Keith Pieper

What started as a simple consolidated service to manage retargeting buys for numerous Microsoft clients evolved into the Cadreon Trading Desk. I convinced an ad network (Turn) to white label and license their platform to Universal McCann to make retargeting management easier. We built on this with data target and private “first look” media deals. My role included everything from the primary value proposition to hiring the first few people and convincing management of the opportunity (ad networks took 10% but added little value). Then I worked with Turn to develop the features we wanted white-labeled and ultimately needed. I trained teams across the country and was the subject matter expert when educating clients, such as Chrysler, Microsoft, and Intel.

I hypothesized the types of data available (parameters, field names, IDs, source domains, etc.) and worked with data scientists to find these signals within the data set. When found, I analyzed how prevalent each signal was and the ease of transforming the data into a useful product.

I completed market research and analysis to uncover the various features available to paid search and social media advertisers in Google, Twitter and Facebook. I polled an informal group of target customers to better understand their needs. I compared the features to current capabilities for understanding our gaps. Each feature was bucketed by similarity (theme) and scored based on strategic importance, timeliness and revenue impact. I collaborated with engineering to understand the effort and cost of delivering each feature. Then I presented a roadmap of the features based on ROI and dependencies, in a general timeline of ASAP, Near Term, Long Term.

Household, cross-device streaming video reach and frequency measurements were highly fragmented. Nielsen is a dominant service for analog television measurement but applies similar methodologies to digital video, which is a very different ecosystem. I led a small team of software engineers to stand up an R&D lab for a better understanding of streaming video data across numerous apps and household devices, including a laptop, tablet, game console, and smartphone. Once we understood how to normalize and interpret the data, I worked with the team to formulate graphical models to represent real-time viewership and eliminate duplication for accurate reach measurement. We identified known Super Bowl advertisers and used the same in-home web data stream to identify visitation to advertiser websites. During the Super Bowl, users could see deduplicated, in-home reach of viewership alongside visitation to Super Bowl advertiser websites, providing a glimpse into the power of Super Bowl advertisements.