This article was originally published in Advertising Age.
About two years ago Kevin Geraghty and his analytics team at 360i developed an intricate keyword matrix to help determine conversion likelihood and other results for a large sports retail client. Today, that matrix forms the basis of the digital agency’s predictive bidding platform and helps inform a third of the keyword bids the company makes on behalf of other clients.
Mr. Geraghty’s appreciation for the calculations and data structures led him into mathematical science and later to study in Ireland, where he got his master’s degree in management science from University College Dublin.
After working on analytics projects for clients such as GMAC Insurance and The Coca Cola Company, as well as Delta, American Airlines and British Airways, Geraghty moved on to 360i, where his job involves econometric modeling of paid media management, marketing-mix modeling and conversion optimization for large brands. He’s been with the agency for seven years.
In his Q&A with DataWorks, Mr. Geraghty speculates on the measurement potential for Facebook’s Graph Search and discusses a data structure with a deadly name.
Advertising Age: Facebook just launched its Graph Search product, which has tons of implications for marketers. What are the coolest Graph Search data opportunities?
Kevin Geraghty: Facebook’s Graph Search will create a number of new data opportunities for marketers. For one, search data will be used as another metric or success indicator. For example, searches on a brand or its posts could become an indicator of brand advocacy or interest, while searches on places will offer up a means of comparison among the Facebook presences of local competitors.
Ad Age: What’s a common misconception about big data you’d like to dispel?
Mr. Geraghty: Big data is not an end in itself. It is not a business strategy. It is, however, a huge opportunity for intelligent marketers to respond in real time to individual customer needs based on their behavior and move from an interruptive to a supportive model of marketing communication. We no longer have to wait for a weekly report and change our offer to target a generic demographic. We can interact with individuals and respond to their interests based on their declared and undeclared preferences.
Ad Age: What do you wish marketers would understand about what data scientists do?
Mr. Geraghty: Data science is not voodoo. We are not building fancy math models for their own sake. We are trying to listen to what the customer is telling us through their behavior. Data scientists are a pragmatic, practical and urgent version of data modelers. They get their hands dirtier with the data, are not tied to particular model formats such as statistical models or control-theory models, and they design solutions that keep pace with the breakneck speed of change in business process and customer behavior.
Ad Age: For laypeople, what do you mean by control-theory models?
Mr. Geraghty: Control theory is useful for making little and frequent adjustments, such as those needed to manage auction media for paid search or display. These contrast with more static statistical models developed for offline marketing problems such as direct mail. An analyst would have several weeks to produce a recommended mailing plan and once the mail went out the door would not make any adjustments on the fly, whereas the data scientist in the digital world has to develop “real-time response” strategies.
Ad Age: What’s the biggest problem with data science people as they navigate the world of marketing?
Mr. Geraghty: Many would not associate creativity with analytics, but creative problem solving is the core of analytics. The biggest problem with data-science people is that too many of them stop at the analysis and don’t push it to deliver the creative solution.
Ad Age: What’s the coolest or strangest type of data set you’ve ever worked with and why?
Mr. Geraghty: The guillotine is the coolest data structure in history. Simply put, it is a set of latency curves laid one upon the other to show how much demand, sales, claims, etc. will arrive at regular time periods as a result of a marketing program or other stimulus. I first came across it as a partial booking curve in the airline industry used to measure how fast a flight would fill up. I then tripped over it as an actuarial triangle in insurance to forecast how fast claims will come in. It has come in handy forecasting social-network growth in response to membership outreach tactics and paid-media performance. Once you get familiar with it, you see it everywhere.
Continue reading Kevin’s interview with Kate Kaye on AdAge.com.
