Launched BrandConnect @ YouTube

TL;DR As lead PM, creator experience, I shipped the automated end-to-end version of BrandConnect, and doubled acceptance rates

Research: Creators need to be paid fairly to do sponsored content deals for brands that they trust. Brands want to reach creators quickly to launch their campaigns with one place to measure performance.

Design: An influencer marketing product that gives creators the ability to set a minimum price, share their offers with their manager, and feel confident to work with brands trusted by Google to run ads.

Go-to-market: Five launches in five markets: US, IN, ID, BR, UK.

Impact: Launched in five countries, doubled creator acceptance rates. Creator average payouts can be multiples higher than annual ads revshare

TV-to-Digital-to-Offline Attribution

TL;DR Built an O2O attribution product that measured 4X Conversion via sequential targeting post TV Exposure

Research: Partnering with a top Insurance advertiser, I tested a hypothesis that shifting TV budget to digital with a sequential message would result in higher conversion rates compared to a 20% placebo.

Design: Utilizing a device graph & an offline sales file, I associated exposed households found in fingerprinted TV commercial data from the client with households reached via digital campaigns and provided a cross channel performance report utilizing last-touch attribution.

Impact: Presented top performing TV & digital creatives resulting in larger media planning budget

Attribution Data Lake

TL;DR I built the first iterations of the Viant Data Lake Product

Research: Recognized opportunity to provide “white-label” data lake for Agencies, fulfilling two critical gaps in their offerings: Persistent device graph for targeting and a data lake “sandbox” for ingesting, organizing and experimenting with attribution models.

  • Agency + Tier II Automotive
  • Agency Technology Arm(s)
  • Advertiser Direct

    • Impact: Increased media budgets

    Patent: Online-Offline (O2O) Attribution

    US Patent US20160019603A1 ‘Attributing offline conversions to online activity’

    Research: Design for linking digital exposure data to actions taken in the physical world on a mobile device that interacts with low-powered beacon technology

    Build: Hack day demo at national sales conference illustrated which hotel locations (read: bars) were more popular with segments of our company: salespersons vs engineers. We won first prize.

    Impact: Limited adoption, but the beacons will have their day!

    Startup: Data Pipeline

    TL;DR I built a Data Pipeline for Disney Mobile, a ~60 headcount startup within Disney Interactive

    Background: Disney Mobile recruited me during pre-launch startup mode. My responsibilities were to create a data pipeline stack to acquire and merge Guest data generated by the Managed Virtual Network Operator (MVNO) vendors into Disney’s owned and operated data stores:

    Constraints: As a Managed Virtual Network Operator, or "MVNO," Disney Mobile relied on external vendors for almost 100% of business function (renting the ‘BSS’ of OSS/BSS)

    Impact: All of the company data and transactions were locked in vendors internal systems. It was my job to load this data into Disney, transform it for reliable reporting and provide flash reports to executive management

    In-Game Analytics & Optimization

    TL;DR I made the popular 'Toontown' in-game log files accessible for business analysts at Disney.com

    Challenge: The producers of Toontown Online, a massively multiplayer online role-playing game, created monthly quests such as in-game scavenger hunts to increase user engagement and ultimately lower subscription churn. However as typical with startup environments, there was no infrastructure in place, just logs and a spare LINUX instance.

    Example: Using only in-game server logs, determine which user(s) completed a scavenger hunt in the correct order and create ad-hoc leaderboards of the usernames that complete quests defined by the Toontown lead Producer.

    Impact: Designed an operational data store so that the activity could be transitioned to a business analyst within 90 days

    Supply Chain: Starbucks DW Consolidation

    TL;DR Starbucks Data Warehouse Doubled In Size

    Challenge: After acquiring Seattle’s Best Coffee (SBC), Starbucks merged its supply chain operations to pick/pack/ship and bill for whole bean and ground coffee from one transactional system. As a Data Warehouse PM I was challenged along with the DW database and application leads to merge accounts for seamless reporting post go-live and SBC deprecation.

    Impact/Key Learnings: You can never have enough test cases, you can never trust test data will look like production: Data warehouse consolidations in tandem with transactional system migrations are hard.

    Supply Chain: SAP to DW Integration

    TL;DR MicroStrategy saved the day

    Challenge: After implementing SAP Financial, NBC/U elected to replace a legacy supply chain system designed for records and CDs with a new instance of SAP for supply chain

    Impact/Key Learnings: You can never have enough test cases, you can never trust test data will look like production: Just like I learned at Starbucks, Data warehouse consolidations in tandem with transactional system migrations are hard–and sometimes the application layer can save the day