image credit: one photo/shutterstock.com
Now that we’ve established that maximizing yield is crucial for video advertising businesses, let’s break down what actually impacts yield—and what solutions exist to help media sellers maximize it. Yield has two levers: price and inventory allocation. Media yield management—and ultimately successful improvement in yield—requires a strong foundation in two key areas: forecasting and optimization.
If the word “forecasting” fills you with dread, you aren’t alone. Forecasting is complicated! There’s a reason some of today’s smartest minds are put to the test in the advertising industry. After all, in what other sector does inventory magically disappear after a given window?
Underestimating video ad impressions results in leaving money on the table, while overestimating video ad impressions results in a liability, wherein a seller must “make good” on the inventory that wasn’t viewed as promised. And yet, video forecasting is often performed in aggregate, using only a historical season average to inform and determine price across all associated ad inventory. Understandably, such a method is insufficient to ensure yield is truly optimized across a portfolio of clients, campaigns, products, platforms and sales channels.
Forecasting can be broken down into supply and demand, which together make up the foundation upon which inventory is allocated and price is determined.
- Supply forecasting: The supply of available ad inventory relative to campaign targets. The dynamics of supply forecasting in media are unique. It is difficult and often impossible for media sellers to simply increase capacity and produce more products for sale when demand increases. Moreover, the number of available impressions that will actually be viewed can’t be known in advance, nor is it controllable by the sellers who offer advertisers access to these impressions.
But there’s no denying the importance of accurate supply forecasting. It can be used to:
- Determine the number of impressions to guarantee to advertisers.
- Plan, price and allocate ad inventory for sale.
- Figure out which new programs to buy or produce (by forecasting estimated views).
- Decide when to schedule programs (by comparing estimated viewership on different days/times).
- Demand forecasting: Inventory that advertisers are expected to request, the price they would be willing to pay, and how the price might change as the flight date gets closer. A reliable, accurate demand forecast is an essential input to deciding whether to sign a given deal or risk leaving the inventory available for future opportunities. Ultimately, demand forecasting is about asking the question, “Is this the highest price I can receive for this inventory?” or “Would increased sell-out by lowering price increase my yield?”
Optimization is the second piece of the yield equation. Optimization isn’t just about finding the right inventory for a given client or campaign; it’s about finding the right inventory, at the right time, at the right price and presenting it to the right viewer. Proper optimization requires the dynamic allocation of inventory into buckets by campaign, as well as by schedule. To enable proper optimization, pricing must be fluid, ensuring every deal is optimized on its own, as well as within the larger business portfolio.
The challenge of portfolio-level management becomes greater as premium, targeted or advanced advertising products are introduced into the system, both in TV and digital. This makes TV advertising yield optimization increasingly difficult to execute with confidence—and nearly impossible with only Excel in your toolkit.
Enter artificial intelligence to save the day.
The only way media companies can contend with these challenges effectively and efficiently is with the help of technology. Today, media companies can use AI to leverage more datasets to better predict—and continually improve—forecasting and optimization accuracy.
When combined with data science techniques for more traditional forecasting models, for example, AI allows media companies to consider the impact of trends, seasonality, adjacency and competitive programming. As a result, AI can suppress the influence of anomalies in historical input data (to avoid bias) on a continual basis, increasing overall yield.
For a more detailed dive into forecasting, download our whitepaper, “The Role of Forecasting in Yield Optimization for Sellers of Television Advertising.”