IABM BaM Award

Arvato Systems has won the "IABM BaM Award" in the "Monetize" category at this year's International Broadcast Convention (IBC) in Amsterdam, in late September. The award was given for the newly developed AI-based forecasting tool, which can be used for targeted ad placement. In more detail, the forecasting tool is designed to analyze historical viewing trends to predict TV ratings. The victory of the Monetize category officially announces our new AI-based tool as a product or service which offers a significant user benefit for the broadcast industry. You can learn more about the reward in the BENET.

In today's world, there are two different markets when it comes to selling commercial slots to companies on live TV. In countries such as Germany, the different slots are price based on time, date, and show popularity. Based on those facts, a price is determined by the broadcaster, and advertisers can then decide if they want to go ahead with the purchase or not. In contrast, markets such as Poland, Netherlands or the Nordics don't focus on only offering particular commercial slots. Instead, companies work closely with broadcasters or advertising agencies to reach a defined number of viewers who are in the company's target customer audience. 

That is where our AI-Based Forecasting tool comes into play. The process of estimating the number of viewers, as well as their demographics, is a challenging and tedious task. Not only do viewer numbers change based on outside parameters (e.g., the weather), but those "commercial packages" need to be individually designed to match the companies defined target group. 

If a company wants to promote a new beer on TV to 100,000 men in the age between 21-35 the beer company is going to contact a broadcaster via his agency. The broadcaster needs to create a campaign based on their estimation of how many slots, at what times, and which shows the commercial needs to be played, to fulfill the requirements. From there, they are going to offer a set price to reach the goal. After the campaign is fulfilled, the beer company receives reports based on Nielsen's TV ratings/viewer of how many people actually saw the commercial. If the real outcome was below the promised numbers, the difference in percentage is to be deducted from the price, resulting in revenue decreases. In contrast, if the real outcome is above the promised goal, the beer company will not pay for this value and this potential revenue is lost.

To decrease tedious and lengthy manual work and to avoid revenue differences, and, more importantly, customer dissatisfaction, we developed an AI-based Forecasting tool. With the help of historical data of two years (2017 and 2018), the AI-tool identified patterns which can recognize parameter around timing and the program. In more detail, data about showtime, length of the program, day of the week and week of the year, or the kind of show, if it is a repeat, what season of a show it is, and even what episode number were collected. 

Thus, our forecasting tool does not only help to streamline lengthy and tedious processes, but it also helps to free up resources within the process of prediction. To predict TV ratings requires the studying of existing data combined with a lot of experience. Also, to successfully convert data and knowledge into forecasting requires gut feeling, which opens the door for misconceptions and mistakes. Real-life events and tragedies may always interfere with predictions, but our tool assists the broadcasters in reducing potential inaccuracy and misassumptions.

If you are interested to learn more, please feel free to visit our product site or contact Bernd.Goeztelmann.