5 Ways Data Science Can Drive Higher Earnings Per Share (EPS)

Posted by Neil Schaffer on Sep 24, 2019 1:00:00 PM

Drive-Higher-EPSimage credit: Wright Studio/shutterstock.com 

In today’s media organizations, core business decisions are often made using Excel as the habitual tool of choice, overlooking the power of modern data science to improve outcomes. Big data can help you and your team forecast inventory more accurately to reduce makegoods, create data-backed rate cards that maximize value for each placement, and unearth insights from new, proactive reporting.

In fact, research shows that a lack of actionable data hurts your earnings per share (EPS) by leaving substantial revenue on the table. A University of Texas study found that the average Fortune 1000 company could increase revenue by $2B+ annually just by increasing data usability by 10%. Granting salespeople better access to data, along with increasing data quality, would increase return on equity (ROE) by another 16%.

At Furious, across several years and multiple clients, we’ve measured that by using data science effectively, media sales organizations can realize average revenue increases of nearly 10%. Furious recently released a white paper, “Spinning Data into Gold: How Data Science Improves Earnings per Share for Media Sellers,” to help CFOs and other leaders maximize the power of data in their organizations. Think about how you could use data science in the five ways below to help drive EPS:

1. Provide more detailed, granular reporting that prescribes action.

When data resides in separate places and is analyzed manually, it is nearly impossible to analyze business metrics across an entire portfolio, while still optimizing for each individual sales channel, inventory type, and advertising partner. A software platform that makes use of data science can greatly improve business outcomes by bringing data together from across your organization and enabling more advanced, forward-looking, holistic calculations.

Furious has created new data-driven metrics to help clients prescribe action, including:

  • Opportunity gauge: A forward-looking report that shows inventory and/or networks you’re not utilizing effectively. By aggregating and normalizing both delivery and measurement data, Furious has been able to identify programs, dayparts, and networks that were underutilized for specific campaigns, freeing up premium inventory across customer portfolios.

  • Rate Card Adherence: A measure of how closely individual sellers (or sales teams) adhere to published rate cards when closing deals. It’s an easy way to take the pulse of your sales organizations and see if there are any changes that need to be made to individual deal pricing, or to the rate card itself.

2. Give a better line-of-sight to tomorrow with more accurate demand, supply, and revenue forecasting.

Incorporating AI allows you to manipulate data in real-time and to incorporate more advanced mathematics (e.g., time-series forecasting) to ultimately arrive at more accurate forecasting.

Take the example of demand forecasting. Today, decisions are often based on information captured in a customer relationship management (CRM) system, such as Salesforce. But the data in a CRM is input by sales people on what they anticipate happening—such as a deal that’s anticipated to close at a certain price—rather than what actually takes place. Instead, with data science, you have insight into many predictive variables, like seasonality, changes in programming, and month-to-month comparisons to supplement and empower human judgment. The result is a much more accurate way to predict demand and price inventory accordingly.


3. Optimize pricing.

Pricing decisions need to be based on more than just relationship-driven sales deals. The goal: dynamic rate recommendations or decisions, which could be embodied as weekly or even daily rate cards. They would be made up of a regularly updated set of price rate recommendations driven by current data, continuously improved by machine learning algorithms. The eventual result is the disappearance of a “standard annual or quarterly rate card” entirely and the introduction of the use of true dynamic pricing of each unit and each campaign order.

Pricing optimization also comes into play in setting prices for upfront deals, where significant money may be left on the table right now.

4. Greater allocation efficiency and inventory utilization.

Operators are faced with the challenge of efficiently allocating inventory between different types, such as national vs. local spot sales. You need the power of data science to inform the right mix to use, and then, importantly, to allow that to change dynamically over time.

This need becomes even more pressing with a larger mix of media channels and with newer ad formats. Broadcasters and cable networks, especially those with digital assets, now have to worry about cross-channel cannibalization. Also, with advanced ad formats like OTT, you can generally command higher CPMs to satisfy asks for specific, layered audience demographics, but there’s a limited supply of views for any given narrow group. Understanding which inventory is irreplaceable (i.e., what is the opportunity cost of giving away certain spots) cannot be solved quickly or accurately without data science.

5. Reduce makegoods and rework.

TV sellers are still left guessing how much viewership a given program or campaign will receive. Audience estimates via surveys remain standard practice. As a result, sellers are often forced to issue makegoods for audiences that an advertiser was guaranteed but couldn’t be fully delivered. While makegoods aren’t avoidable entirely, they take up large amounts of salesperson time and company treasure and should be minimized as much as possible. Data science can significantly improve viewership forecasting—and thus reduce the frequency of makegoods.

Additionally, sellers often input a deal but have no certainty that an entire order will actually be delivered on or if it will be preempted by a more attractive price. Data science can help. Typically, the winning sale is determined by a simple “rate rule,” which simply awards the placement to the highest number. AI can determine a certain value for inventory and will only bump a sold spot if another bid comes in 30% higher, for example. Additionally, a seller who loses a bid can be given a full view of all other inventory that meets their audience demo requirements to choose from, in the event their deal doesn’t ultimately go through.

To learn more about how data can help your organization, download "Spinning Data into Gold."

spinning-data-into-gold-white-paper-download

Topics: Data Analytics, Data Science