Building the document type taxonomy that now grounds all of Google Docs strategy

Overview

Product decisions were being made on a vague, abstract understanding of users — no concrete grounding in what people were actually trying to do in Docs.

Started small: a quick in-product survey to get more precise. It surfaced that people talk about their work in terms of the tangible documents they're building and the overall task they sit inside — a strong first altitude for a mental model of how people approach Docs.

Approach

Built a project proposal and pitched it to research leadership and PM leadership; approved with funding.

~4 months of work combining three methods: in-product surveys (internal and paid users, manually coded into a candidate set of document-type codes), a week-long diary study following what people did with Docs and other text editors and why, and a review of template libraries across the top text editors on the market — plus a vendor running supplementary secondary research.

Landed a concrete, well-documented, quantified taxonomy of document types right before the planning and strategy cycle — timing was deliberate.

Validated and scaled the taxonomy through successive surveys: extensive text-editor users, users matching key archetypes, users across organizational profiles, and a large in-product survey of Docs users directly — producing deep-dive reference material per document type that teams could pull into product requirements or strategy work.

Impact

Adopted immediately and enthusiastically by the team; became a core lens for how features were prioritized.

10 shipped Docs features trace directly to this research, including Help Me Create (document generation) and Templates — landing less than 6 months after the first report.

Grounded Docs product strategy, vision, and planning from day one: design sprints organized around specific document types, and a multi-year strategy built on the taxonomy.

Adopted beyond Docs: Workspace teams including Google Meet and Gmail used the taxonomy to envision and prioritize their own 2025 work, fostering a shared focus on end-to-end workflows across Workspace.

Directly informed AI evaluation prompt specs (coverage weighted by real-world document type frequency), template prioritization, and feature motivation — including note-taking and the pageless format.

Grew in strategic importance over time: follow-on phases added more dimensions of quantification, accumulating to roughly 8,000–9,000 survey responses. A v2 — revising the top-level categories and adding a deeper layer of detail — is currently in progress.