Last updated: November 14, 2026
By Mike Carter, Managing Director, KORE1
A data product manager job description should name the data products the seat owns, who uses each one, the promises attached, who pays for the platform, and a base range of roughly $140,000 to $240,000 by level. This page covers the platform version of the seat, the one whose customers are analysts and engineers inside your own company. Most postings for it read like a consumer-app PM ad with the word “data” added, and the candidates worth reaching notice on the first pass.
One question stalled a search for us this fall. “If the 5 a.m. load runs late, who decides whether the dashboards go out?”
The setup, with the caveat that it’s a composite and the details are changed. A life-sciences software company in Cambridge, Massachusetts, about 450 people, ran its analytics on Snowflake with dbt models, Fivetran feeds, Airflow schedules, and Looker on top. Roughly forty analysts in finance, commercial, and clinical operations leaned on it every day, and the warehouse held more than 1,900 tables, around 600 of which a recent cleanup had found nobody querying in ninety days.
The posting for its first data product manager was lifted almost whole from the team that ran the customer-facing app. “Own the roadmap.” “Drive adoption.” “Partner with stakeholders.” Eleven more bullets in the same key. That was the posting. No stack. No pay range. And not one line about what the platform cost or who picked up the phone when it broke.
The finalist we liked had run an internal data platform at a Boston fintech for three years, and she put her question to us on the first call. Nobody knew. The VP of engineering assumed the product manager would decide. The finance lead assumed she would, since month-end close ran off that load. And the monthly warehouse bill had climbed from roughly $38,000 to roughly $71,000 in a year, sat inside the engineering budget, and came mostly from queries run by finance and clinical operations. She asked about that too.
She took an offer from a company that had written the answers down. Fair enough.
My seat in this. KORE1 earns a fee when a company hires a data PM through our data product manager staffing practice, so a posting that fills itself costs us business, and I’d still rather you wrote a good one. I also came to staffing after two decades in sales and marketing, which makes me, professionally, a stakeholder. I’m the person who emails the dashboard owner asking why the number moved. A platform PM is who I always wished I could find. If you haven’t settled what kind of data product manager you need, read our guide to hiring a data product manager first. A general seat belongs on the product manager job description template, and a role that sits close to engineering fits the technical product manager job description template better.

What a Data Platform Product Manager Is Hired to Own
A data product manager owns data the way a product manager owns software, with customers, a roadmap, and a standard of quality. In a platform seat the customers are internal, which means analysts, data scientists, and engineers, and the manager answers for adoption, reliability, and cost instead of a user-facing feature.
The idea isn’t new. Zhamak Dehghani laid it out in a December 2020 piece on Martin Fowler’s site, arguing that analytical data “must be treated as a product, and the consumers of that data should be treated as customers.” She also named a role, the domain data product owner, who has to understand who the data users are and how they use the data. Six years on, that role has a market price and a recruiter’s inbox full of postings that misread it.
Nice theory. It breaks on the Monday after the hire starts, when the first request is a dashboard, the second is a dashboard, and by Friday the new platform PM has become an analyst with a better title. It happens fast.
The posting’s real work is smaller than most drafts assume. Skip the mission statement and describe the platform plainly enough that someone who has run one can tell what they’d be walking into.
List the Data Products, Who Uses Each, and What You Promise
Start with an inventory. Strong candidates ask for it in a set order. What do people rely on, who relies on it, what happens when it’s wrong, and what has already been promised?
| What you run | Who relies on it | Example line for the posting |
|---|---|---|
| Certified warehouse models | Analysts, and everyone who reads their reports | “42 certified models feed month-end close, refreshed by 6 a.m. Eastern.” |
| Metric definitions or a semantic layer | Every dashboard, finance, the board deck | “One definition of net revenue retention, owned by this seat, used in 31 dashboards.” |
| Event and product telemetry | Product teams and the experimentation group | “Event schema reviewed here before any new event ships.” |
| Feature tables and training sets | Data scientists, and any model in production | “Feature tables behind three live models, rebuilt nightly.” |
| Catalog and lineage | Anyone hunting for the right table, and auditors | “Half of certified tables described in the catalog today, all of them by next year.” |
| Ingestion connectors | Everything upstream of the warehouse | “74 Fivetran connectors and 6 custom ones, two of them flaky.” |
Round numbers are fine, and so is “most” when legal is nervous, but leaving a row out because the answer is embarrassing isn’t, since candidates read an omission as the answer. Honesty is the point.
