Job Type
Full-time
Description
Position OverviewMarketing depends on identifying the audience correctly. Deep Sync supplies the identity data that makes that possible, across offline and online channels, in both consumer and business marketing.
You will own the product surface on which a customer selects an audience, purchases it, and delivers it into the channel where the campaign runs, built on the platform now under development. The role also covers new product development: taking an idea from a hypothesis to a paid offering, and discontinuing the ones that do not justify further investment.
The products we build have to be ready for an AI first world. Customers increasingly reach data through agents and natural language rather than through a screen, which means what we ship has to serve a machine as readily as a person.
The position calls for a senior product manager fluent in advertising and in data, and equally at ease with AI. You should already understand what a match rate is worth to a buyer, the limits of lookalike modelling, and why one individual appears as separate records in separate systems, you should be able to go into the data and verify such questions yourself, and you should be building with current AI tools rather than reading about them. We run a flat organisation and a fast product development cadence. The position is based in Kirkland, WA.
Key ResponsibilitiesProduct ownership and delivery- Surface ownership. Own an area of the product from the problem through to the outcome: the case for doing the work, the requirements engineering builds from, and the usage it shows six months after release.
- Requirements and written decisions. We work in documents. A requirements document states the problem, the decision taken, what is in scope and what is not, and how success will be measured. Presentations are for reporting, not for deciding.
- Scope discipline. A rebuild invites reproducing every feature that exists today. Establish what is actually used, make the case for retiring the rest, and hold that position in review.
- Release management. Select the customers who receive each release first, define the test that closes a milestone, and report the result as it stands.
- New product development
- Opportunity identification. Attributes that have never been packaged, audiences that have never been priced, and questions customers are currently paying somebody else to answer.
- Early validation. A test with two or three customers, a measure agreed in advance, and a short written finding. Evidence before a long build rather than after it.
- Applied AI. Determine where a model or an agent belongs in the product and where it does not, specify how its output will be evaluated, and demonstrate the idea with a working prototype rather than a description of one. Design for consumption by machines as well as by people: predictable interfaces, described data, and output that a system downstream can act on without a person in between.
- Progression or discontinuation. Both are acceptable outcomes. An idea that remains a proposal for three quarters is not.
Market and commercial- Customer contact. The sales, delivery and client facing teams hold most of the knowledge about what fails, and customers hold the rest. Time spent with them is part of the role.
- Consumer and business marketing. The two buy different things for different reasons, and both are Deep Sync customers. Understand what each is measured on.
- Revenue alignment. Which audiences sell, which attributes do not, and what a customer pays for rather than merely tolerates. Work with product marketing on how the result is presented to the market.
- Delivery destinations. The Trade Desk, DV360, Meta, Amazon, TikTok, LiveRamp, the Snowflake and Databricks marketplaces, and the offline channels. Determine which integrations are worth maintaining, and what each has to do before a customer regards it as complete.
Data- Verification. Match rates, coverage by attribute, record counts, and the cost of a source against the revenue attributed to it. No analyst is attached to this role, so write the query yourself.
- Working with the data owners. Identity resolution, the attribute taxonomy and modelling belong to other teams. You are their customer, and you specify what the market requires of them.
Requirements
- AdTech or MarTech product experience. Required. You have built and released product within an advertising or marketing technology business: audience data, a demand or supply side platform, a customer data platform, an identity or onboarding provider, a clean room, or measurement. Adjacent experience is not a substitute.
- Understanding of identity data, online and offline. Deterministic and probabilistic matching, first party data onboarding, the difference between reaching a person through a mail file and through a bid request, the effect of an identifier being withdrawn, and the constraints privacy regulation places on all of it.
- Exposure to both consumer and business marketing, or a demonstrated understanding of how the second differs from the first.
- Sound judgement at speed. The cadence here is fast, which is not a licence to guess. Assess the options, state the trade off, choose, and be able to explain what would change the decision. A decision taken in a day on seventy percent of the information, and revised when the remainder arrives, is worth more than a month spent reaching certainty.
- A product taken from proposal to paying customer. A new product, a new line, or a feature nobody requested that proved to matter. Be ready to describe how it was tested before it was built.
- Data fluency without being a data engineer. Strong SQL, familiarity with Snowflake, Databricks or an S3 and Parquet data lake, and the ability to validate coverage and match rates before a figure is published.
- AI fluency. Required. What we build has to work in an AI first world, consumed by agents as readily as by people, so this is not a preference. You use current AI tools in your own work, for analysis, prototyping and writing. You can judge where a model or an agent belongs in a product and where it does not. And you understand what a model changes about a requirement: how its output is evaluated, how it fails, and what the underlying data has to look like for it to work at all.
- Written clarity. Plain sentences, a stated position, visible reasoning. A writing sample will be requested and read closely.
- A record of stopping work. A feature, a roadmap item or a project already under way. Be ready to describe what it cost and how the decision was carried.
- Distributed working. Part of engineering is based in India. Most collaboration is written, and it succeeds only when the writing carries the detail.
Experience- At least four years in product management, of which at least three in AdTech, MarTech or a data as a service business.
- Software or data sold to other businesses, serving consumer and business marketing.
- A record of releasing into a platform that already carried customers, and of starting something that did not previously exist.
- Practical use of machine learning or generative AI, either in a product you released or in the way you do the work.
- Bachelor degree or equivalent practical experience. No certification required.
- Advantageous and not required: direct mail or other offline channels, data distributed through the Snowflake or Databricks marketplaces, clean rooms, or a company built by acquisition.
Location: Kirkland, WA. This role requires an in-office presence five days a week with remote flexibility.
Compensation & Benefits- Compensation: $155,000-$185,000 annually, commensurate with experience, and qualifications.
- Flexible Time Off (FTO) and Company Holidays
- Medical, Dental, Vision, and Health Savings Account (HSA) benefits
- 401(k) with Company Match
- Company-Paid Life and AD&D Insurance
- Short-Term and Long-Term Disability Insurance
- Voluntary Life & AD&D Insurance
- Critical Illness and Accident Insurance
Work Authorization: Applicants must be authorized to work for ANY employer in the U.S.
Please Contact: Recruiting | recruiting@deepsync.com
