The problem I was hired to solve
Before teams can make good product decisions, they need to understand how users actually think about the product — what terms mean to them, where they get stuck, what they value, and what they ignore. This body of work spans three years of that kind of foundational investment: the research that makes everything else faster, sharper, and less likely to be built on wrong assumptions.
What I did
This work spanned three interconnected areas:
Growth and operations
- Identified the aha and habit moments that correlated with long-term retention — giving the team a basis for onboarding and activation decisions.
- Built a value and use-case matrix to clarify what users were actually getting from Sonar, and where the gaps between perceived and delivered value were largest.
- Researched why users opted for manual project creation over automated setup, informing decisions about where to reduce friction in the onboarding flow.
- Evaluated which configuration features should move from project-level to organisation-level settings — a decision with significant implications for enterprise adoption and admin workload.
- Defined metrics for enterprise experience, supported billing implementation, reviewed permission sync behaviour in SonarQube, and explored the migration path from SonarQube Server to SonarQube Cloud from the user's perspective.
Product understanding
- Ran terminology perception studies across products — understanding how key terms landed with different user segments and where language was creating confusion or misaligned expectations.
- Contributed to quantitative UX research infrastructure: clarification work, UX benchmarking, NPS analysis, and survey metric design.
- Led discovery for IDE and Cursor-related workshops, establishing a research foundation for the next generation of IDE experience.
Code quality and user outcomes
- Developed personas tied to code quality workflows, informing how the team thought about segmentation and prioritisation.
- Researched the new-code period workflow, quality gate setup and usage, and how often users actually resolved issues when quality gates failed — a finding that challenged assumptions about how much developers engaged with Sonar's output.
- Mapped what users wanted from Sonar as a solution (not just as a tool), and ran usability evaluation work across core product surfaces.
What I found
The through-line across this work was a consistent gap between how the team modelled user behaviour and how users actually behaved. Key themes:
- Terminology was a recurring barrier. Words that felt precise internally were ambiguous or misleading to users — particularly across the boundary between SonarQube Cloud and SonarQube Server.
- Activation and habit moments were non-obvious. The moments users described as "when I realised the value" often weren't the moments the team had designed for.
- Enterprise controls were being managed at the wrong level. Configuration that belonged at the organisation level was sitting at the project level, creating unnecessary admin overhead and inconsistency at scale.
- Code quality consumption was shallower than assumed. Many users weren't working through Sonar's findings systematically — they were triaging selectively, which had significant implications for how results should be surfaced and prioritised.
The decisions this enabled
- Gave the organisation a system-level understanding of user behaviour — moving beyond local UI questions toward structural decisions about how the product was organised and communicated.
- Established the persona, value-mapping, and metric-definition infrastructure that subsequent squads built on — including the more recent remediation, CLI, and migration work.
- Supported near-term product decisions (onboarding flows, configuration architecture, quality gate design) and the deeper framing work that helped teams decide what to measure, how to describe value, and how to interpret user behaviour across products over time.
- Helped SonarSource move from a research-light to a research-informed organisation — building the capability and the evidence base that made faster, better product decisions possible.