Just as we enter the upswing on the hype cycle for big data, the academics are stepping in and making it clear that in this case, there might be much ado about nothing.
According to Dr. Peter Fader, co-director of the Wharton School of Business and marketing professor at the University of Pennsylvania, there are some glaring problems with the way vendors and enterprise execs are framing the conversation around big data.
He doesn’t consider himself a data Luddite, and understands that captured information can generate more value. The argument is that too much information is being captured in order to answer questions that don’t require it.
In an interview this week, Dr. Fader poked holes into the big data practices, collecting masses of data to enable granular customer analysis. He says that while the concept itself has high potential (after all, it’s just data), it reminds him of another bubble that infiltrated enterprise IT no so long ago--customer relationship management (CRM). Like big data, he says, the CRM concept focused on collecting and analyzing transactional information, but failed to achieve its given goal.
“…ask anyone today what comes to mind when you say ‘CRM,’ and you'll hear ‘frustration,’ ‘disaster,’ ‘expensive,’ and ‘out of control.’ It turned out to be a great big IT wild-goose chase. And I'm afraid we're heading down the same road with Big Data.”
Fader is also skeptical of social media analytics being used to predict stock prices, calling the practice “ridiculous” even though a number of companies and even fellow academics have shown some success with their various models, especially since they’ve moved beyond the mere sentiment analysis typically associated with big-data-stock-magic ideas.
In this case he says that in a case like the social stock example, since the platforms don’t correlate individual transactions with social media posts, it affects their ability to make accurate predictions. Ultimately, the professor describes the desire to capture information from all sources as a ‘data fetish’.
An example was given about capturing mobile data from someone shopping in a store. Asking questions about how much data is required, the usefulness of capturing second-by-second location data, and how much value can be delivered by integrating that data with other information about the shopper.
Fader says that data scientists would ask these questions with a practical mindset on what information to collect and store. Data zealots on the other hand, would rather collect all the information because it might be valuable in the future. It’s what he calls the difference between old and new school analytics.
The professor made sure to note that predictive analytics are not a new field of study. Lester Wunderman made the phrase “direct marketing” popular in the 1960s using a only a few sources of information. He and other data scientists eventually discovered the recency, frequency and monetary value (RFM) model. By focusing on how recently customers made a transaction, along with how often they visit and the value of what they purchase, companies were able to draw important information about their clients.
Drawing comparisons to stock market ‘chartists’, people who only buy or sell based on stock price, Fader says big data people are pretty similar. Instead of asking why the data has changed, they would rather look for patterns within it. “In short, there is very little real science in what we call ‘data science,’ and that's a big problem.”
Ultimately, the professor believes the best way to understand customers is by asking the right questions from the correct information. Everything else is a flash in the pan.