NEW ORLEANS – A few weeks ago I wroteabout some of the similarities between Hewlett-Packard Co. (NYSE: HPQ) and IBMCorp. (NYSE: IBM) ahead of both vendors holding partner conferences inFebruary. HP doubled-down on its PC business two weeks ago in Las Vegaswhile, today in New Orleans, IBM reiterated selling its PC business some yearsago to Lenovo has turned IBM into a more profitable and nimble company. The twocompanies are on the same page in one major way though: the importance of big data.
One of the more interesting keynotes at HP wasdelivered by Autonomy founder Mike Lynch, who spoke about trends in socialmedia, big data, the cloud, governance and mobility and the growth ofunstructured data. He’s focused on finding ways to turn that unstructured datainto actionable business intelligence, which isn’t easy because unlike data ina relational database, unstructured data us a jumble of often contradictory information.Meaning needs to be derived based on context and other information, constantlyre-evaluated, and action taken in real-time.
At the time, I tweeted that it all reminded me ofthe challenge IBM faced with Watson, its Jeopardy-playing super computer.Watson needed to evaluate Jeopardy answers, which are often confusing puns,decide the meaning and then furnish the correct response.
I was reminded of Lynch and Autonomy this morningat IBM’s conference as Jeff Jonas, chief scientist (cool title) for IBM’sentity analytics group , spoke with partners about big data. It was athought-provoking presentation (with just a little more verve than Lynch’s)which raised many interesting points about the data deluge.
For example, Jonas noted that while data is growingand computers are getting faster, organizations are also getting dumber as they’reunable to act on that data. More data is a good thing though, he said. The moredata you get the better predictions you can make, as you can see where you’vegone wrong.
“Big data leads to new physics since more datalowers both false positives and false negatives,” said Jonas.
In a particularly chilling observation for privacyactivists, Jonas said “when” and “where” will become the most important methodsof data disambiguation. While someone can lie about their social security numberor name to try to game the system, lies that can be caught out as more data isgenerated to weed-out the false results, the same person can’t occupy twodifferent places at the same time.
Jonas said this will become key to data analytics.And the fact is we’re opting into it ourselves, with the desire forlocation-based services via our smartphones. I was told several years agolocation-based was poised to be the next big thing in social media, so I’m sureit will be any day now.
But as fascinating as the big picture presentationsby HP’s Lynch and IBM’s Jonas were, I was reminded of a conversation I had inLas Vegas with an HP business partner: this is all really cool but it’s futurestuff. I sell to small and medium-sized businesses, how is this relevant andmeaningful to me and my customers?
That’s the question that both IBM and HP are goingto have to answer. I can see the benefit for the large enterprise, but mostbusiness partners play a little further down in the market. Not only do youneed to define the use-case for big data and analytics in the SMB (and I dothink you can), you need to scale it down to a format and price level thatmakes it affordable and consumable for smaller businesses. That still seems tobe a work in progress for both vendors, and the messaging there has beenlacking so far.
Lynch mentioned upcoming Autonomy-based HPappliances that would scale enterprise search down into the SMB, while Jonaswas more muted on IBM’s SMB plans. In a similar area, SAP also recentlyannounced an SMB version of its HANA appliance for real-time analytics.
