IBC 2019: Arvato Systems reveals AI Ad Forecasting
Arvato Systems (stand 3.B36) is delighted to announce the launch of its new addition to its BMS AdStore order management system for advertising at this year’s IBC that enables users to forecast ad ratings using artificial intelligence.
Ratings forecasts are critical for broadcasters in defining their prices and sales models. Advertisers want to know how and when they will hit their target GRPs or impressions in an increasingly complex environment as broadcasters look to offer cross platform campaigns to compete with newcomers that are disrupting the market.
There are a large number of factors that may affect ratings and the relative importance of these factors varies significant between peak and off-peak viewing. The new AI-driven analysis in BMS AdStore has been trained not only using such obvious factors as genre, ratings for the previous episode (if applicable) or the same timeslot a day, week or year before, but also unrelated, contextual data such as the weather.
“From the start of the project, we very quickly reached a point where the AI was matching the accuracy of the human predictions,” explains Bernd Götzelmann, Director Products BMS at Arvato Systems. “This included areas — especially particular weeks or time slots — where the error rate was higher both for machine and human. This became a focus point for the AI model to analyze what factors were driving the higher error rate which led to further models delivering even higher levels of accuracy.”
Achieving high levels of accuracy for human forecasters take years of practice and is an increasingly rare skillset throughout the industry. The advanced AI capabilities of BMS AdStore ensure that broadcasters from all over the world can tap into the expertise of the system. By using artificial intelligence to analyze historical data and viewing trends, BMS AdStore can now achieve highly accurate ratings forecasts for upcoming schedules which, when combined with the system’s optimizer, can result in more accurate placement of commercials and improved delivery of KPIs.
“Not only are we now providing more accurate forecasts, we are able to do so faster and to frequently rerun forecasts as necessary as critical external factors change,” concludes Götzelmann.