By: Mélanie Roosen, L'ADN
Hilarious consultant, former professor, and rapper in his off hours, Eric Siegel shows us that data can be fun, and used wisely, quite effective.
With your videos, you managed to provide a very accessible view of data. This is quite interesting, considering data is usually considered a “cold” topic, interesting experts exclusively. How did you turn the geek stereotype into something cool and pop? What is your ambition behind that approach?
When we made the predictive analytics rap music video (www.PredictThis.org), the parody practically wrote itself. After all, the USA’s Chief Data Scientist designated – in a famous article – his own profession "the sexiest job of the 21st century." But aren’t firemen supposed to be the sexiest? That a geek is actually cool is nothing if not ironic.
I’ve always thought it was helpful and fun to explain a technical concept with a supposedly "cool" song. As a computer science professor at Columbia University around the year 2000, I sang educational songs to my students, such as a rock ballad about the angst of debugging your computer program.
Can you tell us more about your book, Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die?
The book reveals how predictive analytics works, and how it affects everyone every day. Rather than a "how to" for hands-on techies, the book serves lay readers and experts alike by covering new case studies and the latest state-of-the-art techniques.
I’m a former academic so this conceptually complete introduction to the topic serves as a textbook at over 35 universities. But it does read like a textbook – it is more in the accessible, entertaining "pop science" mode – accessible and relevant to any reader. OTOH, the final 3 chapters cover advanced topics of interest even to the experienced experts.
Amongst the many correlations you talk about in your book, which one is your favorite? The funniest?
Well, I like to lead with the link between ice cream consumption and shark attacks. As one increases, so does the other. Is this because eating ice cream makes a person taste better to a shark? Probably not. The more widely accepted explanation is that it is seasonal: When the weather is hot, more people swim and also more people eat ice cream.
When you find a connection in the data, it is only a correlation – a link that indeed helps predict – but it does not necessarily tell you anything conclusive about causation. When you try to answer "why" and find the explanation, you are seeking a causal explanation, which cannot necessarily be concluded definitively from the analysis itself. As they often say, "correlation does not entail causation."
Can you give us an example of a brand that has managed to use data to improve its business?
My book’s central table has 181 mini-case studies, so I’m not sure where to start. This includes examples from Airbnb, the BBC, Citibank, ConEd, Facebook, Ford, Google, the IRS, LinkedIn, Match.com, MTV, Netflix, PayPal, Pfizer, Spotify, Uber, UPS, and Wikipedia.
Retailers like Target and many major banks dramatically improve profit by more intelligently targeting their marketing with prediction.
And how about improving the well-being of its consumers?
Healthcare applications are growing rapidly – we’ve even launched an annual Predictive Analytics World event focused on this: the PAW Healthcare conference.
The applications include diagnosis, treatment optimization, hospital admission prediction, targeting compliance intervention (who’s not taking the meds they should be taking?), drug development, drug testing processes, and much more.
How can so-called traditional sectors, such as agriculture, use data efficiently?
Most major sectors are moving into predictive analytics, recognizing the value of optimizing mass-scale operations by way of predicting – for each individual – the likely outcome or behavior. Such predictions directly inform the decision or treatment taken with each individual person, corporate client, voter, automobile to be fixed, building to be inspected for fire risk, etc.
This is a big change to current processes. You can’t just crunch the numbers – you need to take the predictions output by the analytics and use them to drive better decisions. You have to act on them. This means a change to the current process.
Change always meets some resistance. But the value and tremendous results other organizations are achieving pushes this change forward. There is no stopping it.
To match these changes across sectors, we’ve been launching more and more industry-focused Predictive Analytics World events, including PAW Business, PAW Healthcare, PAW Workforce, PAW Manufacturing, PAW Government and PAW Financial Services.
In some sectors, such as health, insurance, or banking, the use of data can impact the relationship between the brand and the consumer. How should companies communicate about the topic?
The choice to not reveal your use of data – to hide what you’re doing – will backfire and only hurt trust. Transparency is critical.
But this must be done prudently. For example, when US retailer Target revealed they’re predicting who is pregnant in order to target marketing, they did it in a clumsy way that resulted in a PR snafu bar none. In my perception, they assumed the public would find it as purely positive as their internal audiences had. Read this article for more info.
In the end, consumers greatly benefit as well. Beyond improving a corporation’s efficiency and profit, the value of predictive analytics for consumers is unquestionable: less junk mail (and better for the environment), more relevant ads, better movie, music, and books recommendations, effective email spam filters (they depend on predictive models), better Google search results, more engaging Facebook feed ordering, more robust healthcare, and increased safety by more effectively targeting the inspection of buildings, manholes, etc.
What would be the greatest danger of data misuse?
How do we safely harness a predictive machine that can foresee job resignation, pregnancy, and crime?
I actually devoted an entire up-front chapter of my book – "Chapter 2: With Power Comes Responsibility: Hewlett-Packard, Target, the Cops, and the NSA Deduce Your Secrets" – to the issues in privacy and other civil liberties.
And how do we achieve value for law enforcement and national security without infringing on rights? See my op-ed in Newsweek on that.
Beyond all that, what if automated security screening discriminates by religion? This isn’t just a prejudicial mindset – it would be the systematic action of pre-judging based on a protected class.
This power/technology is like a knife: It can be used for good or for evil. It is valuable and powerful – that means it can be dangerous, but the idea of universally outlawing it is definitely not on the table.
Eric Siegel’s book: Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie or Die (John Wiley & Sons, 2016)
Eric Siegel singing and dancing: Geek Professor Drops Rap Video, Tries to Dance
About Eric Siegel:
Eric Siegel, Ph.D., founder of the Predictive Analytics World conference series and executive editor of The Predictive Analytics Times, makes the how and why of predictive analytics understandable and captivating. In addition to being the author of Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die, Eric is a former Columbia University professor who used to sing to his students, and a renowned speaker, educator, and leader in the field. He has appeared on Al Jazeera America, Bloomberg TV and Radio, Business News Network (Canada), Fox News, Israel National Radio, NPR Marketplace, Radio National (Australia), and TheStreet. Eric and his book have been featured in Businessweek, CBS MoneyWatch, Contagious Magazine, The European Business Review, The Financial Times, Forbes, Forrester, Fortune, Harvard Business Review, The Huffington Post, The New York Review of Books, Newsweek, Quartz, Salon, Scientific American, The Seattle Post-Intelligencer, The Wall Street Journal, The Washington Post, and WSJ MarketWatch. Follow him at @predictanalytic.