NPR Network Data Model Problem Statement
Making the nuances and effects of large, complex problems digestible, concrete, and actionable.
We observed that:
-
NPR maintains a portfolio of approximately 20 distinct tools, services and products that enable our relationship with stations and partners.
These tools and services developed over the course of 20 years, most of which were developed as a point solution for a specific station-NPR use case
Many different groups of stakeholders, product managers, and software developers in different divisions of NPR work on technologies in this space
Each piece of technology cares about and controls a different set of unique identifiers and station attributes
-
Each technology in the portfolio defines its own unique identifier.
Technologies with origins in broadcast often use call letters as a unique identifier (UID).
Technologies with origins in digital often use orgIDs as the UID.
Some technologies have no definition of a ‘station’ or ‘business entity’ but instead rely on an application-specific subset of data, such as a digital stream id or website URLs.
-
While nearly every tool cares about a different set of station attributes, some attributes – name, call letters, orgIds, etc. – are referenced across multiple systems
• The source of record for each attribute is poorly defined
• Workflows for updating an attribute often rely on manual processes that differ by technology
• Some NPR technologies are connected, but many are not. For instance, stations input their broadcast schedules into one tool, find their scripts in another, and report their show carriage information into yet another tool
-
A single licensee may operate more than one station as part of a network.
• Some stations within a network may be member stations, while others may not be.
• Stations in a network may inherit rights through a defined hierarchy. Associate stations are children of member stations
• The concept of hierarchy is more applicable to the broadcast space, where there are transmitters and repeaters
• Not all tools are aware of defined hierarchies
• Broadcast-defined hierarchies are useful in limited, specific digital use cases, like those involving the physical location of a station service area
Many licensees operate both radio and tv stations.
-
While some stations elect to use call letters as their public-facing brand, many do not.
• Station staff are likely to refer to their company by brand name, rather than by call letters, even when their company operates only 1-2 broadcast signals
• Networks that broadcast the same signal in multiple geographic locations almost always use a brand name
• Some brands include entities that are not part of the same Member station hierarchy or network
-
Stations enter into voluntary relationships with one another (e.g. Journalism Collaboratives, co-producers on a show), often with the intention of producing more or better journalism and programming.
• Relationships between stations often confer rights to edit or distribute content
• Individual humans may fulfill roles on behalf of multiple partner entities, including both NPR and stations
-
NPR provides tools, services and products to entities that are not member stations.
• Some technologies are available only to NPR member stations, while others are available to a more broadly defined set of partner entities
• Some participants in a collaboration or relationship may not be member stations
-
Public media staff often move between stations and related partner entities, and sometimes work or product content for more than one at a time.
• Stations with smaller staffs often need a single person to manage a variety of tools and tasks, while stations with large staffs may have specialists
• Few systems adequately manage continuity for a single staff person across multiple employers, for instance, a journalist who publishes for multiple stations and NPR
• People who work at a station have the best sense of which of their colleagues should be able to access each tool
• Many tools require user management intervention from NPR staff
-
Many stations rely NPR tools for a major share of their operations.
• Stations are required to use many of the tools and services that we provide in order to access the programs and services
• About half of stations do not have the means to shop around for optional products like our CMS or donation processing form
-
Different NPR technologies rely on divergent underlying technologies (Java vs. Angular, etc.), and may even require manual maintenance of bespoke databases.
...WHICH CAUSED THE FOLLOWING EFFECTS:
-
Understanding how to change a piece of station data requires deep institutional knowledge.
Fostering adoption of any development across technologies or teams is stymied by roadmap divergence, competing priorities and systemic challenges to cross-divisional collaboration.
-
Updating station data attributes may require manual updates across multiple systems.
Finding the cause of an error in an attribute may require looking across multiple tools.
Collating the brands, content and staff associated with a specific partner requires a different process for each connected tool.
Tracking how individual people connect across the system is challenging from both administrative and content production perspectives.
Associating a staff person with the correct partner entity occurs in multiple systems and requires multiple mappings.
Many tools maintain discrete infrastructure and staff to manage duplicative data.
-
After initiating a change to a data attribute, a station staff person may wait days to weeks for that change to appear across every tool that they use.
People at smaller stations are more likely to understand how a variety of tools come together, while people at larger stations are more likely to deeply understand a specific tool or task.
-
Technologies often represent broadcast-relevant hierarchies even when their use cases do not reflect the distribution of radio signals.
WITH THIS IN MIND, HOW MIGHT WE:
Create a way of connecting station data across systems that is available across tech stacks?
Define a reasonable scope for support of collaborations between two or more organizations?
Define a source of record for each piece of station data?
Develop a framework for representing our complex business relationships with stations that prioritizes ease of use for humans?