Overview (tl;dr)
What started as requirements gathering research uncovered an institution-wide content management challenge.
The Cambridge Web Platform team asked me to gather requirements for a redesigned staff profile template as part of a Drupal migration. Instead, my research uncovered that staff profile data was fragmented across multiple disconnected systems, creating an institution-wide content management problem that no template redesign could solve.
My role
As Lead User Researcher, I planned and delivered the end-to-end research, from study design and recruitment through to synthesis and communicating recommendations to designers, engineers and stakeholders.
Impact
The findings informed the immediate redesign while giving stakeholders the evidence to distinguish short-term improvements from a much larger strategic challenge.
My Research Flow
Research Intake
My initial chat with the PM on a new piece of work.
Planning & Strategy
Hashing out the brief, meeting stakeholders, signing off on objectives, and recruiting.
Research Execution
Running interviews, testing sessions, and other data collection.
Analysis & Synthesis
Coding, triangulating, and turning raw data into insights.
Knowledge Sharing
This is the decision-making phase, when deliverables and action items are shared with stakeholders.
The Context
The University of Cambridge needed to update and migrate its Drupal content management system (CMS), which was a huge opportunity to redesign the entire Cambridge web ecosystem.
I was part of one the product teams in this Web Ecosystem Programme as the sole lead User Researcher. As part of the CMS migration and design updates, the PM approached me to gather requirements for a CMS content type (template) called Staff Profiles.
The goal was to understand what information different users needed from staff profiles and translate those needs into requirements for a new CMS template, with development due to begin in just two months. The PM also wanted to understand how content management of these profiles currently worked.
The project was made more complex by the University's highly decentralised structure. Individual Schools and Departments managed staff information independently, there was no central list of users or stakeholders, and Colleges managing staff information further acted as their own entities outside of the University.
Confidentiality note
This case study contains only information already publicly announced by the University of Cambridge. Internal artefacts have been recreated or anonymised for portfolio purposes.
Stakeholder Alignment
At kickoff, I aligned with stakeholders across the web team and project leadership to define research goals, success criteria, and decision-making needs. We agreed research should go beyond validating template requirements and explore how profiles were created, managed, and used across the University. The kick-off also helped to confirm the project brief, and thus allowed me to focus on the methodology.
Planning & Strategy
The objective of this project was to identify requirements for the public-facing profile pages on the University's website.
Research questions
- Who even gets to have a profile?
- What information do different audiences need from staff profiles?
- How do staff currently create and maintain profiles?
- Where are the critical pain points and opportunities for improvement?
Target audience
- Profile content creators: web editors, academics, researchers, PhD students
- Profile content consumers: prospective students, collaborators, media, etc.
A successful research outcome would...
...support the PM with determining the best strategy on how to redesign the staff profiles content type, gather design requirements, and prioritise the most critical feature improvements.
Methodology
Desk Research
Reviewing existing documentation and prior research before going any further.
As-Is Profile Audit
Auditing how staff profiles currently work and appear across the University's sites.
Competitor Analysis
Benchmarking how peer universities structure and present their profile pages.
Survey
Open-ended survey to identify needs, pain points and priorities at scale, while recruiting for interviews.
User Interviews
Semi-structured interviews revealing behaviours surveys alone couldn't capture.
Recruitment Strategy
Having the support of School Heads and Secretaries, as well as a survey created enough momentum that we grew interview participation from just 2 to more than 50.
The survey also surfaced stakeholders the project team wasn't previously able to clearly identify. My team initially recommended I reach out to participants through our research pool and internal research boards, but cold outreach only yielded two interview participants. This was unsurprising, given how many emails university staff receive daily.
I shifted strategy and partnered with School Heads and Secretaries, who act as trusted gatekeepers to departmental mailing lists and are selective about what they circulate. Having my research accepted by them carried real weight, as staff would know that only important communications get sent through these channels. Their endorsement meant people were far more receptive to engaging with me.
In order to increase my chances of having my research be accepted for circulation, I explained the value and importance of gathering staff feedback early on in the website redesign as this would be something that would impact everyone. I also created project information sheets, email scripts, and a survey that would double as a recruitment pipeline. The Heads and Secretaries for all but one of the schools accepted to circulate my research and survey to all of their departments.
