UX Research Ops Tree Testing
Self initiated

Studier

A lightweight tree testing platform built in three days to protect a live public-service research timeline.

Studier was created during a sensitive public-service navigation project, when the team needed to run a tree test within the same week. Existing options did not fit the project constraints. The required Qualtrics module was unavailable, external procurement would take too long, and third party platforms raised privacy and security concerns.

I built the first working in three days. It supported the essential workflow for : create a study, set up an IA tree, add tasks, publish a participant link and collect responses. The platform later expanded into a reusable research operations tool with preview, lifecycle controls, custom questions, exports, data cleaning support, guide content and version history.

ClientInternal programme tool
RoleProduct Owner, UX Researcher, AI-assisted Product Builder
TeamIndividual build, used by programme members
Timeline3-day MVP, followed by iterative releases
ToolsCopilot, React, Supabase, GitHub, Vercel
HENEX Lens R27 · Impact Dashboard Agency × Meaning × Interface
Skills
Product ownership Research operations Tree testing Information architecture AI-assisted development Dashboard design Data analysis Privacy-aware research

Background

A research timeline with no safe off-the-shelf option.

A sensitive public-service navigation project needed a tree test within the same week to validate a proposed information structure. The team's usual Qualtrics module was unavailable, external procurement would take too long to meet the deadline, and third-party research platforms raised privacy and data-handling concerns for content that touched vulnerable users.

Problem

How might I create a safe, fast and customisable tree testing workflow that captures enough behavioural evidence for IA decisions without delaying the project?

The project team needed evidence about whether people could find support information in a proposed navigation structure. A standard survey could not capture enough behavioural evidence. The research needed selected paths, , click history, time, click count, skipped tasks and acceptable path logic.

Solution

A controlled workflow for tree testing.

Studier turned tree testing into a controlled research workflow. It was designed around the smallest useful flow, avoiding unnecessary personal data while still capturing the behavioural evidence needed for IA decisions.

Create and publish tests

Set up studies, IA trees, task scenarios, target paths, acceptable paths and public participant links.

Run participant flows

Guide participants through intro, privacy guidance, task-by-task navigation and selected path submission.

Export and analyse responses

Capture path choices, match types, first clicks, time, click count and outputs for data cleaning.

Outcome

A project risk became working infrastructure.

Studier helped the team run the tree test on time, reduce dependency on unavailable tools and create a controlled environment for privacy-sensitive research.

Saved time

Protected the research schedule.

The team could run the tree test within the required week instead of waiting for procurement or adapting an unsuitable tool.

Protected data control

Reduced tool and privacy risk.

The platform avoided unnecessary personal data and introduced clear responsible-use conditions before access.

Made research repeatable

Supported wider programme use.

Other programme members could register, create studies and manage tests from the dashboard.

The finished dashboard: study status, links and response management visible in one place.
Open Studier Platform

Public access starts from a registration screen.

Contribution

I owned the tool, the test workflow and the analysis.

I owned the full process independently, from product framing and MVP scoping to Copilot-assisted build direction, internal testing, research setup, response management and analysis.

Product ownership

Define the useful minimum

I shaped the product need, scoped the first MVP and kept each feature tied to a real research or operational problem.

AI-assisted build

Build through small changes

I used Copilot to move quickly, then tested each generated change against the intended workflow.

Research operations

Set up the test workflow

I created the tree test structure, target paths, acceptable paths and response handling process.

Data analysis

Turn outputs into findings

I cleaned, validated and analysed the exported data so the research could support IA decisions.

Design

One platform for setup, testing and response management.

The platform was built around the core research workflow, then expanded where real use showed a need.

Study dashboard

Create, review and manage studies.

The dashboard lets internal users check status, copy participant links and manage the test lifecycle from one place.

Study builder

Set up custom trees, tasks and questions.

The builder supports study instructions, privacy notes, IA tree setup, task scenarios, target paths, acceptable paths, pre-task questions and final questions.

Responsible use gate

Set clear use conditions before access.

Before users enter the platform, Studier presents clear use conditions for responsible internal testing. The screen sets expectations around approval, privacy, sensitive information and checking exported data before sharing.

Develop

From a three day MVP to a reusable research tool.

Studier did not grow from a fixed roadmap. It grew from real use. Each update responded to a specific research or operational need, tracked across the project's own development log.

Core workflow

MVP workflow

Create study, IA tree, tasks, participant runner and response capture.

Study setup

Study setup maturity

Add pre-task questions, CSV checks, preview and readiness checks.

Task navigation

Locked review mode

Added a locked review state so completed studies could not be edited mid-analysis, keeping response data consistent.

Lifecycle

Lifecycle and review

Add closing time, review mode, task navigation and response path review.

Data cleaning

Republish workflow

Introduced a republish path so a study could be corrected and reopened without losing existing responses.

Data operations

Data operations

Add exports, data cleaning, clear data options, guide content and version history.

Validation

Testing whether people could find support information.

Studier was used to support a public-service tree test across ten support information scenarios. The analysis helped identify IA issues around category boundaries, support pathways, process labels and safety-related content.

Participants35
Cleaned task responses323
Tasks10
Total success50.5%
Influence

One clearer route performed strongly and became a useful benchmark for improving labels and navigation paths.

Delivery

Participant flow, guide content and version history.

The final release made the tool usable beyond a single project, with a public participant flow and internal reference material.

Participant flow

Complete tasks through a collapsed navigation tree.

Participants can open a public link, read the introduction, complete tasks and submit selected paths without logging in.

Guide and internal use

Make the tool usable beyond one project.

Guide content and version history helped the platform become a usable internal tool rather than a one-off prototype.

Open Studier Platform

Public access starts from a registration screen.

Reflection

Build only what the delivery risk needs.

Value

A focused internal tool protected the research timeline and turned IA testing into a repeatable workflow.

Learned

The platform was built under pressure, so advanced reporting and dashboard visualisation stayed outside the MVP.

Next

Add testing rounds, stronger visualisations, wrong path clustering and exportable findings reports.

Primary recipe R27 · Impact Dashboard

Agency Meaning Interface

Studier uses R27 because the project turned a time critical research process into an actionable internal dashboard. It made study status, participant links, response capture, lifecycle controls, exports and IA evidence visible enough for the team to manage progress and act on findings.

Secondary recipe