Much of the focus of Infochimps’ recent announcement that they had productized their internal Big Data platform focused on the “Big Data-Heroku” angle, which is understandable. While the array of open source assets for attacking data problems at scale is large, standing up a cluster can be intimidating for practitioners new to the field. Besides the unfamiliar querying mechanisms like MapReduce, there are a lot of moving parts involved. For those willing – to borrow Mårten Mickos’ metaphor – to trade their time to save money, the experience of building Big Data stacks from the ground up using tools like Chef and Puppet can be enormously rewarding. But as commercial open source vendors have proven satisfactorily over the past decade or so, there’s a substantial number of businesses who prefer to trade money to save time. Hence the attraction and interest in Big Data as a Platform.
But the wider significance of the Infochimps “pivot” may be that one of the original would be data marketplace vendors found the opportunity sufficiently wanting to reposition itself as a more traditional software play, productizing their big data experience. This is the clearest indication yet that data marketplaces may be the latest “Application Service Provider” cycle, as in right idea, wrong time. ASPs, remember, pioneered what later came to be called Software-as-a-Service, selling to a market who was not yet prepared to either software or consume it strictly through a browser. Most could not adapt, and either dried up and blew away or were acquired for their assets, their people or both. A decade later, and SaaS is not only an accepted application delivery model, it is the dominant one in many contexts. The ASPs mistake wasn’t product, it was timing.
Certainly the friction towards the marketing and sale of data as an asset is, at present, high. In comparing current market perceptions of data to enterprise acceptance of open source a decade ago, we’ve argued that mundane issues like licensing make data marketplaces at present largely inefficient, which in turn acts as a drag on adoption in a vicious cycle.
As much sense as data marketplaces make in a world that’s increasingly oriented towards evidence based decision making, it may be that the market just isn’t ready for marketplaces whose primary product is data. That their time will come is not, for me, a question. What’s not yet obvious is whether the time is now, or not yet. The Infochimps news is evidence of the latter, but the pluarl of anecdote is not data as they say. We’ll be watching the various data marketplace vendors, then, for signs of progress or regression. The cost of the latter may be high – much higher than is anticipated, in fact. But that’s a post for another time.
In the meantime, give Infochimps credit for making the necessary adjustments in terms of their approach and positioning. History is littered with the carcasses of startups that couldn’t adapt to being ahead of their time; with their new product, Infochimps is far less likely to be one those.