What the NBA can teach media buyers about data science
by Pat McCarthy, SVP, Marketing, AppNexus
In 1922, an AT&T-owned radio station in New York City, WEAF, executed what was arguably the world’s first radio ad, offering 10 minutes of airtime to any advertiser willing to pay $100. A Long Island real estate firm jumped at the chance, promptly running a radio spot extolling the virtues of a new apartment building being developed in Jackson Heights, Queens.
70 years later, in October of 1994, AT&T struck again, developing what many consider to be the world’s first online advertisement, a banner ad that ran on a website called hotwired.com (soon to be known as Wired Magazine). The ad – slightly humble by today’s standards – met with resounding success, receiving a 44% click-through-rate due to its sheer novelty.
In the 70 years between those two ads, advertising has come a long way. Newsprint made way to radio; radio made way for TV. But in those years of technological flux, there was one factor that held steady: the manner in which media got bought and sold. There was a ritualized process to it: the buyer and seller would meet face to face and thresh out the price of a given campaign for a given type of media. It was a Mad Men-inflected world where every deal was sealed with nods and handshakes – rather than facts.
Before the advent of big data and real-time marketing, marketing was a “creative pursuit” whose practice was far from scientific. Marketers didn’t always have the right facts at hand to ensure their ad campaigns were successful. As such, they needed to become absolute masters at face-to-face negotiations and deal making.
The ability to negotiate at the table will always be worth its weight in gold. But for marketers to truly succeed in an age where big data informs the moment-by-moment success of an ad campaign, new skillsets are required. In today’s marketing, real power stands firmly with those who can distill relevant insight from a fast-breaking stream of data points – and then use those insights to reach successful conclusions.
Increasingly, media buyers will need technology platforms that give them every opportunity to hone their real-time relationship to their audiences. Such platforms won’t simply help them understand to whom, when or where to deliver their advertisements. These platforms will increasingly allow them to know which ads to deliver.
How media buyers can learn better data science from the NBA
Fortunately, plenty of real-life models exist for media buyers to study and brush up on their data science skills. In fact, a media buyer needn’t look further than his or her television screen to see what I mean.
For those of you who follow professional basketball, you can’t help but notice the impact that data science has made on the way the game gets played. Ever since the introduction in 2013 of the STATS SportVu® camera system to each of the 30 NBA arenas, coaches, players and fans alike have been able to draw upon a wealth of new data that has, inarguably, improved how teams play basketball at the scientific level.
Here’s the part marketers really need to pay attention to: in just a few years’ time, the NBA has managed to build a data science technology platform that lets coaches and players quantify the intangible, meaning data points that were too granular to notice previously.
Using the STATS platform, NBA coaches and players can collect data points such as the total number of miles traveled by a given player during a given game and exact number of seconds each player has possession of the ball. Like the world’s most sophisticated ad-tech platforms, they can quickly calculate the where and the when – and provide teams with a means of determine what kind of customized strategies they need to adopt in order to win.
By aggregating these data points together, NBA teams suddenly have enough real-time insight to play at the level of data science. For instance, data science has established that it is generally more efficient for a player to try for a three-point jump shot than venture a few feet further and only be able to score two points.
Nearly everyone who watches ESPN knows about the phenomenon known as Stephen Curry and his record-breaking number of three pointers this season. But Curry’s genius at scoring three pointers is just the most obvious example of a new, League-wide approach to shooting. Let’s not forget the fact that the Houston Rockets (of AT&T Arena fame), a team that also happens to be one of the most data-driven teams in the NBA, achieved a record for the total number of three pointers in the 2014 – 2015 season, at least until this season, when the Warriors (led by Curry) eclipsed them.
While the NBA may have earned record-breaking total revenue of around $5.18B in the 2014/15 season, that figure pales in comparison to the revenue generated from digital advertising within the same timeframe. Digital advertising revenue rose to a staggering $25.7B by the end of the first half of 2015. Not even an NBA team led by Stephen Curry can keep pace with the speed of real-time, programmable ad serving. The only folks who can will be the data scientists. It’s high time marketers played like the NBA, adopting technology platforms able to build championship-worthy advertising in real time.