Senior Product Manager, Brand Outcomes, Video and Native Advertising

About the team

The Ads Product team is responsible for monetizing Roku's platform through innovative advertising solutions across the Roku TV experience, including its premiere streaming service, The Roku Channel, on both Roku devices and off-platform environments. Collectively, these products reach millions of users daily, creating valuable advertising inventory and engagement opportunities throughout the entire Roku ecosystem.

About the role

Brand campaigns are bought on reach and judged after they end. We want them to work harder. The measurement is already in place, with viewer polls and standing holdouts. The tools that act on it are yours to build: an ad system that optimizes toward awareness, recall and lift while a campaign runs, proves the result with holdouts, and lets advertisers check the proof with their own tools.

You will own the products behind that across Roku's in-stream video and native ads. These campaigns are bought on impressions and judged on reach, awareness, recall, consideration and brand lift. You will build the cross-surface campaign experience, the models that optimize delivery toward lift, and the holdouts and third-party measurement that prove it. You will work closely with data science and engineering, and with the performance video product team on where brand ends and performance begins.

The goal is a brand campaign on Roku that works harder than it does anywhere else, and can show it.

For Massachusetts Only - The estimated annual base salary for this position is between $177,500 - $218,300.  Compensation packages are based on factors unique to each candidate, including but not limited to skill set, certifications, and specific geographical location.  This role is eligible for health insurance, equity awards, life insurance, disability benefits, parental leave, wellness benefits, and paid time off.

Problems you'll work on

We think brand advertising is the least optimized part of the ad stack, which is the opportunity. These are the problems we want you to take on.

• A brand campaign should know whether it is working before it ends. Most are planned on reach, paced evenly and judged weeks later by a survey or a marketing mix model, when nothing can be changed. Your job is to turn our polls and holdouts into signals delivery can use mid-flight: which households, which creative, which slot in the pod, and when to stop buying more of the same. When the advertiser's own panel or marketing mix model disagrees with our read, you find out why.
• Many brand impressions go to people who do not care, and we can tell. We ask viewers whether the ad they just saw was relevant to them. Answers swing widely between campaigns from the same brand and by where the ad sits in the pod, and much of the variation comes from the viewer, not the ad. What should the system do about it: suppress, reprice, rotate? Every suppressed impression is supply we do not sell, so you have to price what an irrelevant impression costs.
• Brand lift labels are scarce and expensive. Polls reach a small share of viewers, and every holdout household is a household the campaign does not reach. How small can a holdout be and still give a usable read? Can native engagement, which is dense and fast, give an earlier read on lift to train on? We think it might. You design the test that proves it or kills it.
• Creative is the biggest lever and brand campaigns barely use it. Creative, offer and targeting move together inside a campaign, so the data cannot say which one worked. Run two creatives against each other inside the buy, measure which is more relevant, and ship the winner while the campaign is live. Then ask whether what a household watches predicts which creative will land with it.
• Reach is a count of people. Advertisers want people who changed their minds. Counting a household once across in-stream video, native and screens beyond Roku, with consent, is hard. Optimizing and pricing on cost per incremental reach and lift per dollar is harder, and it is a product we get to define.
• The line between brand and performance is a measurement habit. Advertisers split budgets because their measurement splits. One campaign can build awareness and also cause the sign-up. If a brand buy can show both on the same holdout, should the budget move toward whatever the proof favors? You decide where brand ends and performance begins, with the performance video product team.

What you'll be doing

• Brand outcome products: Own the roadmap for products that let advertisers buy impressions across in-stream video and native and optimize toward unique reach, awareness, recall, consideration and brand lift. Define the metrics that connect impressions to outcomes, such as cost per incremental reach and lift per dollar.
• Cross-surface campaigns: Define how one campaign plans, paces and caps frequency across video and native, including deduplicated reach and shared creative, and how advertisers see the results in one report.
• Optimization models: Work with data science and engineering to define models for incremental reach, predicted lift, relevance and frequency, and build them into delivery, pacing, creative selection and bidding. The models train on brand lift and relevance survey responses and holdout results. Test whether native engagement signals give an earlier read on lift and improve the models.
• Brand lift and incrementality: Build holdouts and brand lift polls into campaigns by default, including experimental design and sample sizing.
• Third-party measurement: Own how Roku data works with third-party measurement partners, panel-based currency and marketing mix models, so advertisers can check our results with their own tools.
• Brand and performance: Work with the performance video product team so advertisers can see how brand exposure affects later conversions, and so budgets can move between brand and performance based on results.
• Working across teams: Work with engineering, data science, sales, research and operations, and make the products clear enough that sales and advertisers can explain them. Weigh the effect on video and native inventory, pricing, direct and programmatic buying, and the viewer experience.

We're excited if you have

• 5+ years of product management experience, including complex or platform-scale products
• Experience with brand or upper-funnel measurement, such as brand lift, reach and frequency, incrementality testing, marketing mix modeling or attention metrics, and you can read a lift study and question its sample size
• Experience building ad products that combine targeting, measurement and optimization, and a working understanding of how ML models are trained and evaluated
• Experience with ad products on more than one surface or format, such as video, native, display or feeds
• A background in digital advertising, media, marketplaces or data platforms; connected TV or premium video experience preferred
• Familiarity with third-party measurement partners and how agencies evaluate brand campaigns
• A record of shipping products in ambiguous areas and showing the results in numbers

#LI-JC5