Why Spotify Built Data Science Personas
Spotify’s Analytics Platform builds tools for data practitioners — notebooks, dashboards, and metrics. The team’s goal is to create products that genuinely connect with users, but that requires a clear picture of who those users are. That picture has gotten blurry as the industry has expanded the “Data Scientist” title to cover a wide range of work with different needs.
Many people who work with data don’t have that title at all. The catch-all term “data practitioner” was too vague to steer product development effectively. To fix that, Spotify’s team built a set of personas based on a mixed-methods study: qualitative interviews with employees plus quantitative analysis of behavioral data. The goal was to make the personas instantly understandable and actionable, so product leaders could use them in real decisions.
The result was six archetypes covering Spotify’s data practitioner universe. Along the way, the team learned several lessons worth sharing with anyone doing similar foundational research.
Lessons from the Persona-Building Process
Make Research an Iterative Loop, Not a Relay Race
Instead of handing off from qualitative to quantitative work, the team ran them in a continuous cycle. User interviews led to draft quantitative rules for classifying personas; those rules were then tested against more interviews. Each interview either validated the classification logic or exposed a tie that required refining the rules before the next round.
Activate Findings Immediately
Foundational research often stalls when it’s time to act on it. Spotify’s team addressed this by making tailored recommendations a fixture of their share-outs. Slides identified which persona each team should focus on, opening a discussion about how to respond. Within a quarter, the framework evolved from a conversation starter into the basis for setting goals and metrics.
Keep Stakeholders in the Loop Throughout
Research teams often go quiet during a project and emerge only with final results. To avoid that, Spotify’s team aligned on scope before starting, kept an open Slack channel for updates, and invited anyone in the product area to observe interviews. At mid-project milestones, they shared “Postcards from the Field” — unpolished qualitative observations — to keep teams engaged. Later, persona previews were shared with product leads for feedback.
Prefer Simple Methods Over Clever Ones
The classification methodology was deliberately simple, relying on rules like “has used dashboarding tools more than 90% of data scientists.” This made results easy to explain and easy for stakeholders to trust.
Extend the Work Beyond the Original Scope
To keep the personas alive after the project, Spotify built a dataset that classifies every internal user on a daily basis. That dataset now feeds recruitment efforts and deepens product metrics by segmenting all key results by persona type. The team also worked with designers to produce branded illustrations, making each persona a memorable, shareable artifact.
The Six Personas
Curious where you fit? Here is how Spotify describes its data practitioner archetypes:
Paola the Product Strategist — Paolas are the most common data science persona at Spotify. They analyze data and tell stories with it to inform strategy, often partnering closely with non-technical stakeholders. Elsewhere they might hold titles like Product Analyst or Data Scientist, Analytics.
Eli the Extensive Explorer — Elis are the second-most-common type. They are data scientists, machine learning engineers, or research scientists who dig deep into data for complex research and models. This persona merges traditional data science with ML engineering roles.
Ivan the Influencer — Ivans are leaders supporting data teams and shaping best practices and processes. They may be team leads or managers.
Daryl the Data Viz Artist — Daryls craft interactive, visual stories and spend most of their time building dashboards for non-technical stakeholders. This is a niche role that often lacks its own title at many companies.
Sigrid the Systems Engineer — Sigrids build infrastructure that helps technical stakeholders use data for analysis. They commonly go by Data Engineer or Analytics Engineer.
Dalia the Data Dabbler — Dalias do not carry a data scientist title but use data for analysis or decisions. They typically tackle straightforward questions, often with help from the other personas.
Most Spotify data practitioners identified strongly with a single persona, but some reported being a hybrid of two archetypes.