Then address the 600-tables problem before anyone raises it. Can the new PM turn something off, and who has to agree? Some strong candidates will take a seat where finance signs off on retiring anything tied to the close, because somebody got burned once. They’ll pass on a seat where the question was never raised, since that debate is how a platform PM spends the first year. It’s a long year.
Contracts next. Say whether you have any. “Establish data contracts” reads very differently to a candidate depending on whether you have none or three hundred. A contract can be mechanical. In dbt’s model contract documentation, a model whose logic or input data doesn’t match the declared columns and types simply “does not build.” The Open Data Contract Standard, at version 3.2.0 as we write, gives a contract a service-level section and a data quality section of its own. Tell candidates which of those you actually use. “Contracts written in dbt for the certified models, nothing upstream yet” is a sentence a platform PM can plan around. “Own data quality” isn’t.
Say Who Pays for the Platform
Cost is where this posting differs most from the generic product manager ad, and it’s where most drafts go quiet.
Between warehouse compute, storage, the ingestion tool, the BI licenses, and whatever the data scientists spin up on the side, somebody owns that total, somebody sees it monthly, and somebody can say no to a new tool, and in plenty of companies those are three different people. Three people, three answers. A senior candidate wants all three names, or at least all three titles. They’ll also ask whether cost is a goal or a limit, which are different jobs. One cuts spend. The other keeps the platform inside a number someone else picked.
Put a monthly spend band in the posting, say who holds the budget, and say whether the seat approves new tools or only recommends them. In the Cambridge search, the bill was the second question the finalist asked. At that company it had nearly doubled in a year, and the people causing it had no idea.

Where the Seat Stops
The platform PM decides what gets built next and what stops running. Data engineers build it, and if that’s the person you need, our data engineer job description template is the page to use. The analytics engineer writes the models the PM prioritizes, a line the analytics engineer job description template draws clearly. Architects choose the shape of the whole thing, covered in the data architect job description template, and the person who writes the rules about who may see which column is a different seat again, which our data governance analyst job description template handles.
If your draft lists duties from three of those, you’re hiring a department. That’s a different search.
Say who the seat reports to, too. We see three arrangements, the head of data, the CTO, and the product organization, and the candidate’s second question changes with each. Under the head of data, they ask about autonomy. Under the CTO, priority against engineering work. Under product, whether anyone there understands the problem. Each changes the interview. Once the posting is out, our guide on how to interview a data product manager covers what to ask the people who answer it.
Name the Tools, and the Dates Attached to Them
List your stack in the section about what the hire inherits, and ask for experience with “a cloud data warehouse” in the requirements. Naming Snowflake as a hard requirement shrinks the pool for no return, since a platform PM who ran Databricks for three years will pick up a new warehouse within a month, and the posting then reads as if nobody on your team has ever moved between tools. Pools are small enough already.
If a catalog is part of the platform, put its date in the posting. Microsoft is retiring Microsoft Purview Data Governance, including the Unified Catalog, with end of support on November 12, 2027 and governance moving to Microsoft Fabric and the OneLake catalog. Google discontinued Data Catalog on June 1, 2026, with Knowledge Catalog as its successor. A platform PM who inherits either one inherits a migration, and they’d rather read that in the posting than discover it in week three. Our governance posting guide walks through the same catalog changes from the policy side.
Data Product Manager Job Description Template
Fill in every bracket. Parenthetical notes are instructions to whoever drafts, and they get deleted before the post goes out. Drop any section that doesn’t apply.
