A Data Product is any product or service that leverages data to solve a problem, improve a process, or enhance an experience. Data products can be internal or external, depending on who the end users are. Some examples of data products are:
- Customer, Employee, Vendor, and Product 360's
- A recommendation system that suggests relevant items to customers based on their preferences and behavior
- A fraud detection system that flags suspicious transactions and alerts the appropriate authorities
- A dashboard that visualizes critical performance indicators and trends for business decision-makers
- A chatbot that answers common questions and provides personalized support to users
Data Product Management
Data product management is a relatively new discipline combining data science, engineering, and business skills and expertise. Data product managers are responsible for creating, launching, and optimizing data-driven products and services that provide value to customers and stakeholders.
When developing Data Products, the following questions should be analyzed against business needs:
Are there existing data products?
Are there existing data products, services, or similar solutions which adequately address the need for the data in question? Too often we identify a gap in our data product needs. Instead of searching for an existing solution or service which could be enhanced to meet the gap, we build a new product or service which, ultimately, undermines the objective of the product management discipline.
How will the success of the data product be measured?
In other words, what does "good" look like? Using Agile methods will ensure we are only building features and capabilities we need and which we can measure for value delivery.
When answering this question, data quality is a key consideration. Organizations must be careful not to require all data to be perfect before release. Understanding the requirements for accuracy, consistency, timeliness, etc.... for the specific use cases the data product is meant to address is paramount.
How much synergy does the proposed data product provide?
If a dozen use cases are being addressed, priority should be given to those data products which address the highest number of use cases. Of course, value, feasibility, compliance, and other concerns need also be considered.
An excellent example of a high-synergy data product, one most organizations should consider for early development, is Customer-360. A customer-360-type data product delivers value to many use cases and digital products. This is also an area where many companies spend millions of dollars to produce mediocre results (e.g., Master Data Management), thus there are potentially both strategic (i.e., customer-facing) and operational (internally facing) benefits to be achieved.
Is the Data Product relevant?
Is the Data Product relevant and meaningful to its consumers; be they people or machines? A good data product caters to a reasonably broad array of requirements and use cases, but limits exist. It should include what's needed for the known universe of use cases which, if created correctly, includes some not-so-distant future needs but is restricted to a solid business case.
Again, Agile methodologies for Data Product Development ensure fluid and manageable processes for identifying, prioritizing, incorporating, and releasing new features and capabilities. Thus, when we build our Data Products, we should constrain them to what is known and ensure they meet the needs of the systems and people who will use them.
What is Data Product Management?
Data product management is the process of defining, developing, and delivering data products. Data product managers work at the intersection of data science, engineering, and business, and they need to have a deep understanding of all three domains. Data product managers perform various tasks, such as:
- Identifying customer needs and market opportunities for data products
- Defining the vision, strategy, and roadmap for data products
- Collaborating with data scientists and engineers to design, build, and test data products
- Launching and iterating on data products based on user feedback and data analysis
- Measuring and communicating the impact and value of data products
Data Product Teams
Data Product Teams, also called Pods, should be created in an agile structure and could include the following roles:
- Data Product Owner
- Data Stewards
- Data Product Analysts
- Data Engineers
- Data Architects
- Data Designers
- IT Leads, technology engineers, and other DevOps personnel
Other roles can and should exist either full or part-time, depending on the specific needs of the business and the particular data product.
Skills and Competencies of a Data Product Manager
Data Product Managers differ from traditional product managers in several ways. First, they have a deep understanding of data and analytics, and how they can be used to create innovative solutions. Second, they have a solid technical background and can communicate effectively with data scientists and engineers. Third, they have a strategic vision and can align the data product roadmap with the business goals and customer needs.
Data product managers need to have a combination of technical, analytical, and business skills in addition to soft skills such as communication, collaboration, and leadership. Some of the critical skills and competencies of a data product manager are:
- Data literacy: Understanding, interpreting, and communicating data effectively
- Data intuition: Identify patterns, insights, and opportunities from data
- Data storytelling: The ability to present data in a clear, compelling, and actionable way
- Data ethics: The ability to ensure that data products are fair, transparent, and respectful of user privacy and security
- Product management: The ability to define, prioritize, and execute product goals and requirements
- Stakeholder management: The ability to align and influence various stakeholders across different functions and levels
- User empathy: The ability to understand and anticipate user needs, pain points, and expectations
- User research: The ability to conduct qualitative and quantitative research methods to validate user problems and solutions
- User testing: The ability to design and run experiments to evaluate user behavior and satisfaction with data products
Building a successful data products capability requires investment in both the technical infrastructure and the human skills to manage and evolve these products over time.
