Skip to content

What is the role of a Data Product Manager?

The importance and demand for data have inspired lots of data tools and technological innovations to life. But with that also came extreme complexity. A16z’s modern data industry landscape is a good overview of how complex today’s data stack can be for any size of the company. While previously, as long as you had ‘data’ you could analyze and service your needs through Excel, R, and SQL interfaces, today that forms a small part of the modern data stack. 

With an immense amount of data, you have internal metric needs for tracking performance, and product metrics to help you decide on product goals, and even your customers and partners require you to provide data. To manage all this complexity, we have seen the role called Data Product Manager (DPM) emerge at companies. As data becomes more pervasive and central to business operations, it also becomes more valuable and powerful.

For this purpose, the evolving and emergent role of the DPM has peeked employer interest in recent years. In the US market, Zippia currently lists over 9,000 open DPM positions. Over 85,000 positions in data product management are currently open in the European Union and posted on LinkedIn. The DPM role is in high demand, rendering it a lucrative six-figure position. Glassdoor reports the expected compensation range between $83,000 and $203,000 with a $130,000 median.

The responsibilities of such a role vary widely based on the organization's data maturity and built-in data culture. Data products and services are becoming an integral part of any business model; thus, the need for a role to lead such initiatives has naturally grown in importance. In this article, we discuss the different traits, skills, and characteristics possessed by a good DPM.

What is a Data Product Manager?

DPMs can be considered the glue between all parts of the business (internal and external) that requires data.

A data team, especially a mature one, is often subdivided into several partitions, including (but not limited to) data engineering, data science, and data operations. As such, one of the primary duties of a DPM is to ensure goal alignment and work with stakeholders to define what a new product will do. The stakeholders can either be internal teams or external clients - depending on the nature of the business. For example, if the data science team is building an in-house fraud detection system, then the stakeholders would most probably be the risk and fraud team.

The DPM also identifies potential bottlenecks early on in the process so that they can be mitigated before they become more severe obstacles later. One typical example of a data project is starting an ambitious project without owning or having access to the required data. This issue ultimately results in a delayed or, worse, failed project. Other common bottlenecks in the data pipeline include data veracity issues, misaligned priorities, incomplete or unclear requirements, and also talent shortage.

In addition to being responsible for communicating both internally and externally on behalf of their data product, some DPMs are also responsible for ensuring consistency across all business products (including those not owned by the data vertical). Building products that integrate seamlessly not only increase product quality but also improves customer trust. Business products should follow the same productization process, the same design principles, and provide the same user feel and experience. This is an important role when it comes time to optimize these products in the future (or potentially, even retire them). The products you build must also be consistent across other widely adopted products and follow the same standards.

A simple example is the ‘settings’ icon; everyone understands that the gear icon represents some form of ‘settings’ menu. So one should think twice about using a different icon. A common mistake of data product inconsistencies is perhaps the delivery method. For example, having one product delivered via a web service and another in the form of a desktop application. It would be drastically better if both products shared the same platform - and bonus points if they could communicate together!

The roles and responsibilities of a DPM continue to evolve. A frequent question that businesses and aspiring DPMs instinctively ask is: “What does it take for a DPM to be truly effective?”. When we dissect the DPM role, we can immediately start to pinpoint critical skills for this role.

  • Product-driven problem-solving
    • The DPM solves problems by managing data products. While DPMs can often be mistaken as those that provide data access to stakeholders, what DPMs are in charge of is solving some core problems through data-driven products. With ever more complex data stacks, DPMs are no different from PMs that build software.
  • Domain knowledge and curiosity for new stacks
    • Innovation in the tech and data industry is frequent and fast. DPMs must stay on top of the latest data workflows and stacks to provide the best possible data product to the teams. Understanding the pros and cons of different stacks, commercials, management cost, ease of adoption, and flexibility is key. But without the domain knowledge and curiosity to search and discover new techniques, the organization can easily fall behind working with legacy and outdated tools that force individuals to find workarounds for their data needs.
  • Internal navigation skills
    • What makes DPM difficult is that data has so many stakeholders. Data engineering, science, analyst, executive, product, operation, sales, and marketing, everyone is a stakeholder to DPMs. Sometimes even external parties like customers and partners that require data from the provider. Given the remote working environment today, iterative communication is costly (a simple Slack message can require days for a reply). DPMs, to make things progress, need to be ready for internal navigation, persuasion, and alignment. Sometimes internal navigation can be the most taxing part of the day. Behind the glories of analytical frameworks and products lies extreme efforts of internal navigation!


