Productizing TV Viewership Data
Background
An audience measurement and consumer research company that offers audience data on TV, radio, newspaper, online, gaming and more. The data comes from a combination of offline reporting, in-home metering, and wearable portable device, collected across multiple markets and demographics. The resulting exposure data is then factored up to reflect a reasonable facsimile of viewership behaviors and locations.
Challenge
The company was able to track viewership down to 15 second intervals across both TV program and commercial data, but the national volume of this data across all TV networks became unwieldy and slow to interact with. This inhibited the company’s ability to offer products to their customers that would allow them to ask granular questions of the data with fast response times. These challenges were also reflected in the daily operations, where some offline customer generated requests took days to turn around.
Solution
Our team began receiving the raw, exposure level data from the company every week, factoring up the results as the company did. We then instituted a program that intelligently and dynamically separated the data into logical chunks, allowing queries to be performed far more efficiently. We built a platform designed to service a multi-tenant audience, who could break down the data across markets and demographics, across both TV programs and commercials, down to the 15 second level incredibly quickly. This allowed both the company and their customers to perform actions that previously took days in a matter of seconds, and gave the company a new source of revenue through the platform we built.