Why UK Universities Partnered with Explorance to Solve NSS Comment Analysis

June 2026 saw the launch of the Explorance MLY National Student Survey (NSS) model, specifically designed to deliver the most accurate, fully automated analysis of thousands of comments for 10 main NSS theme categories.

MLY sets itself apart by providing transparent, traceable processing, and categorisation, offering users comprehensive access to every insight through a fully interactive, user-friendly dashboard experience. No other solution matches this level of specificity, transparency and interactive analytics for NSS data in higher education.

The scale of the NSS is remarkable. Conducted annually across the UK, the survey gathers feedback from final-year undergraduate students about their university experience. Since its launch almost 20 years ago, more than 5.5 million students have shared their views.

In 2025 alone, over 500 higher education providers participated, with approximately 357,000 undergraduate students responding, which amounts to a response rate of nearly 72%. Each year, this generates an enormous volume of qualitative feedback.

Delegates from across the UK and beyond joined Explorance for Transforming NSS Insights: University Collaboration Behind the MLY NSS Model, a panel discussion featuring Explorance MLY customers who played a central role in shaping the model. The session explored practical approaches to interpreting NSS open-text comments and demonstrated how MLY helps institutions transform student feedback into meaningful action.

With NSS 2026 results due to be published soon, now is an ideal time for institutions to prepare to act quickly on the insights they receive. We spoke with representatives from three institutions involved in developing the MLY NSS Model:

• Dr Natalie Holland, Research Officer, Liverpool John Moores University
• Lorna Robertson, Student Surveys Officer, University of Strathclyde
• Ufuoma Elegbede, Analyst, Strategy and Insights, University of Nottingham

Here are some of the highlights from their discussion:

Challenges in Analysing NSS Open-Text Responses

The NSS generates thousands of open-text responses each year. Panelists discussed the main obstacles they faced when extracting actionable insights from this volume of qualitative data and how their institutions previously approached comment analysis.

Natalie: "We have been using software to support qualitative analysis for several years. However, one challenge with previous solutions was that, while they were excellent at processing large volumes of text, they were not specifically designed for student feedback. The outputs often required significant interpretation.

We spent considerable time reviewing the themes generated by the software and then returning to the underlying comments to understand how those themes had been derived. Ultimately, the biggest challenge was time. Even with specialist text-analysis software, analysing a large NSS dataset still required substantial manual effort.

At the same time, there is considerable institutional pressure to understand the results quickly. Senior leaders, faculties and departments all want to know what students are saying as soon as possible - particularly when results are less positive and action planning needs to begin immediately. For us, speed and efficiency were the key challenges."

Lorna: "Our biggest challenge was resource capacity. We needed to produce high-quality analysis while also reporting results in a timely manner. NSS data has a very high profile within the institution and is subject to significant scrutiny.

Historically, our approach was highly manual. Team members reviewed and coded every individual comment themselves. As you can imagine, that took a considerable amount of time. As a result, our qualitative reporting often lagged several weeks behind our quantitative reporting. While stakeholders could quickly access scores and survey metrics, the richer insights contained within the free-text comments arrived much later.

That was far from ideal because colleagues wanted immediate insight into what students were saying and what actions should be taken. Introducing MLY was already a significant improvement for us last year, but the customised NSS model has taken that process to another level."

Ufuoma: "Historically, we worked with the comments exactly as they were received from Ipsos, which conducts the NSS on behalf of the Office for Students and the UK higher education funding bodies. Typically, comments were divided into two broad categories: positive and negative. They were then made available through a Tableau dashboard, where users could apply filters or search for keywords.

For example, someone could search for 'library' and see every comment containing that word. However, we identified two significant limitations with this approach.

First, student feedback is not always binary. A comment might say, 'The course was excellent, but feedback on assessments was often delayed.' In that example, the student is expressing both positive and negative sentiment. When comments are forced into a single category, valuable insight is lost.

Second, keyword searching is inherently reactive. You need to know what you are looking for before you start searching. If students raise an important issue using language you did not anticipate, there is a risk that insight will be missed.

One of the things we quickly realised when using MLY for other institutional surveys was the importance of connecting qualitative and quantitative data. We wanted to understand not only what the scores were telling us, but also why students were giving those scores. Our previous approach made that difficult."

