BETA
This is a BETA experience. You may opt-out by clicking here

More From Forbes

Edit Story

Going Beyond Big Data To Knowledge

Following
This article is more than 10 years old.

“Ipsa scientia potestas est” (“knowledge itself is power”), Sir Francis Bacon

Don’t believe the big data hype

Over the last month, I have written about all the hype surrounding big data including how it equals a big headache for executives and how the promise of big data bypasses the C-Suite. This final installment deals with the pending big data disillusionment which may result from the hype and failure of big data to deliver meaningful, strategic solutions for senior level executives.

The almost daily barrage of big data breakthroughs are usually advertising or PR executives extolling the latest hyper targeting opportunity from analysis of social media tweets, likes, online clicks or posts. The promised breakthroughs result from surveillance of digital actions, most of which have little or no relevance to any specific business. In the context of all the bits of digital data swirling the Internet, only a fraction is pertinent to any particular company. The conventional wisdom is to attempt to sort through the volumes of useless data to get to small or relevant data to create a model for more targeted online media buying. I guess one could argue that a model can be developed for just about anything. However the key question is, “Does the model really work in the real world?”

Questionable data bears questionable models

Obviously, the accuracy of the data is important for the validation of any model. Unfortunately, several important issues call into question the validity of the eventual small data used in media models. First is the issue of click fraud which has already been acknowledged by digital media buyers as costing U.S. marketers over $11.6 billion in wasted ad dollars this year. Second is the problem of sorting out fraudulent and useless data to extract the smaller set of useful data. And third is the issue of understanding the meaning and intent of digital data determined as useful. What is the intent of a “like”? Does it translate into marketplace behavior? Since consumer behavior is more complex than random digital actions, what type of unverified assumptions must be ascribed in place of fact in the model? A recent paper published in Sociological Science found that support for causes via social networks may only be “click deep.” One of the authors, UC San Diego’s Kevin Lewis is quoted as saying, “[Social media] seems to have failed to convert the initial act of joining into a more sustained set of behaviors.” "Likes" don’t necessarily translate into behavior.

It’s clear that most of the big data commotion gets down to media applications based upon questionable models. Maybe that is why a recent Gartner survey found that through 2015, 85% of Fortune 500’s will fail to exploit big data for competitive advantage.

From big data to knowledge

Data is the starting point and basic building block in a knowledge-based organization. Since the majority of big data uses today are machine-to-machine ad serving applications of “real-time” digital or internal data, knowledge isn’t required. Strategy requires a broader view of data. Strategy requires data that serves as fuel, but logic and experience still need to be applied to generate knowledge-based systems. Knowing not only what happened, but why it happened (diagnostic), what will happen (predictive) and how we can make it happen (prescriptive) is important for moving beyond big data to knowledge.

Prosper Insights & Analytics has been developing big data applications that integrate and analyze hundreds of consumer databases from sources internal and external to companies. These applications provide predictive and prescriptive knowledge which decision makers can use to define strategies, understand competitors, predict sales and better allocate media.

----------------------------------------------------

Gary Drenik is CEO of Prosper Insights & Analytics, a company that prides itself on turning data into solutions. www.ProsperDiscovery.com