With the Hadoop World 2015 Summit kicking off today, now is a good time for IT services providers to take a step back and appreciate the end game most vendors have in mind when it comes to anything relating to Big Data.
While vendors such as IBM, Oracle, SAP, Microsoft, and Amazon Web Services are happy to sell the IT infrastructure needed to build a Big Data application, it’s becoming increasingly clear that most of these vendors would prefer to deliver those applications via the cloud as part of what has become known as Data-as-a-Service (DaaS).
Rather than simply making Big Data applications available in the cloud, most of the major vendors are working toward a model where they actually process analytics on behalf of the customer. A Data-as-a-Service model is designed to charge customers for continuous updates to multiple classes of analytics applications. Using application programming interfaces (APIs), the analysis created by those applications is then fed back into other applications to better inform any number of business processes.
Benefits of Data-as-a-Service
The upside for the customer is that they gain access to information that enables them to make better business decisions without having to invest in all the infrastructure and expertise normally required to obtain it.
It will be up to each individual customers to then decide how much they trust the analytics results being provided to automate a particular business process. Some will decide that they want business managers to retain control. Other will embrace a more prescriptive approach in which analytics will be used to automate an entire business process from end to end.
In effect, the rise of DaaS will meld IT services and traditional business consulting in a way that promises to deliver much higher levels of truly actionable business intelligence. Internal IT organizations are, of course, building their own Big Data projects, but the data science expertise required to turn those projects into useable results is in short supply.
More significantly, most businesses are not all that interested in owning the actual analytics. What they really want is the business outcome enabled by the analytics. It might, for example, be nice to own the algorithms used to determine the most optimal shipping route for any given product. But the more important thing is finding a way to slipstream the algorithms into a business process with the least amount of friction possible to enable the business to actually achieve that goal.
Big Data's dirty little secret
Of course, before any of that can happen, organizations need to have access to massive amounts of data to analyze. The dirty little secret of analytics all these years is that most of the results being presented were based on the aggregate of 10 percent or less of all the available data.
From a statistics perspective that’s enough information to make an educated guess. But if you want to maximize a business outcome every day then you need access to 90 percent or more of all the raw data. Naturally, that’s where Hadoop comes in to create the “data lake” that makes storing all that data economically feasible.
Opportunity for IT service providers
Put it all together, and it’s clear that Big Data analytics coupled with Hadoop, the cloud, and APIs are about to change the world we live in. The opportunity for IT service providers is to be among the first to figure out how to monetize those new capabilities in a way that can actually be consumed in a meaningful way by the business.