Why Institutions Chose to Collaborate on the MLY NSS Model

Each institution had different reasons for partnering with Explorance on the AI-driven MLY solution for NSS qualitative analysis. The panelists explained what motivated their involvement and the advantages MLY offered over previous approaches.

Ufuoma: "We began using MLY across other institutional surveys during 2025, so applying it to the NSS felt like a natural next step. When Explorance approached us about collaborating on the development of a customised NSS model, we saw it as an opportunity rather than a risk.

We had already gained experience with MLY and developed a strong understanding of its strengths. Participating in the project allowed us to share our experiences, challenges and requirements directly with the development team.

One aspect we particularly valued was the collaborative nature of the process - not only with Explorance but also with other universities. The user-group discussions and shared learning opportunities were extremely valuable. Being able to contribute to a solution that could improve how institutions analyse student feedback across the sector was something we were very keen to support."

Natalie: "For the previous two years, we had been developing our own customised model internally and had found it extremely valuable. However, it was also a time-consuming process.

One of the biggest challenges was understanding how existing MLY themes had been developed and then determining how best to map those themes to NSS terminology and reporting structures. The opportunity to collaborate with other institutions and tackle that challenge collectively was incredibly appealing.

It allowed us to explore different perspectives and identify the most effective ways of structuring and interpreting NSS feedback. One of the most exciting opportunities was the potential for greater sector-wide consistency.

If institutions can begin with a common framework for analysing NSS comments - while retaining the flexibility to customise it for their own needs - it opens up opportunities for wider collaboration and potentially benchmarking. Over time, institutions may be able to compare themes emerging from their own student feedback with national trends and sector-wide patterns. For us, that longer-term potential was just as valuable as the immediate benefits of the model itself."

Lorna: "Collaboration was a major factor for us. Prior to this project, I had already worked with colleagues to create a customised model specifically for Strathclyde. Through that process, I gained valuable experience and insight into how MLY themes could be mapped against NSS themes.

I was keen to share that experience while also learning from others. I was also interested in seeing how closely our internally developed model aligned with the broader Explorance approach.

One of the most challenging aspects of creating a customised model is the mapping process itself. Deciding how MLY's extensive set of themes should align with NSS themes requires time and careful consideration.

Fortunately, MLY makes the technical side of customisation relatively straightforward. Themes can be renamed, reorganised and restructured through a simple drag-and-drop interface. Although the mapping process requires upfront investment, it is absolutely worthwhile because it saves a huge amount of time in the long term."

How the Custom NSS Model Changes Qualitative Analysis

The development of the Explorance MLY NSS custom model introduces new efficiencies and capabilities. Panelists described how the model will change their approach when data maps directly to core NSS qualitative themes and when they plan to use it.

Lorna: "I would not necessarily say that our overall approach will change dramatically, but the customised model will make the reporting process far more efficient.

"The biggest benefit is that it removes a large amount of manual coding because the themes have already been mapped and structured within the model. We will be using it during our upcoming NSS reporting cycle - which is surprisingly close now, given that results are due next month.

We are also looking to make greater use of MLY's additional functionality, particularly the summary features that identify key recommendations, areas for improvement and institutional strengths.

These features help us surface important insights from large volumes of free-text comments that we simply did not have the resources to analyse previously. As a result, we will be able to integrate richer qualitative insights into our reporting and provide more meaningful information to stakeholders across the University.

I would add one important caveat: MLY's summaries provide an excellent starting point, but they should always be sense-checked by a human reviewer. Institutions will often want to incorporate their own terminology and contextual understanding before sharing findings more widely. Even with that caveat, it significantly enhances our reporting capability.

We are also considering using the NSS model as the foundation for customised models for other institutional surveys in the future."

Ufuoma: "For us, this represents a genuinely transformational change. As I mentioned earlier, our previous approach relied heavily on keyword searching, which meant insight discovery was often reactive. With MLY, we are moving towards proactive discovery.

Instead of having to know what we are looking for in advance, the platform helps identify themes and patterns automatically, allowing us to uncover insights that might otherwise remain hidden.

Building this capability manually would require weeks of coding and analysis. MLY enables us to move from a highly manual process to one where insights can be generated much more quickly.

We are particularly excited about features such as recommendations. It is one thing to categorise comments and identify themes, but being able to surface clear recommendations gives senior leaders a much faster route from data to action.

