Case Studies Pixability Enhances Ad Performance with Snorkel Flow's NLP Capabilities

Edit This Case Study Record

	Pixability Enhances Ad Performance with Snorkel Flow's NLP Capabilities

Pixability Enhances Ad Performance with Snorkel Flow's NLP Capabilities

Technology Category	Analytics & Modeling - Machine Learning Analytics & Modeling - Natural Language Processing (NLP) Analytics & Modeling - Predictive Analytics
Applicable Industries	Software
Applicable Functions	Business Operation Sales & Marketing
Services	Software Design & Engineering Services System Integration
Challenge	The time-consuming process of manually labeling high-cardinality training data blocked Pixability from expanding their NLP capabilities. Read More
About Customer	Pixability is a leading YouTube & Connected TV Ad Platform that focuses on helping customers maximize their reach and optimize their video ad spend. The company is known for its advanced advertising solutions that leverage data and technology to deliver targeted and effective ad campaigns. With a strong presence in the digital advertising industry, Pixability serves a diverse range of clients, including major brands and agencies, by providing tools and insights to enhance their advertising strategies. The company's commitment to innovation and customer satisfaction has positioned it as a key player in the ad tech space. Read More
Solution	With Snorkel Flow, Pixability distilled knowledge from foundation models to build smaller, deployable classification models with more than 90% accuracy in just days. This approach significantly improved ad performance and brand-suitable targeting. By leveraging Snorkel Flow, Pixability was able to automate the labeling process, which previously required manual effort and was a major bottleneck. The solution involved creating a 600+ class multi-label NLP model that provided greater granularity and support for custom content categories. This allowed Pixability to offer more precise targeting options to their clients, enhancing the overall effectiveness of their ad campaigns. Read More Log in to view content
Contents

Technology Category

Analytics & Modeling - Machine Learning

Analytics & Modeling - Natural Language Processing (NLP)

Analytics & Modeling - Predictive Analytics

Applicable Industries

Software

Applicable Functions

Business Operation

Sales & Marketing

Services

Software Design & Engineering Services

System Integration

Challenge

The time-consuming process of manually labeling high-cardinality training data blocked Pixability from expanding their NLP capabilities.

About Customer

Pixability is a leading YouTube & Connected TV Ad Platform that focuses on helping customers maximize their reach and optimize their video ad spend. The company is known for its advanced advertising solutions that leverage data and technology to deliver targeted and effective ad campaigns. With a strong presence in the digital advertising industry, Pixability serves a diverse range of clients, including major brands and agencies, by providing tools and insights to enhance their advertising strategies. The company's commitment to innovation and customer satisfaction has positioned it as a key player in the ad tech space.

Solution

With Snorkel Flow, Pixability distilled knowledge from foundation models to build smaller, deployable classification models with more than 90% accuracy in just days. This approach significantly improved ad performance and brand-suitable targeting. By leveraging Snorkel Flow, Pixability was able to automate the labeling process, which previously required manual effort and was a major bottleneck. The solution involved creating a 600+ class multi-label NLP model that provided greater granularity and support for custom content categories. This allowed Pixability to offer more precise targeting options to their clients, enhancing the overall effectiveness of their ad campaigns.

Impact #1	500K programmatic labels sourced from FM responses and keyword analysis with zero ground truth.
Impact #2	600+ class multi-label NLP model that provides greater granularity and support for custom content categories.
Impact #3	Achieved 90% accuracy on a model with 26x more classes, demonstrating the scalability and effectiveness of the solution.

Impact #1

500K programmatic labels sourced from FM responses and keyword analysis with zero ground truth.

Impact #2

600+ class multi-label NLP model that provides greater granularity and support for custom content categories.

Impact #3

Achieved 90% accuracy on a model with 26x more classes, demonstrating the scalability and effectiveness of the solution.

Benefit #1	500k programmatic labels
Benefit #2	600+ class multi-label NLP model
Benefit #3	90% accuracy

Benefit #1

500k programmatic labels

Benefit #2

600+ class multi-label NLP model

Benefit #3

90% accuracy

Download PDF Version

Overview

Pixability Enhances Ad Performance with Snorkel Flow's NLP Capabilities

Operational Impact

Quantitative Benefit