Analytics and “Big Data” have become two of most pervasive words in the technology sector over the past several years. They are often associated so closely together it would make it appear that one is not desirable without the other. While it’s likely the case that if an organization possesses big data there is some value to be found utilyzing analytics there is a growing recognition that possessing big data is not a requirement for organizations to find value in analytics.
Small Data is a term currently gaining momentum in the analytics space. It’s definition is still evolving however there are three (sometimes contradicting) principles most definitions make mention of. The first is that Small Data is “the last mile of Big Data analytics” – it is the focused, actionable insights at the tail end of Big Data analytics. The second is that Small Data refers to the many small footprints we leave in the online world, from social media and online purchases to searching and browsing, which can be aggregated on a customer profile level. Finally Small Data refers to the size of the data set being analyzed. It’s the final principle that drives home the point that possessing a very large data set isn’t necessary to achieve value from analytics.
The reasons why very large data sets may not be available can be highly varied; from being a new or growing company to operating in a specialized sector or emerging market. In many of these cases the conundrum is that before an organization reaches the point where “Big Data” is available is when the analytical insights may be of the highest value – when the organization is attempting to gain a foothold. Combined with available external sources Small Data can be used to gain powerful operational and geo-demographic analytic insight among many others.
There is a great opportunity to help business grow using analytics on a smaller data sets. Low cost architectures (including Hadoop) are available to make Small Data analytics work while providing a scalable platform for the transition to Big Data in the future.