For example, they can immediately see suggestions around what students would like the institution to start doing, change or stop doing. That ability to move rapidly from data to insight and then to action is extremely valuable.

We are also keen to explore the relationship between quantitative and qualitative data. While it is not always possible to create a direct one-to-one relationship between survey scores and individual comments, qualitative feedback can certainly help explain why particular outcomes are occurring.

If we identify areas of strong performance - or areas where performance is lower than expected - we can examine the associated themes and comments to better understand the underlying drivers. That is something we are very excited about moving forward."

Natalie: "One of the original reasons we adopted MLY was its recommendations functionality. The platform does not simply analyse comments and provide thematic and sentiment analysis; it also identifies the recommendations students are making.

That saves a significant amount of time and helps us focus quickly on the issues students consider most important.

In terms of the customised NSS model specifically, one of the biggest benefits is consistency of language. When reporting quantitative NSS results, we already use established terminology and frameworks. We wanted to avoid presenting qualitative analysis using completely different terminology that might confuse stakeholders.

The customised NSS model addresses that challenge by aligning qualitative analysis with the language already used across NSS reporting. As a result, we can share findings more confidently because stakeholders can clearly see the connections between quantitative metrics and qualitative themes. That consistency improves understanding and helps ensure insights are more accessible and actionable."

Advice for Institutions Considering the MLY NSS Model

For institutions not yet using the Explorance MLY NSS model, adoption offers clear benefits. The panelists shared their recommendations for peers looking to get the most out of the platform.

Natalie: "If you are already using MLY to analyse NSS comments, then for me it is a straightforward decision. The model aligns terminology, structures themes in a way that reflects NSS reporting, and enables richer analysis using language that stakeholders already understand. It simply makes the outputs more meaningful and easier to communicate.

One additional tip relates to filtering and segmentation. One of MLY's greatest strengths is its ability to analyse feedback across different student groups. Of course, you can filter by department, school or faculty, but we also use it extensively for demographic analysis.

This allows us to examine how students from different backgrounds experience their education and identify variations in experience that might otherwise remain hidden. That kind of analysis can directly support Access and Participation Plans, equality and inclusion initiatives, Teaching Excellence Framework (TEF) submissions, strategic planning and other priorities.

Like any AI-based solution, MLY continues to improve over time. The more institutions use it, the more robust it becomes.

What MLY does exceptionally well is remove the most time-consuming stage of the process. Instead of spending weeks categorising comments, researchers can focus their time on interpretation and understanding. In practical terms, it probably reduces the overall workload by around half."

Lorna: "One feature I particularly value is MLY's sharing capability. Once you have generated rich insights from your data, one of the ongoing challenges is ensuring those insights reach the right people.

Historically, that process could be cumbersome. MLY makes it much easier because you can create user groups and control access to information. For example, if you want a particular faculty to see only its own results and insights, you can configure access accordingly.

This creates a much more efficient process for distributing information and ensures colleagues can engage directly with the data most relevant to them. At Strathclyde, we are still exploring how best to use these capabilities, but it is certainly an area we are keen to develop further.

Being able to share dashboards directly with stakeholders has enormous potential for improving engagement and accelerating decision-making.

I would, however, offer a word of caution. As with any AI-powered solution, it is important to recognise that not every comment will necessarily be categorised automatically. There will always be some limitations, and institutions should be prepared to undertake additional review where necessary. It is important to understand both the strengths and limitations of the technology."

Ufuoma: "One of the first things I would recommend is speaking with the Explorance team. Throughout our experience, we have found them extremely open to discussion, collaboration and feedback. If you have questions about the platform, its functionality or how it might work within your institution, it is well worth having those conversations.

I would also encourage people to explore the resources available through the platform and wider Explorance community. There are webinars, case studies, user stories and examples from institutions already using MLY. These provide a useful understanding of what to expect before implementation.

Ultimately, though, the most important thing is remembering why we are doing this. Our goal is not simply to categorise comments. Our goal is to improve the student experience.

The technology helps us move from raw feedback to actionable insight, but the real value comes from how institutions use those insights to drive improvement. And I would echo Lorna's earlier point: while the platform does a lot of the analytical heavy lifting, institutions should still review the outputs, sense-check the findings and adapt them where appropriate.

The flexibility is there - you can adjust themes, make changes and tailor the model to your own context. It is not a one-size-fits-all solution."