or at least the next important phase of search engine optimization and e-marketing.  I blogged at Intelligent Enterprise and Ventana Research about the conference and the growing awareness that many operational business intelligence (BI) implementations will need predictive analytics to make good on the “actionable information” promise of these applications. However, I am interested in how data mining and predictive analytics are being applied to social media and will be writing about this topic from time to time here.

Lauren McKay of CRM Magazine wrote about a Predictive Analytics World keynote by Usama Fayyad, CEO of Open Insights and former chief data officer for Yahoo!.  Fayyad talked about the success of “behavioral targeting,” or analysis of search and Web browsing, and how it might be applied to social media interactions and increase the ability of online businesses such as Yahoo! to react and respond within a relevant time period. The implementation of predictive analytics tools and techniques could allow companies to improve their response rates not only by knowing more about their customers but also gaining greater mastery of time. Right now, it seems like a lot of online businesses are shooting in the dark about how long they should keep their marketing offers alive. If they could analyze (and begin to predict) the buzz about interests and products in social media, they could gain a better understanding of the life expectancy of marketing initiatives.

“Big as this industry is, we still haven’t figured out this business,” Fayyad said humbly in his keynote. The potential benefits and dangers of social media – or really, the whole field of behavioral information that will grow to include location data as people rely on mobile devices for search and social interaction – are enormous. We don’t want to get too breathless; after all, inertia rules how most large organizations market their products and services. But, tapping behavioral information could eventually rewrite many rules of marketing. A big question, along with cost-benefit analysis of course, is how much nosiness people will tolerate. However, that’s a topic for another blog.

As I kick off this blog from my hotel room in Las Vegas, there are few clouds in the sky. The morning sun radiates through the desert dust kicked up by the activities of a million people. The pale yellow sky sits above the tall wings of the casinos and the dark, flat shadows they cast on the shorter buildings below them. Thus, without any physical clouds drifting by to offer themselves as models, I’ll have to use my imagination to think virtually…which, of course, makes perfect sense.

“In the cloud” has become a term of art tossed about perhaps a little too easily by vendors, VC-backed start-ups, IT information strategists and lately journalists in the non-IT, mainstream business media. In a sense, the idea is to connect the unknowing with the all knowing: Users and/or developers can execute whatever they need to do – on mobile, laptop or other devices – while up there on some cloud the necessary computing power, scalable software and expert professionals respond without users or developers having to know how it’s getting done.

The business arrangement may remind the grizzled veterans among us of timesharing, but keep in mind that this is 2009, not 1969. Today, few have even heard of green screens. Users expect rich interfaces and some ability to personalize and customize the services; and to succeed, the computing environment in the cloud can’t waste any of its precious resources. Providers of the cloud computing services will have less and less tolerance for sloppy software that doesn’t use resources efficiently. At the same time, we will surely see a spike in demand for tools that enable real-time performance monitoring, analysis and response to problems inside and outside the clouds.

In other words, most of today’s systems can’t just be put inside a mass of water vapor or frozen crystals (thanks, Wikipedia) and called “cloud computing.” And there’s more to cloud computing that simply “pay as you go” models, aka software as a service (SaaS). For this reason, cloud computing could spur innovation even beyond the important developments centered on virtualization. To satisfy users’ demands, database architectures, for example, may have to make more creative use of memory-based and near-line storage as well as search or other programs that are different from the query style of SQL. As well, organizations will have to devote greater attention to business rules, policies and governance issues related to information sharing, privacy and quality.

Cloud computing is changing the vendor landscape. We now have “coopetition” arrangements between the likes of Amazon and IBM, with the latter recently announcing that DB2 works in the Amazon Elastic Compute Cloud. IBM’s recent announcements from its Blue Cloud Initiative focused on Tivoli for managing data centers set up as “private” clouds; John Foley of InformationWeek offers a nice analysis here. Other vendors big and small are adjusting their strategies based on how strongly they believe the market will be for cloud computing and where they can take advantage: in other words, the marketing pieces are still in motion at Hewlett-Packard, IBM, Microsoft, Oracle, SAP and elsewhere.

Many bloggers have referenced the useful analysis recently published by researchers in the Electrical Engineering and Computer Sciences Department at the University of California, Berkeley. I recommend reading this document, even though it won’t be the final word on this still-developing phenomenon. I will be using my vantage point in this blog to explore and examine where the clouds are going, particularly as they relate to information management, business intelligence, analytics and policy directions for information governance. Yes, even on a cloudless day in Las Vegas.