Data product managers are responsible for the end-to-end lifecycle of a data product (a.k.a, the single source of truth)

The typical data product development lifecycle is composed of three main stages:

  • Inception and Planning
  • Development
  • Productionization

A DPM is responsible for keeping track of projects, ensuring they are properly planned, executed, and launched successfully. The DPM must align the data strategies with the business objectives. For example, let’s say that a company is looking to improve customer retention. Using machine learning, they can offer customer discounts based on their shopping history. The DPM must understand the business enough to determine which customers should be eligible for discounts, which are the relevant metrics for the machine learning algorithm, what kind of discounts are available, and how to communicate the results back. In short, they are responsible for discussing and planning the deliverables and ensuring the successful delivery of such requirements.

Inception and Planning

Inception and planning a data product is perhaps the hardest phase of an internal data project. You fight with priority, the possibility of success, budgeting, and also accountability planning. Given everyone wants their data now, it is difficult to manage expectations of how long of a journey it could be to get a proper data product up and running internally. ETL is probably the clearest of all but also the most expensive investment in time and money that you burn all the energy getting to a data mart.

In some cases, the DPM is also responsible for deciding whether a data product should be decommissioned. Deciding whether to continue investing in a product or pursue alternate routes (e.g., Amplitude) is a critical decision. 

Data products are often highly specialized, so a good DPM needs to know what kind of information they need before starting any project. Unfortunately, the modern data stack evolves so quickly that the DPM can be caught in playing catch-up.


Often internal data products or projects fail because the initial stakeholder alignment is easy (who doesn't want data?) execution can seem relatively easy to get support, and someone commits to it part-time without clear accountability. The DPM should possess excellent people and communication skills and work effectively across departments (i.e., engineers, designers, and marketers, to name a few) to ensure project completion on time. A DPM acts as an intermediary between the various teams working on the project.

When working on a product, everyone involved must understand the role they play in delivering a successful result. If not, it can lead to wasted time or a sub-par product. DPMs should ensure that each team member has sufficient information about the project. Especially in an agile environment, priorities, requirements, and timelines tend to witness several micro-adjustments as the project matures. The DPM must ensure that all data efforts contribute to the overall product growth and success. They are responsible for ensuring that any new features or changes align with the respective company goals.


What is all the planning for if it does not get traction? The DPM must ensure that all parties deliver according to their agreed timelines. Upon completing all development efforts, the focus must now shift to getting users active on the new product. 

A great DPM excels in engaging internal users to the new product, and creating or migrating workflows over to the new product. Frequently, data projects start with a high level of excitement across all teams, only to be fizzled out over time, and a great DPM is excellent in making sure it stays afloat until the new product blends into a new or existing workflow. Without a healthy level of communication between the data verticals and the stakeholders, interest might be lost in the project. The DPM must keep a strong relationship with the stakeholders and all other parties involved, frequently updating everyone to ensure that the project buy-in remains high. 

Deliver success metrics

Are you implementing a new data product to your internal dashboard or adding Amundsen? What is success for you? Another key trait for DPMs is a strong knowledge of what makes a product successful - and, most importantly, understanding how to measure success. Often DPMs are knee-deep into the data stacks, ensuring data quality and consistency that internal delivery KPIs are less prioritized. Given there are a million possibilities of how a data project or product can go, it is critical to define (ideally beforehand) success metrics of even internal data projects. Nobody should work to implement Datahub only to see it reach 10 MAU (sound familiar?).

