Big data blinds: from 4Vs to 4Cs

“There’s no shortage of widely held myths around Big Data. Perhaps the most dangerous is that Big Data knows all and that it obviates the need for human judgment.” – Phil Simon (2016)  

WHAT are Big Data blind spots and why are they so dangerous? Marketing managers can easily be blinded by the promising 4 V features of Big Data: (1) volume, (2) velocity, (3) variety, and (4) veracity. In reality, these features come with their own corresponding 4 C challenges: (1) confirmation, (2) control, (3) communication, and (4) confidence. In an ideal world, Big Data analytics will provide previously unavailable real-time customer insights to improve decision making. “Big and Lean is Beautiful: A Conceptual Framework for Data-Based Learning in Marketing Management” explores the idea that Big Data actually provides decision-makers with tools for self-deception if they are not careful enough. To mitigate these dangerous effects, we have developed a lean start-up methodology.

Big Data’s 4Vs

SO WHAT? Could Big Data lead to biased learning because it is high in volume, variety, velocity, and veracity? Yes, as we demonstrated for Volume, Variety and Velocity in http://marketingandmetrics.com/combining-big-data-and-lean-startup-methods-for-business-model-evolution/

Veracity is the newest, and arguably most important, addition to the Vs. The truthfulness and accuracy of any dataset must be present for insights generated from the other 3Vs to be valuable in a professional context. Conversely, veracity can easily falsely validate these insights, making it the most dangerous V as well. The other features of Big Data have similarly troubling challenges, as shown in Table 1.

NOW WHAT? When employing Big Data in the decision-making process, managers make two major assumptions about the available data: (1) it is large, fast, multidimensional, and truthful and (2) it is highly representative of target settings. The above table shows the first assumption needs to be carefully validated. As to the second, managers can increase the match between learning and target settings by A/B testing, blindness to and falsification of hypotheses, regularly updating analyses, transparent presentation of statistical insights, and statistical literacy and risk intelligence. By combining the lean start-up methodology with Big Data-driven decision-making, we believe that marketing managers will be able to both leverage the benefits of Big Data and avoid the dangers that come from its misuse.

For the full paper: http://marketingandmetrics.com/category/big-data-business-models/

Advertisement