Lightning talk 1:
The Challenges of Engineering Data Science

with Sunny Manoleva – Data Scientist and Software Engineer at The Trade Desk and Rory Graves – Staff Software Engineer, AI Lab at The Trade Desk

All Data Science relies on Engineering to achieve its goals.  Engineering for Data Science is fundamentally different from traditional engineering.  In this talk we will take perspectives from an engineer working in data science and a data scientist working in engineering to give insight into this fun and challenging cross over area.

Lightning talk 2:
Supply-path optimisation: a Bayesian approach to take rate estimation
with Gianluca Campanella – Lead Staff Data Scientist at The Trade Desk

In the ever-evolving landscape of programmatic advertising, Data Science is pivotal in shaping effective pricing strategies for real-time bidding. This engaging talk delves into two crucial facets of this dynamic field: bid shading and supply path optimisation.
Bid shading involves strategically placing lower bids than the advertiser’s true valuation. This tactic aims to secure ad inventory at lower costs while preserving supply, capitalising on the cost savings linked with the second-price auction model, which has largely been discontinued in 2019.
Supply path optimisation is an approach to streamline the intricate ad supply chain. In an ecosystem with multiple suppliers offering access to the same ad impressions, advertisers need to navigate a web of options to determine the ‘best’ path for buying ad inventory. This decision hinges on factors like pricing transparency, ad quality, and overall performance.
This presentation will not only uncover the intricacies of pricing strategies but also highlight their profound impact on advertisers’ ROI and campaign success, enabling informed decisions at scale.

Lightning talk 3:
Forecasting at The Trade Desk: From Trillions of Potential Impressions Down to Just a Few Metrics
Ed Norton – Data Scientist at The Trade Desk and John Riddell – Data Scientist at The Trade Desk

“Forecasting is the hardest problem in advertising” Jeff Green (TTD CEO)

In this talk we’ll go through the challenges that forecasting presents and how we’ve overcome them at TTD. We’ll show how we’ve set up a system that produces forecasts in seconds and how we validate that our forecasts are trustworthy for our clients.

Panel discussion:
Data-Driven Decision Making for the Open Internet with The Trade Desk
with Gianluca Campanella – Lead Staff Data Scientist, Sunny Manoleva – Data Scientist, Rory Graves – Staff Software Engineer, John Riddell – Data Scientist and Ed Norton – Data Scientist, from The Trade Desk.

Supported by