Job Title
[Data Product Manager, Data Platform] (Add “Senior” or “Principal” only if the seat leads other product managers or owns a budget above $[N]. If the work is dashboards for one team, you’re posting an analytics manager. If it’s one customer-facing feature built on data, use the general product manager template.)
About the Platform
- [N] analysts, data scientists, and engineers use the platform every week
- Built on [warehouse], [transformation tool], [orchestration], and [BI tool], at about $[band] a month
- [One change coming in the next two years, such as a warehouse migration, a catalog replacement, or a new AI workload]
Data Products You’ll Own
- [Certified models], [N] of them, feeding [finance close / board reporting]
- [Metric definitions], used in [N] dashboards
- [Event schema / feature tables / catalog], whichever are real today
- [Ingestion connectors], [N] managed and [N] custom
Who Uses It and What They Need From You
[Finance closes the month on three certified models and needs them by 6 a.m. Commercial analytics needs new sources added within two weeks. Data science needs feature tables that don’t change shape without notice.] (Three or four real examples beat a paragraph of “stakeholders.”)
Promises You Keep
[Certified models refreshed by [time] on [N]% of days. Test failures reviewed within [N] hours. Requests answered within [N] business days.] (Write only promises that somebody measures today. Candidates will ask to see the number.)
What You Decide
- The platform roadmap, with [the largest bets reviewed by the head of data]
- Which datasets get certified, and which get retired, [alone / with finance sign-off for anything tied to the close]
- [Tool approvals up to $[N] a year]
- What counts as “on time” and “correct” for every certified model
What Sits Elsewhere
[Data engineering reports to the head of data engineering and builds what you prioritize. Access policy belongs to the governance team. Dashboards belong to the teams that ask for them.] (Write this even when it feels awkward. Every finalist asks.)
The First Twelve Months
- A baseline for on-time delivery, quality test coverage, and monthly spend, built by day ninety
- [A reliability result, such as certified models on time 95% of days]
- [A cost result, such as spend held flat while usage grows]
- A decision on what to retire, made and carried out
Background We’re Looking For
- Owned an internal data platform or a set of shared data products, with named consumers
- Written SQL and read dbt models or their equivalent, enough to argue with the people who build them
- Set and measured a freshness or quality target, and explained a miss
- [Years of experience, honestly scaled to the level above]
Helpful, Not Required
- Run a warehouse migration or a catalog replacement
- Held a cost target for a cloud data platform
- Written data contracts that producers actually signed
Pay, Benefits, and Location
Annual base of $[low] to $[high], the range we expect in good faith to pay at this level. [Bonus target N%.] [Benefits in two lines.] Working from [city], [remote / hybrid N days a week].
Three Sources That Disagree, and Massachusetts Wants the Number Anyway
Nobody publishes a clean figure for this title, and the sources that try disagree in a way that tells you something.
| Source | What it measures | Figure | The catch |
|---|---|---|---|
| ZipRecruiter, Nov. 14, 2026 | Data Product Manager | $159,405 average, $141,000 to $197,000 middle range | Its low end is $51,500, which is a different job wearing the title |
| Salary.com, Nov. 1, 2026 | Product Manager, Data Science | $149,018 median, $138,538 to $161,236 middle range | Nearest title it tracks, and a different one |
| Bureau of Labor Statistics, May 2025 | Computer and information systems managers | $175,140 median, 16% growth projected for 2025 to 2035 | Has no product manager line |
BLS has no line for data product managers, so the managers’ series is the closest federal anchor and a loose one. ZipRecruiter’s $51,500 floor is the tell about every aggregator on this title. Three words that describe a platform owner at one company describe a reporting coordinator at the next, and the averages blend both. On our desk, a mid-level platform seat clears at roughly $140,000 to $180,000 base and a senior one at $180,000 to $240,000, the same bands as in our hiring guide. Two postings that sat on ZipRecruiter’s page for this title on November 14 show how wide it gets inside one metro. A data product manager lead role in Los Angeles listed $179,000 to $286,000, and a technical product manager for data and internal products listed $180,000 to $230,000. Averages hide that. Our salary benchmark assistant checks a base figure against your metro in a few minutes. The tracker-by-tracker version of that picture is in our data product manager pay analysis.

