An Introduction to our Technology

Hundreds of hours of video are uploaded from millions of devices each minute. Connecting content creators and brands with audiences is critical for realizing the value of the social video ecosystem.

ZEFR

Our Deep Tech

ZEFR’s ability to capture and make sense of virtually every single video is what separates our capabilities from our competitors. This requires highly scalable and reliable systems for interfacing with multiple API platforms.

Search & Discovery Engine

Our proprietary recursive search engine traverses a larger search space than our competitors with variable depth logic, powered by machine learning. This branching search process discovers and filters videos based on potential value to content owners and advertisers.

Machine Learning - Natural Language Processing

Our NLP algorithm then evaluates the metadata around each video to determine the intent and content.

High Performance Learning

Each video and its enriched metadata are then classified, tagged and organized into a petabyte warehouse. Big data processing technology distributes the data clusters through Elastic MapReduce (EMR) and Spark technology, allowing for large data deployment and configuration at scale.

ZEFR’s VideoID Engine is the world’s first multi-platform video graph. The data is used in two primary applications: Digital Rights Management and Advertising.

BrandID®

ZEFR’s BrandID® technology creates a signal that aids in monetization.

Segmentation

Our learning engine optimized hundreds of millions of videos over 12,000 topical segments that complement traditional content classification systems as well as Google Knowledge Graphs.

Forecasting

Each video that relevantly matches to a segment is then run through a forecasting engine which uses statistical modeling to predict activity 30 days into the future with an accuracy of over 85%.

Brand Safety

High opportunity videos are then assessed by a proprietary brand safety engine that evaluates videos in over 200 languages.

ZEFR’s BrandID® product then launches vetted videos into both auction and reservation environments for activation in contextually targeted campaigns for top brands in key industry verticals.

RightsID® for Content Owners

ZEFR’s RightsID® platform functions across diverse social video ecosystems to organize and characterize intellectual property within sports, film, music and television.

Asset Management

The system combines Natural Language Processing with audio and visual recognition technology to sort and structure millions of assets each day. For example, within one week, the system will sort over 7 million music assets --recognizing that remixes, live recordings, and fan karaoke may be in the same IP.

Scorecard

Once the assets are organized, the ownership policies determined by the rights holder are translated into a rules engine the ensure the system upholds the intent of the rights holder’s digital rights plan.

Claimability Scoring

The system takes the data developed to score each video on the likelihood that it contains IP. The system then pairs this score with our NLP algorithm to allow the system to learn and grow with the ever-changing social video ecosystem.

BrandID + RightsID


Built on deep technology--these products combine as the most comprehensive social video identification and monetization tools available to brands and content owners.