The rise and fall of Big Data

WM Circle Logo

By: RickFerguson |

Posted on October 23, 2017

The rise and fall of Big DataAbout a decade ago, the phrase “data is the new oil” swept the globe as the twin corporate power centers of IT and Marketing realized that, in many cases, their companies had more data than they knew what to do with – and the phrase “drinking from the fire hose” became a metaphor for the struggles many companies had with extracting actionable insight from data. About five years ago, the usual suspects in IT consulting and cloud-based analytics began to trumpet the phrase “Big Data” as a way to sell into companies hoping to extract value from the massive amounts of data at their disposal. Five years later, and “Big Data” has become a phrase seldom uttered in polite company. What happened to Big Data – and how can marketers best leverage the data at their disposal?

By Rick Ferguson

Most marketers may not know it, but the phrase “data is the new oil” originated in the customer loyalty space. The earliest use of the phrase is credited to Dunnhumby founder and Tesco Clubcard architect Clive Humby, who coined the phrase at the ANA Senior marketer’s summit in 2006. That pronouncement was swiftly copied and coopted by pundits from around the globe, which naturally led other pundits to adopt the phrase “Big Data” as a play on “Big Oil.”

As Slate’s Will Oremus points out, peak Big Data arguably occurred in 2012 with the publication of the New York Times article “The Age of Big Data,” because, as we all know, once the New York Times catches wind of a trend, that trend has already been around for a while. Since then, the phrase has become passé – in part, as Oremus notes, because we just call it “data” now, and in part because the rush to rely solely on data for business decision-making has often revealed the limitations of data-based decisions, as this money quote reveals:

“The haste to implement and apply big data, via what’s often called ‘data-driven decision-making,’ resulted in grievous mistakes. Some were blatant: There was the time Target sent coupons for baby items to the family of a teenage girl who hadn’t told anyone she was pregnant. Or the time Pinterest congratulated single women on their impending marriages. Or the Google Photos snafu, in which the company’s vaunted A.I. mistook black people for gorillas due to a lack of diversity in the data it was trained on. (It’s worth pointing out that, in this case at least, the ‘big data’ wasn’t quite big enough.)”

The problem with the phrase “Big Data,” as well as the phrase “data is the new oil,” is that these phrases fetishize data – leading marketers to believe that any problem, including the problem of building loyal customer relationships, can be solved through large-enough data sets and sophisticated-enough algorithms. Our over-confidence in these tools often hinders our ability to see the forest for the trees – and we often fall victim to what data scientist Shane Brennan calls the “Ten Fallacies of Data Science” which we will helpfully summarize here (all descriptions courtesy of Brennan):

  1. Assuming the data exists. Analytics work pre-supposes the availability of data that forms the basis for the work, but often the data is non-existent, lacking common identifiers, or aggregated at too high a level.
  2. Assuming the data is accessible. If the data exists, it is often inaccessible, or accessible only after the expense of considerable time and effort.
  3. Assuming the data is consistent. If the data is accessible, it is often not accessible in a self-consistent and well-defined format.
  4. Assuming the data is relevant. The required dataset may be out-of-date or not available at the level of granularity required for the analysis.
  5. Assuming the data is understandable. Some or all of the dataset may be indecipherable, and bad data may get ignored in the outputs or cause additional problems.
  6. Assuming the data can be processed. IT departments may not provide the tools necessary for the analysis, often due to corporate purchasing rules and security regulations.
  7. Assuming the analysis can be repeated. If marketing requires a data analysis to be repeated, doing so is often difficult or impossible due to data aging or the structure of the analysis.
  8. Assuming the data transfer is properly secured. When sharing analysis results within the company or with outside suppliers, the files are often not properly encrypted, resulting in security breaches.
  9. Assuming the analysis can be understood. Many data analysis outputs are poorly understood by marketers who, afraid to admit to their lack of understanding, make poor decisions as a result.
  10. Assuming your conclusions are unbiased. Many marketers enter into data analysis with a preconceived notion of the answers – and suffer from selection bias as they note data supporting their suppositions while ignoring data that doesn’t fit.

If your marketing or analytics teams falls victim to any one of these fallacies, then the result may be “garbage in/garbage out,” which leads to marketing campaigns or loyalty offers that are at best ineffective, and at worse actively harmful to your business. Is the answer then to eschew all data analysis in favor of, say, gut feel, astrological charts, or throwing darts at a board? Not at all. Companies all over the world use data analytics successfully to fuel personalization, construct relevant offers, and build differentiated experiences. As Oremus notes, it isn’t “Big Data” anymore – it’s just data, and our use of it has transformed marketing for the better.

Avoiding Brennan’s Ten Fallacies is, rather, about making the most efficient use of your analytical resources. By understanding the limits of data science, you can ensure that you focus your efforts on data that most correlates with current customer value, or is most predictive of future customer value. Call it “small data” or “customer-centric-data,” or just continue to call it “data” – if customer behavior shifts in a profitable direction because of your analysis, then you’ll know you’re on the right track. Be diligent, question your assumptions, and be aware of your biases. Big Data may be over, but data science, like any scientific pursuit, is forever.

Rick Ferguson is Editor in Chief of the Wise Marketer Group and is a Certified Loyalty Marketing Professional (CLMP).