Tracking audiences’ responsiveness to promotions using machine learning
The first look of a show so eye-catching, it intrigues you. A promo so gripping that all you can do is book a date at 8, with your television set. Enchanting, enticing and alluring the audiences into their world, welcome to the world of promotions. Only a few promos create that lasting impact, leaving no choice to the audiences but to tune in. Can you create a winning strategy that will help you tap the right market?
Currently, there are solutions, focused in understanding the overall effectiveness of promotion campaigns, but it that enough? Survey agencies are using viewer level data to find out the conversion of various promo campaigns. It fails to understand the factors that affected the conversion and the degree of effect of each factor at a granular detail.
The effectiveness of promotions can be gauged by disintegrating it at different levels taking into account an exhaustive list of factors. What if you were to know that for an upcoming show, a Rom-Com targeted towards younger population of tier-one cities, placing it on a celebrity talk show on a sister channel, during its prime time run is definitely going to make it buzz-worthy and pull in newer audiences.
The factors that affect the success of promotion can be broadly categorised into – Content and Placement. Placement depends on factors like the day part, channel, type of show and segment it airs in. The type of promo based on its content could be classified as plot-centric, character-centric, based on conflict resolution or it could be episodic, thematic or umbrella. How we use these factors to determine their effect on the reach of the show decides how well the strategy works.
When Machine learning and Big Data technologies come into the picture, they enable us to create a process that learns on its own and keeps improving over each iteration, run over large volumes of data. With the launch of technologies like Google Cloud Vision, Google Speech to text API and Parts of Speech tagger for Hindi language, it is possible now to objectively capture the content of a promo and quantify it. Advanced audio watermarking technology embedded BAR-O-meters in set top boxes capture the household level viewership data related to promo placement, audience response and conversion. Once this knowledge of content is combined with the audiences’ response and mapped to demographic data, we have struck gold.
Audience behaviour is constantly evolving, the process of capturing and tracking the changing behaviour, over a long period of time across multiple campaigns and a large viewer base with the help of machine learning and big data technologies is the future of promotion strategy. AthenasOwl Smart Analytics can leverage cutting edge data science capabilities to help you devise a promotion strategy.