Last week, DSSNYC took over Viacom Headquarters in Manhattan’s Time Square with a series of talks led by the top experts in data science. While the talks covered a variety of different topics ranging from tone-recognition with AI to customer retention, the entire event was centered around how data science practices can support, strengthen and grow the media and entertainment industry.
The million dollar question in the media landscape has always been: “How do we best satisfy our consumers?” At DSSNYC, the question became: “How can data science provide new and improved solutions for customer satisfaction?”
Anna Coenen, Data Scientist at The New York Times, said that having a solid idea of who one’s customer base is remains most important. “We can’t always personalize to everybody because we don’t know where they’re coming from and we aren’t using any third party services,” she explained. Further elaborating on this issue, Coenen divulged that the question of when and where to use algorithms is at front and center of The New York Times’ strategy, specifically pertaining to how personal it is optimal to go.
Ed Scott, CEO of ElectrifAI, would claim that the more brands zero in on their customers’ personal needs, the better. “If you don’t know your customers really intimately, you’re at a real risk for losing them,” he said. Scott’s knowledge of this issue is based on the structure of the media industry at large: “In the media, every capital structure is leveraged. The churn is an enormous problem.” Scott believes that the consistency of data retrieval can help brands combat attrition as seen in his experience with ElectrifAI: “Because we have so much data every two weeks, we can get a pretty accurate picture of who is going to churn,” he said.
There are numerous conceptual ways in which brands can better get to know their consumers and clients, but Lauren Lombardi, Senior Data Scientist at Nielsen, had a widely applicable strategy to share: “Instead of asking what the client wants, at Nielsen, we’ve started to ask why the client wants it,” she said. According to Lombardo, focusing on “the why” ensures that any successful deliverable can be replicated and tailored effectively to a client’s goals.
See talks like this in person at our next Data Science Salon: Applying AI and Machine Learning to Finance, Healthcare and Hospitality, in Miami.
Focusing on the deeper goals behind consumers’ requests and being transparent about how the provision of their individual data can help them achieve such measures is a sensitive but vital conversation happening between brands and consumers. While moderating a panel on the future of AI and ML in the media and entertainment landscape, Amy Yu, Vice President of Viacom, recognized that “the right to be forgotten doesn’t exist anymore now that data is everywhere.”
Many DSSNYC attendees came from small media teams, even those who were part of a larger blanket organization. That said, dispersing limited resources in the most efficient manner is a struggle that everyone relates to regardless of their company’s size. Speakers and panelists agreed that succeeding in this area requires brands to fully understand the inevitable trade-offs that a targeting decision can produce. “Data scientists must manage the trade-off between speed and accuracy,” said Lombardo. “A data science team can either focus on making the model three percent more accurate or three times faster.” Where a brand falls on the spectrum between speed and accuracy varies from organization to organization, but universally, brands need to understand the mix of priorities that produces the best results.
Ethics was a topic on the forefront of most panelists’ minds when debating how to allocate company resources. “It’s important to have robust processes in place that take data ethics into consideration when there are business needs at play,” explained Gilad Lotan, Head of Data Science at BuzzFeed. Coenen elaborated on this, stating that solving the most pivotal of issues affecting a company requires having a well-rounded approach. “If you start to dig deeper into a particular business problem, you may come across something that didn’t seem obvious immediately, but allows you to make a better algorithm,” she said.
All the conversations at DSSNYC implied that the main key to successful resource management is knowing which data is necessary to use and which is not. Lotan’s closing advice was geared towards saving brands money and time: “There are so many unintended consequences of storing data by default. Be rigorous about why you need to store different data.”
“For the bottom line of your business, if your products are used wrong, they won’t work right and your clients will think you don’t have a good product.”
Lombardo’s words proved a concise segue into a salon-wide discussion on product management. She stressed the importance of foreseeing and understanding products’ potential impacts throughout the entire building process. Joshua Miller, Head of Data Science at Samba TV, applied a strategy similar to Lombardo’s and attributes the success of his product to its ability to span across time: “Because we are able to leverage past, present and future data in our model, we get the precision very high,” he explained.
All in all, good product management is contingent on brands’ ability to understand and control the development of their products. Chris Whitely of Comcast Media presented a pivotal question for the audience to consider in relation to their own products: “If I build it, can I control it? If not maybe I shouldn’t be building it at all.”
DSSNYC 2019 culminated with the first ever DSS Elevate (DSSe) panel, featuring five powerful women in the data science industry. The panel united concepts of innovation and AI under the blanket of overarching themes such as mentorship, confidence and inclusion. Noemi Derzsy, Senior Inventive Scientist at AT&T Labs left attendees with an encouragement “never to be upset about constructive criticism.”
Curious for more?
Don’t miss the next Data Science Salon in Miami, September 10-11, 2019.