Right now, analytics suffers from too much math, too much marketing hype -- and too few sheep.
When you write about analytics you get all sorts of notices about things poised to change business as we know it. Analytics, of course, involves math. Math is one of the three Ms that matter in business, along with marketing and making useful things. But there's often a chilling vagueness around what analytics actually means to the bottom line.
That's peril number one, finding the benefit of analytics.
I spoke recently with the founders of Sideband Networks, a startup that combines analytics with some novel software it has developed to let companies parse their networks as traffic flows through it. "The value is in the speed at which we can perform the analytics and correlate past and present data or past and present analytics, to give you some actionable data," said Sherman Tang, Sideband's CMO.
But just what will companies do with that actionable data? Tang and his co-founder, Sideband CEO Zane Taylor, had a lot to say about the problem they are solving, but not so much about what solving the problem does for companies. I can guess that it might let companies predict when they're about to get hit with a denial-of-service attack, or ferret out data leakage while it's happening. But they didn't say that.
All they said was real-time analytics will change how networks get viewed.
I should praise Sideband for not hyping a product that is still under development. A bunch of you will probably read this and post below all the reasons that it would help to know what's happening on the network in real time. But where is the use case?
If the first peril is an inability to say how analytics matters, the second peril is waiting for it to matter.
At the Connected Insight Summit hosted by Activate Networks, I heard a talk by Nicholas Christakis, the Yale academic whose social networks research is at the heart of Activate's technology (originally, it was called MedNetworks.)
Christakis told us that when we know the structure of a social network, we can predict how disease (or gossip) will spread. It's great stuff, and Christakis even had a slide showing how a network of physicians can be analyzed in a way that creates natural, efficient structures for accountable care organizations (ACOs). ACOs are a promising way to contain healthcare costs, so Christakis' work could make it easy to build networks that reduce costs without causing patients to give up their current doctors.
Such a thing could be huge, given healthcare's status as a combination grail/black hole for technology spending.
Here's the problem: I talked with Activate CEO Larry Miller during a break in the conference, who told me that no one has structured an ACO using social network analysis in the real world.
Maybe Christakis summed up the problem when he said of social network analytics: "There's an obsession right now of looking for shepherds. But you don't just need shepherds, you need sheep, and you need to be able to algorithmically identify them."
No sheep, no restructured ACOs.
Then there's the third peril: analytics as a small bucket. I talked recently with Frederic Laluyaux, president and CEO of Anaplan, which uses analytics to create a business modeling and planning platform. It creates a much faster way for companies to do their financial planning than a tool like Excel. Big firms like Diageo and Hewlett-Packard use Anaplan to bring their financial planning and operations closer together. Laluyaux even had a business analyst write a song praising the product (I think that's an example of a sheep). Laluyaux thinks there's a $10 billion market for products like his. That's big, but also tiny, in the grand scheme of the world economy. Is that what analytics gets us?
Real-time, social network-driven analytics sounds big. Getting there should, at the least, enrich traditional data. "The classic approach to open data is data mining, which doesn't let us take into account how people behave," Activate's Miller told me. Capturing how people behave matters because groups of people then become what Christakis calls "sensors" for the real world. It sounds fascinating and important.
What we need to do now is get there. Maybe we need some sheep to take a leap. Right now, analytics suffers from too much math, too much marketing and too little making of useful things.
IT leaders must know the trade-offs they face to get NoSQL's scalability, flexibility and cost savings. Also in the When NoSQL Makes Sense issue of InformationWeek: Oregon's experience building an Obamacare exchange. (Free registration required.)
6 Tools to Protect Big DataMost IT teams have their conventional databases covered in terms of security and business continuity. But as we enter the era of big data, Hadoop, and NoSQL, protection schemes need to evolve. In fact, big data could drive the next big security strategy shift.
Big Data Brings Big Security ProblemsWhy should big data be more difficult to secure? In a word, variety. But the business wonít wait to use it to predict customer behavior, find correlations across disparate data sources, predict fraud or financial risk, and more.