MediaGamma

Converting data streams to revenue streams: #worldclassAI



In a data-rich world where AI is a ‘must-have’, we’re a world-class AI spin-out from University College London, one of the premier global centres of computing science academic excellence.

We help brands to find audiences and engage with them, and publishers to build audiences and monetise them.

Products

With our Audience Decision Engine and automatic machine learning pipelines, we answer the fundamental questions of the industry:

Checklist Icon
Pre-Bid Click Verification

We built our pre-bid fraud detection in-house because we felt more could be done to minimise budget wastage. We benchmark ourselves against the industry’s leading players and continuously come out on top.

Graph Icon
Click Through Rate / Cost per Install Estimation

It took our data scientists billions of data points and months of machine learning to develop our world class CTR / CPI estimation product. DSPs who use this product generally see on average a 30% reduction in eCPC / eCPI for clients.

Chart Icon
Bid Rate Optimisation

Why pay more for an impression in the exchange when you don’t need to? We have built the industry’s first bid rate optimisation tool which ensures you pay the right price for the right impression.

Suit Icon
Audience Lookalike Modeling

Our sophisticated machine learning algorithms can deliver desktop and mobile audience lookalikes with unrivalled scale at over 90% accuracy.

Group Icon
Audience Extensions

We provide publishers with the data science to maximise revenue whilst providing the tools to build their own data management platform (DMP) solution.

Gauge Icon
Floor Price Optimisation

Publishers can now maximise yield from every impression they sell with our industry first dynamic real-time floor price optimisation product. With the advent of header bidding, the opportunities to maximise yield are complex, yet vital.

About

In a data-rich world where AI is a ‘must-have’, we are world-class AI experts who help digital professionals understand and monetise digital user behaviour.

We help advertisers optimise against metrics such as Customer Acquisition Cost (CAC) and Lifetime Value (LTV), by finding audiences and engaging with them; and publishers to optimise revenue and yield, by building audiences and monetising them.

Our AI team is a spin-out from a premier global centre of computing science academic excellence, University College London.

We readily translate digital advertising challenges into data science solutions with tangible results, using the most advanced machine learning techniques such as predictive, self-learning algorithms that make decisions in real-time.

Get in touch to find out more.

Rael 2 S
Rael Cline

Co-Founder, CEO

LinkedIn or email

Jun 2 S
Dr. Jun Wang

Co-Founder, Chief Scientist

LinkedIn or email

Angus 2 S 2
Angus Bankes

Board Member

LinkedIn or email

Shuai 2 S
Dr. Shuai Yuan

VP of Data Science

LinkedIn or email

Tom
Thomas Marks

Senior Developer

LinkedIn or email

Me
Davide Vitelaru

DevOps Engineer

LinkedIn or email

Imageedit 2 4993425406
Andreas Gompos

Data Scientist

LinkedIn or email

Me Bw
Allan Brisbane

VP of Engineering

LinkedIn or email

Pasted Image At 2018 03 29 06 31 Pm
Victor Malachard

Executive Chairman

LinkedIn or email

Lab

We’re a world-class AI spin-out from a premier global centre of computing science academic excellence, University College London

We’re proud of our world-class credentials, consistently presenting our breakthrough research at prestigious international conferences and journals, and winning awards in the process.

Please do contact us - we'd be happy to see how we can work together.

Converting data streams to revenue streams: #worldclassAI

Bidding Algorithm
Unifying 3 functions (user response prediction, bid landscape forecasting and bidding strategy) to nearly double campaign ROI.

Look-a-Like Modelling
Using a novel "doc2vec" model to improve the accuracy of look-a-like models.

Deep Learning to Estimate User Clicks/Conversions
Area Under the Curve (AUC) metric to measure the CTR performance

Feedback Control of Real Time Bidding
Stabilising Effective Cost Per Click (eCPC) whilst increasing click through rate

Dynamic Pricing - Programmatic Guaranteed & Real Time Bidding (RTB)
Up to 21% revenue increase for publishers

Real Time Bidding Technical Monograph
Overview of datasets & research problems

Publisher Floor Price Optimisation in RTB
Increase in publisher revenue of 35%

Audience Lookalike Modelling
5.5% increase in CTR performance

Optimising Cost per Action (CPA)
Up to 32% ROI for advertisers

Display Advertising Options
Revenue increase of 40% for publishers