Big data and the dot-com deja vu
People are supposed to learn from their mistakes and if Mitt Romney is right, corporations are people too my friend. So what has the current cohort of corporations vying for a position in the big data arena learned from the dot-com crash a mere decade or so ago? We don't know yet, but there are signs. Here is a small one, followed by a big one.
One thing they haven't learned based on the incredible sense of deja vu I have experienced since leaving the telecom world for the bright lights of big data is that tagging your brilliant startup with an odd-ball name that includes random capitalizations and phonetic shortcuts is neither a sign of creativity, nor of differentiation. It just frustrates journalists and their fact checkers, and plays havoc with automated spell checkers, especially on Apple devices. Without passing judgment on the companies themselves, here are a few examples of relatively new company names that work and that WerkLessSo. Cloudera? Nice. Tableau? Smooth. Palantir? Elvishly good. But WibiData, Axemblr, Qunb, and FeedZai are like the ITC^DeltaCom of the telecom CLEC era.
Ultimately though, names are unimportant in the short term and if other similarities to the dot-com or telecom era hold true, the short term is all many companies in this space will experience. The short term is all many of them want or expect to experience. The goal of many an entrepreneur will be to build a high-demand, big data solution faster and better than anyone else, and catch the eye of one of the large players in the industry, get acquired, buy their Cessna and start all over again. Nothing wrong with that.
In the meantime, players of all sizes are scrambling to form the most relevant partnerships through which they can get to market with full and integrated solutions. It is reminiscent of the System Integrator led approach to consolidation, automation and competition in the early competitive telecom space. And the money beginning to flow into this space to fund innovation is also strikingly familiar.
As heady as the 90s were in telecom and the Internet, and as much as big data relies on the advances begun then, the kinds of problems big data analytics has the potential to solve for the world as a whole could make this a far more exciting space. Companies were innovating then, but they also struggled to enable the new world of broadband because they were simultaneously transforming systems that were not designed to support it. And therein lies the lesson companies need to learn--and in many cases, to take to heart. This is the big sign.
As the competitive telecom market was learning to compete, the vendors serving that community were busy competing as well. This created an atmosphere that stifled the very thing that could make telecom service providers most successful, the very thing they begged their vendors to provide for the last 15 years: Open, non-proprietary, integrateable solutions. Vendors stonewalled for years and continue to stonewall in making concessions on the use of proprietary hardware, software and interfaces in order to protect their market share. Some of these vendors are now serving the big data space today.
It is a really good sign for this market that someone like Doug Cutting, creator of Hadoop and chief architect at Cloudera, can say as he did this week at IT World that, "I was somewhat surprised to see that over the last few years … I expected there to be some proprietary competitor, that one of these big IT vendors would come up with their own solution and say no, this is the way you need to do things. But it hasn't happened. All of the players have said we're going to standardize on Hadoop."
That may not hold true forever, and it is good when other ideas are at least proposed, but it also is encouraging to think that perhaps people and corporations do learn their lessons, and that one of the lessons from the 90s is that being proprietary, like being protectionist, doesn't always work in the long run. - Tim