Quasi-Mystical Arts of Data — Three Models of Data Products
Anna Filippova defines three distinct models for how data teams can deliver value, each with different implications for team structure and skills.
The three data product models
- Analytics Consultant — service-oriented team burning down tickets like a help desk, embedded experts furthering business unit goals. Classic intangible service.
- Data as a Service — analytics engineering workflow producing generalized data applications (interactive dashboards, data syncs) for broad organizational audiences. Software-as-a-service model.
- Insight — the most impact-oriented but squishiest model. Goal is deep understanding that drives decision-making. Collapses the boundary between BI and data science. Requires a “Product Manager of insights.”
The analyst’s shadow roles
The “great analyst” archetype is actually playing multiple shadow roles simultaneously:
- UX designer — ensuring the right information is obtained in the most accessible way
- Systems engineer — traversing metric data lineage from source to dashboard
- Persuasive writer — translating findings into ideas for action
- Product manager — articulating ROI of data investments for future needs
The systemic problem
Instead of building cross-functional teams of experts, we ask individual data professionals to be experts in everything. When this inevitably fails, we ask them to “say no, a lot.”
Connects to analytics is a profession, data team operations, analytics craft.
Open questions
- Is model #3 (Insight) achievable without restructuring how data teams are staffed?
- What does “Product Manager of insights” look like as an actual role?