CAMBRIDGE, Mass.-The role of the CIO is constantly in flux, but the rapid emergence of big data, analytics and cloud technology--and the accompanying proliferation of data itself--has further strained IT innovation and added complexity. All this comes at a time when companies expect IT to do more with less and the balance of IT services spending is poised to tilt away from the CIO.
With that at stake, the leading CIOs and CEOs who spoke at last week's MIT Sloan CIO Symposium gave the senior IT and business executives in the audience food for thought about how to use data analytics and cloud technologies to improve businesses processes without putting additional strain on IT resources.
Data Be Nimble, Data Be Quick
Guess? has many product lines, but the decision to sell one specific garment or accessory over another can lead to tremendous sales. The data analysis required to make this decision used to require running queries against a spreadsheet. It took hundreds of hours and resulted in many failed queries-and phone calls to Senior Vice President and CIO Michael Relich. Knowing that Guess possessed the master data, Relich and his team developed 2-D barcodes that buyers could scan with an iPad to pull up data on a particular garment.
This project wasn't part of any strategic plan or long-term mobility initiative. Rather, it stemmed from an obvious need. Many of today's big data analytics projects do.
For William Miller Jr., senior vice president and CIO at semiconductor chip maker Broadcom, this means getting in front of demand rather than merely reacting to it.
At Broadcom, this happens when the IT and engineering teams get together in an operating environment. This has led to the deployment of enterprise social network technology such as Jabber and Yammer, which lets engineers self-organize and then disassemble when a task is done. "It makes engineers maximally productive so they don't spend a lot of time setting up their own environments," Miller says. "It's a highly productive way or organizing."
Broadcom's engineers also work in a private cloud environment, which gives them the flexibility to run jobs any time of day (or night). This flexibility really shows its values when Broadcom makes an acquisition and employees of the new company come in on Monday to see that their assets were fully integrated into the grid over the weekend. "That reassures us that these investments are worthwhile," Miller says.
Data Visualization Gets to What's Really Needed
Big data investments are worthwhile only when people can visualize the information that analysis begets. The president of Liberty Mutual Global Specialty used to receive 80-page operational reports every quarter, and it took several hours for the executive team to go through each one, says Mojgan Lefebvre, senior vice president and CIO of Global Specialty.
Turning that long report into a list of key performance indicators required several steps, Lefebvre says. First was deciding key questions to answer. Then came agreeing on and prioritizing those questions. Next was building a set of reports. Data visualization was the final step. It's about "getting data down to what's really needed," Lefebvre says, adding that it took more than one quarter to get the report right.
"Getting people to adopt a data warehouse is a huge issue," Relich says. It's especially critical in an industry such as retail, where buyer's intuition is a well-honed skill. The solution for Guess has been to deliver meaningful data in an evolving format-first to the BlackBerry, then in Flash-powered dashboards made by graphic designers and, finally, in an iPad app. Giving buyers store-by-store sales data expedites decision making and reduces the number of planning meetings, Relich notes.
Real-time data visualization need not be limited to numbers. At Liberty Mutual, onsite risk engineers can take pictures and upload them to a database. From there, the images are checked against insurance rating engines. This lets underwriters make pricing determinations in hours as opposed to days, Lefebvre says.
Private, Hybrid Clouds May Answer Security, Latency Questions
The device that an employee uses to take those pictures brings up many questions for IT leaders, especially as the honeymoon phase of the BYOD movement ends and the security of the cloud services many use in conjunction with mobile devices is called into question.
Related: How to Craft the Best BYOD Policy
Panelists in a discussion about the evolving cloud agenda suggest that, in addition to security, complexity must be considered as organizations consider the cloud.
After years of cloud vendors easily selling services to human resources, sales and other departments, often to the chagrin of IT departments, IT leaders are trying to rein in corporate assets and improve user management.
Now, says Richard Dorough, managing director of PwC Forensic Services, it makes sense to evaluate the data itself, the service level agreement and the cloud service security before moving anything to the cloud.
For Scott Blanchette, senior vice president of information and technology services for Vanguard Health Systems, this serves as a de facto endorsement of hybrid and private cloud deployments, especially for healthcare and other heavily regulated industries.
The value proposition for cloud services, Blanchette says, needs to be better, faster, cheaper, more flexible and more secure than what Vanguard can do internally. It's no surprise that security and privacy are consequential barriers to entry for cloud service providers; Google won't do it, he points out, while it took four years for Amazon Web Service to get FedRAMP certification and meet the U.S. Department of Health and Human Services' cloud security standards.
More MIT Sloan CIO Symposium News: Listen to Your Customers-But Don't Expect Them to Have All the Answers and Here's What Steve Jobs Thought of CIOs (CITE World)
Its better, then, to look at the cloud on a workload-by-workload basis, says John Roese, senior vice president and CTO and EMC. The "coldest of the cold cryptographic storage" might be a good candidate for the public cloud, he says, but data that will be needed for real-time analytics may be subject to low latency if, say, it goes from Amazon's servers to yours and then to a mobile device.
One way to combat this, Roese says, could be a virtual desktop infrastructure that makes a mobile device nothing more than a data presentation layer.
Teach Your IT Talent Well
These points are all moot, of course, if an organization lacks the talent and the IT leadership to advance data analytics and cloud initiatives that meet users' needs.
Since Vanguard is based in Nashville-"not Tech City, USA but Music City, USA"-the company's data centers are in locations such as Boston and Silicon Valley, Blanchette says. The company also recruits from other vertical industries, since IT professionals can learn healthcare more quickly than they can learn tech.
Once that talent's on board, it's a matter of nurturing it. This means teaching IT staff to articulate the value proposition of what they do, Broadcom's Miller says. It also means going beyond simply responding to user requirements and finding ways to change core business processes. It's about return on investment, Miller says, not just software and infrastructure.
Or, as Relich puts it: "I tell my team, 'Don't give them what they ask for; give them what they need.'"
Read more about byod in CIO's BYOD Drilldown.