What Big Data Is Not... at Least Not Yet!

Faced by the spin and hype surrounding big data, a key for managers and executives wishing to execute a successful big data strategy is to sift the possible future realities from the possible future fictions.

The CIO's conundrum
A worthwhile and natural starting point for examining big data hype is Gartner’s 2012 Hype Cycle. This indicates the distance a technology is from becoming a common productivity tool. Gartner has depicted three tools it judges as being 10 years away from the plateau: Information valuation; the Internet of Things; and the semantic web. The first two being classified as "technology triggers," while the third being already located at the "peak of inflated expectations. " For CIOs grounded in realities of the technology life cycle, their concern is that colleagues on the business side will believe that the future is here now and demand it. They also worry that these business users will then become disillusioned when the promises made in the popular press fail to materialize, thus marginalize the deployment of big data technology as a whole and its long-term potential.

Information Valuation evaluated
First, Information Valuation is great in concept but currently extremely hard in reality to execute. The ability to value individual pieces of information accurately without a free market is challenging. It took the advent of marketplaces such as eBay to enable people to value the items in their basement and see that they were someone else’s treasure. Once the market had been established, pricing fell into line with normal supply demand economics. For information, this may not be so easy, as data becomes a new currency firms will parallel with the valuation of intellectual property assets such as patents. They will license it, ask for royalties, or perhaps even sell it. However, as in the case of patents, much of the data is unique, linked back to a particular source such as a customer, or only valuable when bundled with other data, and as such pricing remains problematic.

However, markets in the intellectual capital space have developed and are now about ten years old, and as such Gartner’s predictions for information valuation are perhaps on track. This knowledge thus allows executives and managers to accommodate the valuation of their data into their long-term planning cycle and thus avoid the disappointment of the short-term reality not living up to the hype.

The Internet of Things on the horizon
Second, the Internet of Things, a term born in 1999 and attributed to Kevin Aston, refers to tagging every object in the world with an IP addressable tag. Again, in theory this is possible as IPv6 allows 2128 objects to be tagged. Eventually this will likely happen, as IP addressable QR codes or tags can be applied to objects and tracked given enough resources. However, we are not there yet, and there are many hurdles, both technical and regulatory, to overcome before this is a reality. For example, if a retailer sells a customer a bottle of water with an identifiable code, and that empty bottle is thrown away and not recycled, it would be possible to trace that back to the individual and for a government agency to fine that individual. In such situations, data, its ownership and its oversight, will need qualifying by society as a whole. Otherwise, perhaps big data will become associated with Big Brother, rather than big business value.

The tangled Semantic Web
Thirdly, the Semantic Web, a common framework helping data to be more easily shared and reused, has been discussed in knowledge representation circles since the 1960s, with recent efforts, from 2001, led by Sir Tim Berners-Lee and W3C.org. However, while the theoretical base of a true semantic web is evolving slowly, there has been progress towards a standard for a Semantic Web Stack encompassing URI, UNICODE, XML, RDF, SPARQL, RDFS, and OWL development including terms of rule validity through RIF/SWRL. Even with this progress, critics point to its weaknesses such as the ability to manipulate the semantics around an object; the use by censors to limit or monitor informational content; and the increased loads on both humans and systems through the addition of a meta layer to the data. As such, in the near term, while big data is going to rely on XML and its associated technologies, it is not going to be utilizing a fully implemented version of the semantic web and its associated applications any time soon.

Overall, big data is here to stay -- it is going to become an increasingly central activity to all organizations wishing to remain competitive. However, executives need to be reminded that a key to embracing any new technology is to understand its limitations and not chase after some holy grail that may or may not exist; staying grounded in the art of the possible is a secure way of working towards maximizing the long-term ROI for any technology and delivering to the business on promises made.

Related posts:

• What’s Your Big Data Capability Maturity Level?
• Get Your Big Data Ducks in a Row
• Redefining MDM – an Open-Source Future
• Islands in the Stream of Extreme Data
• The Internet of Things: Making Big Data Bigger
• Hadoop: Executives & Managers Need to Read the Warning Labels

— Robert Plant, Associate Professor, School of Business Administration, University of Miami