Viral Video Predictive Analytics for Ad Tech startup

Customer: NDA
United States


Our customer is a Silicon Valley-based AdTech startup. They provide a global technology and data platform that maximises both performance efficiency and contextual suitability for advertisers on YouTube.


They were pursuing a very ambitious goal – to find videos which have a high probability of going viral (getting millions of views) and use them as a placeholder for video ads.


Zentadata's Serverless Engine and Data Studio were used to run data analytics on top of the Amazon S3 Storage. With the help of Data Studio, the customer’s Data Science team was able to implement feature engineering and predictive analytics over hundreds of millions of YouTube videos on a daily basis. The assembled data was used to build ML models and find videos with a high probability of going viral.


The customer was able to build a product with an exclusive competitive advantage – the ability to find unique viral videos and use them for product placement.
100 millions

YouTube videos analysed daily


Increased ad performance CTR/VTR

+10 customers

Paying customers acquired from Fortune 500 list