(732) 400 5521

The Bicycle and the Jet Plane

By Catherine McClelland
Tuesday, April, 7 2015

The Analytics Misnomer

The Analytics Group recently gave a presentation at the New Jersey IIBA on the subject of big data and analytics technology. When I told a friend of mine (a developer with a sardonic sense of humor), he said:

“Analytics” is the practice of using faulty measurement methods to concoct dubious “marketing strategies” so you can collect more nebulous data on the returns of such a “strategy,” then rinse and repeat so you can keep your job, right?

My company has the word “analytics” in its name, so I won’t pretend that I don’t have a vested interest. That said, he and I differ on what analytics is. To answer his question, I’ll draw from part of the presentation that we gave to the IIBA.

A casual LinkedIn search pulls up 46,000 people in the Greater NYC area who describe themselves as “business analysts.” It seems likely that a few of them do data analysis as part of their job.

Facebook and Google, two of the richest companies in the world, make their money from targeted ads. Despite being an ardent ad-hater, I have clicked on ads that they’ve sent me, because the content matched something I was interested in.

Now, this is the same strategy that companies have always used: get the word out about their products, ideally to people who are interested in buying them. Companies have always collected data, then interpreted and applied it to make better business decisions. This interpretation is called analytics, whether you’re using nothing more than your own memory and instinct, or the latest data modeling software. Data is data, whether it’s entries written by quill pen in a ledger, or terabytes of information in a data warehouse.

The biggest misunderstandings about big data and data analytics are that these ideas aren’t new. Big data is just data that traditional methods cannot handle because of the size, quantity, type, or speed required to process it. We’re facing this transition now that our legacy technology is straining to keep up with large amounts of unstructured data. However, a similar transition occurred in the 1980s when Excel was invented. There was too much data of too many different types to efficiently process information in hard-copy paper form. Computers met a need to quickly and accurately perform computations. As the technology matured, software such as Excel was a solution for what the “big data” of the time.

Hiring managers at Litle, for instance, give prospective data scientists a research problem based on their own real-world data. The dataset, in spreadsheet form, contains 10 million cells, and the problem is designed to require several days of modeling and research to solve. It’s a problem that’s impossible to solve using traditional methods and nearly unfeasible without modern analytical software; the hiring managers provide only the raw data, but they expect their candidates to know the software available and give a report of which programs they used in the final presentation.

The bottom line is that data analytics is just the interpretation of data. It wasn’t invented yesterday. It’s not a marketing strategy. The transition we’re seeing is a revolution of the technology to cope with larger quantities and more types of data, at higher speeds. This has happened before. In my opinion, the technology itself is less important than the concept behind it.

I made an analogy in my presentation that the shift in analytics technology is like a bicycle to a jet plane. A bicycle, a horse and cart, and a car each achieve more or less the same result: linear travel. They’re improvements and variations on the same theme. They have inherent limitations—they’re slow and limited to the ground. A plane is also a form of travel, but the difference is that it can achieve results that were previously unthinkable because it’s a reimagining that goes beyond the limitations of earlier technologies.

Jet planes are expensive, and pilots take a lot of training. But you also can’t cross an ocean with a bicycle.