Helping researchers discover relevant scientific papers

Helping researchers discover relevant scientific papers

Company

Springer Nature

Role

Senior UX Designer

Duration

8 months

a cellphone leaning against a wall

The challenge

Springer Nature is a global scientific publisher looking for new ways to engage researchers and encourage them to return.

But the growing volume of scientific research made it increasingly difficult for researchers to keep up with papers relevant to their work.

I was part of a team tasked with exploring whether personalised recommendations could help researchers discover relevant scientific papers. The concept had already been proposed, but we didn't yet know whether researchers would find it valuable.

Researchers were already discovering papers through Nature.com. The opportunity was to help them find relevant research beyond the article they were reading.

A different way of working

This was the first team of its kind at Springer Nature. We were working towards a shared outcome rather than simply delivering a list of features. Although personalised recommendations had already been proposed, we had the freedom to challenge the approach and decide how best to test it.

I worked alongside a Product Manager, UX Researcher, Data Scientist and two Engineers. We spent time with researchers at the Francis Crick Institute, understanding how they worked, the challenges they faced and what the data told us about their behaviour.

We each brought different skills, but nobody was confined to their job title. Everyone was involved in research, understanding the data, exploring ideas and testing assumptions.

Testing the idea without building a new product

One of the early proposals was to build a dedicated hub where researchers could access personalised recommendations.

But we didn't yet know whether researchers would find the recommendations useful, so building an entire product felt premature.

I challenged whether we needed a dedicated hub before we'd established that researchers would find the recommendations useful. We agreed to test the concept within their existing journeys first, using Nature.com to reach a large audience.

Nature.com gave us access to a large audience, allowing us to test the concept and learn whether it was worth investing in something bigger.

Based on what you read we could use this information and other data to recommend similar articles.

Designing around how researchers work

Mapping the research journey helped me identify where recommendations might be most useful.

Researchers often started with Google Scholar, arrived at an article and read the abstract before deciding whether to continue.

We explored presenting three recommended articles at this point, allowing researchers to discover related content or sign up to receive further recommendations by email.

I explored different layouts, placements and interactions to make recommendations noticeable without interrupting the reading experience or making them look like advertising.

We used usability testing, engagement data and stakeholder feedback to evaluate the different approaches. This helped us understand how researchers interpreted the recommendations, how they interacted with them and which designs were worth taking forward.

I mapped where recommendations could fit naturally into researchers' existing reading habits.

I explored different placements for recommendations within the article page, balancing visibility with the need to avoid interrupting the reading experience.

When testing challenged our assumptions

We tested the concept with researchers at the Francis Crick Institute to understand whether the recommendations were useful and how people interpreted the experience.

Because I'd used Nature's branding, researchers assumed the recommendations were exclusively from Nature journals.

The whole point of the service was to help people discover relevant research, regardless of publisher.

I went back to the visual identity, exploring how I could make the recommendations feel trustworthy while giving the service an identity of its own.

This led to Recommended, a separate identity designed to work across different channels while remaining connected to Springer Nature.

"These are articles from Nature? Not sure I'd want to just see articles from Nature."

Testing revealed that researchers assumed recommendations only came from Nature journals. I introduced a separate identity to make it clearer that the service recommended papers from across publishers.

I explored a distinct identity for Recommended that could work across different products and channels while retaining a connection to Springer Nature.

Taking recommendations beyond the article page

We also looked at how recommendations could reach researchers beyond the website.

Our data showed that although researchers typically worked on desktop, they also browsed articles on mobile while commuting or travelling.

I designed a mobile-friendly email to deliver personalised recommendations, alongside a landing page that explained how the service worked and encouraged people to sign up.

The aim was to make recommendations useful beyond a single visit to Nature.com.

Personalised recommendations sent as an email for researchers to browse during their commute and when travelling, with a landing page explaining the service.

What changed

The recommendations service reached researchers in over 200 countries, with 480,000 subscribers in the first eight months, representing 6% of nature.com's monthly active users and exceeding our original 4% target.

Building on what we'd learnt

Following the launch, Springer Nature wanted to extend the service into other products. I explored how recommendations could be integrated directly into researchers' existing workflows, including through an inline application.

Ongoing user feedback also helped shape new opportunities, including a plugin that would allow researchers to track and manage their scientific activity. By giving users control over their reading history, including the ability to remove papers, we could use their feedback to improve the relevance of future recommendations.

Additional ideas for how researchers could manage their reading history and improve future recommendations.

What I took away

This was one of my earliest experiences of what I'd now call an empowered product team.

We understood the problem, spent time with researchers, challenged each other's thinking and made decisions together. Everyone brought different expertise, but nobody was limited by their job title.

What made it different was having the freedom to work out how to achieve an outcome, rather than being told what to build.

It's an experience that's stayed with me and continues to shape how I lead design teams today.

Copyright 2025 by Laura Gosney

Copyright 2025 by Laura Gosney

Copyright 2025 by Laura Gosney