Now the part that makes the range non-negotiable in one state. Massachusetts’s Wage Transparency Act took effect October 29, 2025, and the Attorney General’s guidance and the statute itself, Chapter 149, Section 105F, set out the rules. The pieces that matter for a data platform seat:
- It covers employers with 25 or more employees whose primary place of work is Massachusetts, counted as an average across the prior year’s pay periods, part-time and temporary staff included.
- A role that can be performed remotely to a Massachusetts worksite is covered.
- A “posting” is any advertisement meant to recruit applicants for a specific position, and the statute says that includes recruitment done “indirectly through a third party.” The employer “or its agent” has to disclose the range.
- The range is the annual salary the employer “reasonably and in good faith expects to pay for such position at that time.”
- An employee offered a promotion or a transfer into a position with different job responsibilities gets the range too, and so does any applicant or incumbent who asks.
The penalties start gently. A warning for the first offense, up to $500 for the second, up to $1,000 for the third. But the statute counts one or more postings made by the same employer within 48 hours as a single offense, so a batch of five platform reqs with no ranges is one strike, not five. And through October 29, 2027, the Attorney General’s office gives employers two business days to cure a defect after a Notice to Cure letter, which is useful, though I wouldn’t plan around a grace period.
Questions We Get on Data PM Drafts
Data product manager or data platform product manager, which goes in the title?
Data product manager. Candidates search that phrase, and the word platform belongs in the first paragraph, where it does the filtering.
A longer title buys a narrower audience and nothing else. If the seat has real seniority, add “Senior” and say what it leads.
Our senior analytics engineer makes $165,000. Can the PM seat pay less?
It can. You’ll feel it in the first round. A platform PM who earns less than the people whose work they prioritize usually can’t hold a line with them.
We’re promoting someone into the seat. Does Massachusetts still want a range?
Yes, under Section 105F(c), which requires the pay range for an employee offered a promotion or a transfer into a position with different job responsibilities.
Moving a senior analytics engineer into a platform PM seat is exactly that, so hand over the range in writing before the conversation about the title and not after it.
Does the law reach postings a recruiter writes for us?
It does. The statute defines a posting to include recruitment done indirectly through a third party, and it makes the employer “or its agent” responsible for disclosing the range.
Which means the range has to exist before the search starts. That’s a decision for you, not for the agency, and it’s better made against a real budget than against whatever the first candidate says they want.
What do we measure in year one when there’s no baseline?
Build the baseline in the first ninety days and write that into the goals. Count certified tables, measure on-time loads, and put a number on monthly spend.
The targets come after. A goal like “raise on-time delivery from 81% to 95%” needs the 81%, and at most platforms nobody has it yet.
We’re mid-migration to a new warehouse. Does that go in the posting?
Say so, with the date you expect to finish and what’s moving first.
The candidate who has done a migration will want the job more for it, and the one who hasn’t was going to find out in the first week anyway, so you may as well be the one who tells them. If you’re also swapping the catalog, put the old tool’s end-of-support date next to the new one. Telling the truth early rarely costs you a good finalist.
A Last Check Before It Goes Up
Hand the draft to your sharpest analyst and ask one thing. Could they tell, from the posting alone, who to call at 5 a.m.? Simple test. If they can’t, a candidate who has run a platform can’t either. If the seat will grow into leading other product managers, the principal product manager job description template is the next rung, and if the work is mostly models and AI features, the AI product manager job description template fits better.
KORE1 has recruited for technology teams since 2005, in 30+ U.S. metros, with recruiters who average 15+ years in the work. Of the people we place, 92% are still in the seat a year later. Our technology searches average 17 days to fill, and senior data PM searches trend toward 26 to 34 days. A platform seat is usually a direct-hire search, though a contract product manager can hold it through a migration. If you’d like a second set of eyes, send us the draft, and we’ll point to the line a candidate would push on first.

