Is there such a thing as industrialized fortune telling? Yes, sort of! But it’s not at all related to any crystal balls.
Our interview partner for the Clicktale blog today is Eric Siegel, founder, Predictive Analytics World (PAW) and author, "Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die - Revised and Updated".
As Eric writes in his book, people look at him funny when he explains what he does for a living. But learning from data in order to predict the likelihood of events that haven’t happened yet is not only 100% scientifically valid, it is already common practice throughout businesses and organizations of all kinds.
Here I paraphrase Eric’s answers as he provided them to me during an extensive one-on-one conversation.
Q: Eric, how do you define predictive analytics?
A: Predictive analytics is technology that learns from data (which is an encoding of experience) to render predictions for each individual in order to drive operational decisions. This technology crunches numbers to do so, and the predictive scores it outputs positively influence outgoing actions and treatments, e.g. a recommendation, offer, or action. Machine learning is the academic or R&D term for the underlying technology. As for other terms, big data is just a (grammatically incorrect) way of saying “a lot of data.” By itself, the analysis of data doesn’t provide value – it must be acted upon, deployed. Predictive analytics is the commercial application of machine learning – the most actionable value you can get from data.
Unlike forecasting, which is the calculation of aggregate-level trends, predictive analytics generates predictions about individuals, e.g. the likelihood a specific website visitor will conduct a certain transaction, given their behavior online up to the moment of prediction. With that probability level in hand, a website can then make real-time decisions to improve the chances of winning their business. For example, they could decide to extend incentives for future repeat visits or purchases.
Q: You’ve been running the Predictive Analytics World conferences for over 7 years now. How far have predictive analytics come in this time?
There has been tremendous progress. Predictive analytics is used much more widely now, as seen in the many case studies coming out of the Predictive Analytics World conferences. For example, my book’s central table of mini-case studies has grown from 143 to 182 examples since the original edition three years ago. These come largely from presentations and sources I accessed at the PAW events – there are just too many to include them all.
The use cases have diversified as well. Targeting marketing, fraud detection and credit/risk scoring were the original business use cases. But today, examples span all industries, which take advantage of predictive analytics for positively influencing outcomes.
- Shell predicts oil refinery safety incidents.
- There are municipalities that are predicting which manhole will blow up.
- Fire departments predict the risk for each individual building.
- Healthcare institutions predict outcomes from treatments.
Because the use cases are so diverse, we now also have vertical editions of the Predictive Analytics World conference, namely specialized events for Financial, Healthcare, Manufacturing, Government, and Workforce applications.
Q: The use of digital behavior data as input into predictive analytics was still a leading edge use case a decade ago. Where does it stand today?
It’s commonplace at enterprises today. Web behavior data is included as input for many deployed predictive models. After all, behavior predicts behavior!
Conceptually, the training-data predictive analytics takes as input is a table where each row represents a unique entity (e.g. a customer) and the columns represent what’s known about that entity. Data from each individual’s online behavior makes that table wider in a sense. In other words, digital adds more behavioral flags and variables for the machine to crunch alongside other data, such as profile, demographic and transactional data sources.
Q: In what ways is predictive analytics used to improve customer experiences?
Many applications of predictive analytics serve to improve the customer experience – for example:
- Google search results – they aim to be based on your satisfaction.
- Facebook is working to maximize your engagement with your personalized newsfeed; you’re only going to keep browsing if you have a positive experience.
- It’s similar with AirBnB’s search results, which recommend accommodations more likely to interest you.
- Netflix predicts which movies you will like, i.e., by predicting which movie you would rate highly if you were to watch it.
Q: Where do you see predictive analytics going?
Data represents the experience that an organization has accrued, e.g., interactions with prospects as well as existing customers. Organizations are what make the society function and they are increasingly doing what they do based on better deploying predictive analytics.
It’s a predictive analytics world!
The Clicktale team would like to thank Eric for his leadership and contribution to advancing the world we live in by being a focal point for applications of predictive analytics.
On the lighter side, want to see the world’s coolest predictive analytics rap? Watch Eric, aka Dr. Data in the predictive analytics geek rap video now:
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. He is the author of the award-winning Predictive Analytics: The Power to Predict Who Will Click, Buy, Lie, or Die, a former Columbia University professor who used to sing to his students, and a renowned speaker, educator, and leader in the field. Eric 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. He and his book have been featured in Businessweek, CBS MoneyWatch, The Financial Times, Forbes, Forrester, Fortune, Harvard Business Review, The Huffington Post, The New York Review of Books, Newsweek, The Seattle Post-Intelligencer, The Wall Street Journal, The Washington Post, and WSJ MarketWatch.