Posted by: Wendy Schuchart
big data, Business Intelligence, data privacy, Risk management
NEW ORLEANS — The highlight of IBM PartnerWorld 2012 was not the sales updates nor the award winners, but a half-hour presentation by IBM’s chief scientist, Jeff Jonas. The Las Vegas resident genius invigorated the 1,000-strong crowd of IBM’s partners and guests, starting with the simple tale of a very complicated puzzle experiment.
Jonas took a puzzle and removed 10% of the pieces, threw another four partial puzzles into the mix, then a duplicate of the first puzzle, also with pieces missing. He watched how long it took four teenagers to realize that they had been duped (a little under three hours) and how long they sorted out duplicate “data,” as well as data that didn’t belong to the picture they were creating.
For instance, what was a Las Vegas neon sign doing in a puzzle that clearly depicted “hillbillies” on a porch? Jonas explained how this represented exactly the constraints our big data analysis efforts operate under. Midmarket companies aren’t playing with a single puzzle showing the neon landscape of Las Vegas or a charming vignette of some country types playing jug-band music.
As IBM’s chief scientist explained, until you have context, you wonder if the puzzle piece with flames on it is showing a fire in a fireplace or a fire in the kitchen. As the puzzle experiment at PartnerWorld demonstrated, we are in danger of throwing out “bad data” that could become useful in the future. Scott Lowe discussed this phenomenon in our tip last week on Big Chaos.
As a big data analytics junkie, I see the inherent value in using technology to make these connections. For instance, at PartnerWorld, Jonas cited the example of a top five major U.S. retailer at which two out of every 1,000 new hires had been charged with theft from that very same retailer. It boggles the mind that HR wasn’t talking to the loss prevention department, and yet it’s easy enough for a giant enterprise to make such a glaring oversight. Jonas calls this enterprise amnesia and cautioned that companies must stop trying to squeeze data out of the puzzle pieces. Instead, “the data must find the data and the relevance must find the user.”
Of course, IBM PartnerWorld exists to encourage midmarket CIOs to use its big data analytics, under the wing of Jonas, its resident genius and chief scientist. I do wonder, though, if midmarket companies aren’t being coaxed to apply the big-data-analytics square peg into a round hole. It’s rarely a case of pulling out a snazzy GUI to call up a magic answer, as much as departments outside IT would like to imagine.
Big data analytics requires some level of finesse, not to mention some intuitive leaps, to yield the gold in them thar hills. Are midmarket companies ready to take the dive? I suspect many midmarket companies’ big data analytics might be driven from departments outside the CIO’s control. But smart CIOs will have anticipated this movement and taken some proactive measures to insure that the right action is taken at the right time.