Picking the correct KPIs is a skill in its own right though. The end game here is to determine a KPI to monitor so that you not only assess the level of success but also communicate using the KPI to drive continuous improvements without having to justify each new project iteration especially post-release.

Courage to Drive Change

What are you going to do if you put out a self-serve BI dashboard and there are no active users? What if you realize that you missed out on a critical stack or pipeline but you no longer have resources assigned to your product post-release? Data projects, often being an internal product or later-stage undefined area, face difficulties in getting priority. This is why DPMs require an element of courage to drive change within the organization. Data products and projects can take a while to blend into the day-to-day operations of a company as its success depends on the utilization of previously unavailable data. There is a learning curve for the company and you as the DPM must carry on to drive adoption and data literacy even at times when there are internal challenges to adopting data-informed practices.

Driving the internal success of modern data platforms

The DPM is usually best positioned to own and manage the objectives of the company’s data platform. For the sake of the scope of this article, the data platform can be thought of as the single source of truth for the entire business.

A key piece and often overlooked part (due to so many technical details to manage) of managing the data platform is ensuring data engineering, science, analytics, metrics, and business are aligned (best if in real-time sync). This enables access to reliable information for decision-making and reduces the iteration time from data engineering and data science. It is worth noting how important it is for businesses to have consistent definitions across different teams so they are on the same page. One common mix-up companies face is varying definitions of concepts or even KPIs!

The DPM must ensure that the data platform remains a single source of truth for all stakeholders, ensuring there is no disconnected component in the platform. For example, having a Notion page with your taxonomy defined is prone to breaking easily since metrics, queries, and tables constantly change, and relying on someone to update the Notion page each time (let alone governance) is unrealistic. Another common example is table freshness, what good is there for setting up a catalog if you cannot instantly determine if the table is fresh? Make sure you are creating a data platform that becomes a system and source of truth that all stakeholders can rely on, rather than having to manage and contribute based on documentation and processes.

Building relationships and stakeholder management

As the DPM, you are the main champion of your company’s data products. 

You must be able to articulate the product vision to stakeholders, team members and customers. 

You must represent the needs of stakeholders, build domain expertise and drive data governance change. You should also be prepared to communicate with stakeholders who may not have the same level of expertise as you do on a given subject matter.

This means you should demonstrate how the product contributes towards achieving business goals and objectives in a compelling way. You should have the ability to explain complex technical concepts in layman terms. This will help everyone understand the true underlying product value.

As a DPM, you will often be working with various teams within the organization. You will need to build relationships with other managers and team members from various departments such as marketing, sales, and engineering to ensure project success. 

Having good relationships allow you to get feedback quickly, enabling faster decisions. This will allow you to understand interdepartmental goals and objectives, which can help build better products that meet those needs over the longer term. You also need to have an understanding of current industry trends and strategic directions. What are competitors doing? Are there any new technologies or tools that could help improve how the business operates? Knowing what is going on out there will help make sure that no one has an advantage over you when it comes to data-driven decision-making. You would be in a much better position to assess product-market fit.

Building a failed product is significantly easier than a successful one. The DPM needs to say no, more often, and perform negative engineering to ensure the success of at least one or two use cases of the data product. 


Finding that delicate balance between what the business wants and what the data can deliver is perhaps one of the most difficult challenges modern organizations face. The DPM role is intended purely to fill that gap. Bridge the business vision and the technicalities of the data vertical and increase data literacy to contribute to business success. 

This is an exciting time for DPMs, as we see a massive increase in demand for DPMs. Data is a fundamental asset to any business, and ensuring that a company takes full advantage of its data is your responsibility as the DPM.

The demand for DPM is clear, and the versatility of the role also makes it rather attractive. A few years back, the Harvard Business Review described the commercial role of a data scientist as “the sexiest job of the 21st century”. The DPM is what enables all of this at scale!