Foundational Design

The Core Triad:
User, Data, & Task.

Prior to creating intricate metrics or constructing systems, every effective Data Product must understand the overlap between the user, their task, and the specific information needed to complete it.

Data Product User, Data, and Task Metrics

The Three Pillars of Design

The success of a data product depends on the alignment of three key elements. If even one is missed, the product's adoption will suffer.

The User

Who is using this data? It is important to have a thorough understanding of their technical expertise, daily routines, and preferred methods of information consumption.

Key Questions

  • Are they an Analyst, Exec, or ML Model?
  • What tools do they already use?
  • What is their data literacy level?

The Task

What does the 'Job to be Done' entail? Data only holds value when it is utilized for a particular action, decision, or automated process that propels the business forward.

Key Questions

  • Is this for operational action or strategy?
  • How frequently does this task occur?
  • What happens if the task is delayed?

The Data

What specific data is needed to assist the User with the Task? This determines the schema, latency, history, and quality SLA needed.

Key Questions

  • Do they need real-time or daily batch data?
  • What level of granularity is required?
  • Are there PII/Security constraints?
Designing Output Ports

Where the Triad Intersects

The responses of the User, Data, and Task trio greatly influence your engineering approach. Output Ports Delivering the data may require multiple output ports in order to accommodate various triads.

Your product will fail if you create a stunning real-time API for a CFO who simply needs a monthly PDF report.

Mapping the Triad to Architecture

Scenario A: The Executive

User: CMO (Non-technical)

Task: Monthly Budget Allocation

Data: Highly aggregated, historical

→ Build a curated BI Dashboard View.

Scenario B: The Application

User: Recommendation Engine (Machine)

Task: Serve live product suggestions

Data: Granular, sub-second latency

→ Build a high-performance REST API.

Stop Building in a Vacuum

Before writing any SQL code, begin defining your data products by placing a strong emphasis on the User and their Task.

Review Metrics Framework