NASSCOM Community

Big Data Business Models – Deep Dive with the Master, Victor Mayer-Schonberger

Blog Post created by NASSCOM Community on Jun 27, 2016

“Good afternoon ladies and gentlemen, you are the masters and I am the class” was how Victor Schonberger delivered his powerful opening remarks. That, business is now replete with examples of huge applications of Big Data, is but only one perspective. Big Data is bigger. Way bigger!

So is it then a new tool? Like Python programming? Actually it isn’t.

We should leave this idea of tool behind to try and see it from a whole new perspective. This “new perspective,” it must be emphasized is incredibly powerful and it provides a glimpse into the world in a very different way, than what we are normally accustomed to. In process, decision-making can be substantially improved upon, to drive tremendous economic value.

What is at the Core of Big Data? 

Apple Pie (sic).

For years together, marketers would fret over the question: Which is America’s favourite pie? Conventional wisdom said, it was apple pie.

Why?

It was because supermarkets in America showed the highest sales of the 30 cm apple pie. But, if you asked around, most Americans would say, their favourite pie was cherry pie, apricot pie or any other, but hardly would one get the response - apple pie. A 30 cm apple pie is actually very big in size. Big enough to feed an entire family. For want of any other option, Americans were forced to buy the large apple pies.

Perhaps we asked the wrong question.

Instead, we should have asked, what is the right size of apple pie that supermarkets should store? Analytics revealed that the ideal size should be about 11 cm, so that people who actually liked apple pie could buy as per requirement, and not be compelled to buy in excess. Once this issue was addressed and pies were baked accordingly, sales shot up further. Prices came down too, as economies of scale kicked in.

Big Data helps us ask the right questions.

Victor also gave the example of Duolingo, a phone app that teaches new languages. Data patterns revealed that the Spanish were not able to learn English, as it was meant to be learnt through this app. They got down to fixing the problem, and today Duolingo has a special app for the Spanish to learn the English language. A customized solution to a unique problem. 

Walmart stores in America always see a spike in the sale of strawberry pop tarts, whenever there is a warning for hurricane. Why does this happen? Is it because the sugary content is a comforting thought at the time of extreme stress? The data does not tell us why. We don’t always need to know why and establish a causal effect relationship.

A certain experiment was carried out on premature babies to avert impending death. Besides the emergency medical support provided, the aim was also to see if Big Data could be used in this case, to save lives. Digital sensors were primed into action, which showed up to 1200 data points per baby per second. In process, an interesting pattern emerged. 24 hours before the onset of a major infection, the vital signs would stabilize. Almost like a lull before the storm. The doctors did not know why. But it was enough to tease out the value from this data to save lives. It was like a warning 24 hours in advance. 

Reusing Data to Generate Parallel Revenue Streams

Earlier, data collection was done with a purpose in mind. For instance, would red shoes sell? The data collected was used to validate this. Or not. That is what has changed with Big Data. The value of data is much bigger. The sum of all uses and re-uses of data is not restricted to one area alone.   

  • Inrix is a leading traffic intelligence provider. Its services are widely used by commuters to avail the best routes and avoid traffic jams. Obviously Inrix has a repository of gargantuan sized data sets, given the sheer volume. Some part of that data was used to draw out a correlation between traffic movement around shopping malls, and the revenue that was generated from the malls. Predictive Analytics came into play to predict future sales as well. A classic case of how data is being re-used to generate value for an altogether different business line. 

 

  • SWIFT is yet another example. The number of times and value of money transfers can be extracted to determine financial health, using Analytics. The Rolls Royce Jet engines generate 3 gigabytes of data per flight per engine. This data resides with Rolls Royce, which can later be used to negotiate better deals with the airlines. And, they have. Clearly, the Jet engine maker eked out an advantageous position. Yet another example of how data can be re-used to create value which was unforeseen earlier. A mobile company in the Netherlands, is able to detect signal changes at the time of disturbances in weather conditions. This data is bought by companies which predict weather, and used appropriately to give out weather forecasts.

So, the whole point is that data can be used to create entirely different revenue streams.

Look at Kindle for instance which is increasingly turning out to be the preference for readers. The data generated resides with Amazon which is then used profitably. This data does not automatically flow to authors or publishers. It has to be traded. Amazon takes advantage of this data to create parallel revenue streams. 

For small companies, the investment hurdle is now miniscule. With SMAC, companies need not invest in servers anymore, and can avail the pay-as-you-use model.

What it takes to be Successful in Big Data

 

  • Skills
  • Data
  • A mind-set which understands that data can be used to create immense and unimaginable value. It is about being iterative; transparent; and also to have a certain degree of humility. The I—know-it-all attitude can no more work, as data insights can show up completely different results which often trumps conventional wisdom.

 

Challenges 

There should be a high degree of trust in minds of consumers that the data being given out will not be misused. They are the source. Unless the trust angle is strictly established, there is a strong possibility that the source may dry up in future, as consumers may refuse to share data.

It is difficult to make traditional companies see the value of Big Data at present. There are others which may not be large enough to take risks or invest in data analytics themselves. They can however pair up with analytics companies to extract value from data, and work this out through a revenue sharing model.

Let us give Big Data the meaning and value it deserves. 

Outcomes