Tools
Research execution
The secondary research was led by myself, however, I collaborated with Content Designers to collect previous audit results that their team had done, and also worked together with my team's Content Designer to conduct the competitor analysis.
The survey was sent out to almost all departments and we collected more than 300 responses across all the user segments I wanted to engage with. From the survey, more than 50 volunteered to be interviewed.
For user interviews, I conducted 24 remote, semi-structured interviews utilising discussion guides and visual stimulus related to their profiles and that of a competitor. This would help spark discussions. Interviews were recorded with Marvin AI with the user's consent.
One of our faculty members has 8 public profiles...that we know of. We're only able to access one of them, the rest haven't been updated in years. If a journalist or potential funder were to search for her name...well, let's hope they click on the right link.
Analysis & synthesis
Marvin was my primary tool for rapidly synthesising user interview data. Its AI allowed me to quickly update my PM on learnings that were surfacing as the research went on.
Once interviews and testing sessions wrapped, I moved into synthesis. This involved coding transcripts for recurring themes, mapping pain points against user goals, and clustering findings into an affinity diagram to surface patterns across different user groups. I used Marvin's AI summariser to sense-check my own analysis and predictably, it consistently missed insights that mapped directly back to the research objectives, which gave me more confidence in what I'd found manually. These themes were then triangulated against learnings from stakeholders and the survey data.
I'm sometimes asked how I triangulate so many different types of datapoints, and honestly, it's all about mapping these back to the original research objectives and questions. Each relevant datapoint fits together around the objective like puzzle pieces, and together they give the full picture, a clear evidence-based story. Which is why that initial research brief is so so important!
Insights & Recommendations
From the secondary research
Insights from the secondary research pointed us in the direction of a potential solution called Symplectic Elements that was being used by other universities.
From the survey
The survey pointed out that the profiles were not meeting user needs.
From the user interviews
I identified needs from users, who their target audiences were, where their pain points occurred, and how they managed profiles.
Synthesised insights and recommendations
Short-term recommendations
For the immediate requirements of the CMS migration and content type creation, I recommended that the team not spend any time re-defining content management workflows for staff profiles, as the issue was much larger than what could be solved within our constraints. Instead, I provided that team with:
- Requirements for content fields, their definition, format, and the matching user needs these solve.
- Mapped data sources and ownership so our engineers understood where information originated from (such as publications)
- Standardised content presentation and structure for a more consistent profile experience across departments
- And as a few survey respondents said, "If you just make it prettier, that would already be a big win"
Long-term strategy
I identified that staff profiles were an integral part of being a member or affiliate of the University. Profiles provide visibility, accreditation, legitimacy, networking, opportunities, and much more. It was important that if the University was actually going to create a lasting solution for profiles, it would need more resources and time than what the CMS migration programme had. It would also require stakeholder collaboration across all of the Schools and its departments. Ultimately, what users needed was:
- A single source of truth for staff information, one for the public and one internally
- Improved discoverability of expertise, so that people can connect and collaborate on overlapping areas of work
- Introduce improved systems for content governance, management, and review processes
Reflections & Learnings
This project reinforced the importance of looking beyond the immediate design request to understand the wider systems behind user problems. It also challenged me to find the right balance between shipping quick fixes worth the engineering/design time and holding out for more foundational fixes down the line.
I also want to acknowledge that the scope for this project was too big. I was speaking to far too many user groups and in the end, it would've been just as effective to focus on stakeholders, faculty, and web editors, without also speaking to PhD students/prospects and Research Fellows. I'm glad I did on a personal level, as I learned a lot of interesting things about their experiences, but nothing that would have made any substantial impact to our research objectives and what my team had the resource to achieve.
In the future, I would push back and persuade stakeholders to focus on an initial subset of primary users groups to see what we uncover before tackling secondary user groups. It would have saved me significant time and energy.
How I'd approach this today
Advances in AI have significantly changed how I approach research. While we already used Hey Marvin's AI for transcription and thematic analysis at the time, I'd now bring AI in much earlier to accelerate research operations, use agents/skills to run secondary user research, and create more engaging deliverables faster.
(In fact, I'm now building an AI-powered research operations tool of my own, focused on research planning and secondary research.)