In 27 years at SAS, Keith Collins says his whole career has been about scaling to the enterprise. He wonders, "Why now do we call it big data? What will the next generation call it? Bigger data?"

Keith led a panel of business leaders at the Premier Business Leadership Series today who talked about the challenges of today's big data scenarios and discussed what organizations are doing to improve their business with big data. The panelists for this session were:

• James P. Adamczyk, Chief Technology Officer, Accenture
• Maura Hart, CIO, Winn-Dixie
• Mike Keppler, Senior Vice President of Global Sales, Marketing and Revenue Management Systems, Marriott International
• Michael Singer, Senior Editor – Technology, Economist Intelligence Unit

Reflecting the theme of the opening video sequence, which displayed a sea of ones an zeros, this panel discussed not just big data but how to find answers from within the deluge of digital noise. In fact, this group did a great job of summarizing many of the overall themes and announcements from the conference, including the conference press announcement about High-Performance Analytics and the recent Economist paper, Big Data: Harnessing a Game Changing Asset, both make good background reading for the topic.

Accenture has been tracking two trends in vertical applications: embedded analytics in process software and the challenges and opportunities of big data. Michael Singer defined 'big data' as being more than just volume (terabytes, exabytes, petabytes), it is also about variety (complexity), context, connectedness and velocity - so even a relatively modest data source can present huge challenges as it is part a complex decision environment.

Yesterday General Colin Powell talked about making decisions within the time frame of the "Decision Cycle." Today, high-performance analytics is about enabling decision cycles of minutes, not hours.

Adamczyk used insurance claims as an example, where Accenture is seeing many clients build analytics into the adjustment process to identify potential fraud while processing the claim. Near real-time analytics makes this possible, instead of waiting until the end of the month to dig through historical data and identify fraud after the claim has been paid. AllState and CNA spoke about similar efforts yesterday.

Now this does not mean that all data is equal. Bryan Harris yesterday spoke about relevance as a key measure of value, but then again it is hard to anticipate what data will be useful when innovative 'data consumers' find creative uses of data once it becomes available - an example of which is the open data initiative in the UK, which is enabling bottom-up innovation. The panel also spoke about the '4th V - value' the economist research showed that those companies that looked beyond reporting the things they always reported on found new value-creation opportunities in their data.

Mike Keppler also pointed out the huge uptick in data derived from mobile devices, which also brings an expectation of immediacy. "Mobile users expect answers now with decision cycles of seconds, not minutes, hours or days. For example, a mobile user who is looking to make a room booking using their smart phone expects the same high-quality, high-speed response as if they were using their PC or in the hotel lobby. They want the same deals, offers, insights and service they always get."

Marriott still sees one quarter of its sales through the online channel globally but growth in that channel has levelled off recently, says Keppler. "With mobile devices, even though they represent a small percentage of the total, they represent a 300 percent growth rate over last year."

The implications of mobile are far reaching. Online guests often book far in advance. Mobile users, however, typically book within 48 hours. "Now, the guest could be standing in the middle of Times Square looking for a place to stay, and we're with them wherever they go. Our mobile app is marketed as the perfect travel companion. We have to provide them with differnet things that maybe you didn't think about in a PC environment."

Michael said that the many business surveyed by the Economist complained that they simply cannot process their data fast enough. The batch processing cycle is greater than the required decision-cycle.

Maura Hart agreed: Winn-Dixie is a business that needs to make these decisions quickly in a fast-moving, highly competitive market and so can't accept long processing cycles. "Customers internally would like data as close to real time as they can get. There's an assumption from all the transaction systems taht we can get the data in real time."

Ultimately, says Keppler, "It's not just about how fast you can get the data and how complete it is. It's about: what is the value of that data to the business?"

Adamczyk says his experience backs up this point. "Make sure the problems you're addressing are solving your biggest problems. Loop back with the business to make sure we're applying techniques when looking at the data. I know that sounds obvious but you'd be surprised the number of times it doesn't happen."

Finally, Singer suggests the secret sauce is looking at data from different perspectives. "Make sure you're not just looking at the same data again and again. Those who broke through that mold told us they started looking at different parts of the data and started getting real insight."

So, ultimately, it's not the size of your data that matters. It's what you do with it - and how you prioritize your analytics projects that matter. Yes, data sizes are large and varied and unstructured today - and data consumers need answers faster than ever before - but the right technologies and the right business strategies still make it very possible to overcome those challenges.

Peter Dorrington also contributed to this blog post.