Talks 2026

#GamesUR Conference Talk Descriptions

Did the Player Survive the Pipeline? Four Checks Before an AI-Assisted Finding Reaches the Roadmap

Presenter: Pietro Guardini

One player says, “This boss is impossible.” An AI co-moderator asks a follow-up. Speech-to-text standardises the wording. A local LLM codes frustration; an evidence-retrieval tool links the comment to deaths and retry behaviour; several agents propose competing explanations. A final model concludes: “Players are frustrated by the boss encounter. Reduce the boss’s health and damage.”
Every hand-off looks reasonable. But did the player’s experience survive the pipeline – or did small transformations turn ambiguity into false certainty?

AI discussions in Games User Research often focus on a single task: transcription, probing, coding, or synthesis. This talk asks what happens when those tools form a chain. Drawing on academic research and GUR Summit talks from 2010 to 2025, I follow one playtest moment from session to strategy. The contribution is a practical audit for checking whether an AI-assisted claim remains trustworthy. I reviewed studies on AI-assisted research to identify twelve pressure points. Each was mapped against GUR Summit talks from 2010 to 2025 to identify relevant methods, examples, counterexamples, and unresolved questions. The result is a four-part, evidence-informed framework teams can test and adapt. The talk ends by applying all four checks, leading to the revised recommendation: “keep the challenge, improve the clarity of Attack C, and test changes in the next build.” Attendees leave with a one-page 4R Audit for classifying AI outputs as signals, hypotheses needing further investigation, or findings ready to influence the game.

Presenter

Looking into the Starry Night: Two Risk Assessment Tools

Presenter: Michal Mycka

This talk introduces two methods from risk management practice that can be easily adopted by Games User Researchers. These methods provide a structured way to challenge team assumptions, identify blind spots, and support better decisions earlier in the development process. Although widely used in risk management and strategic planning, these techniques remain largely absent in GUR. While typical GUR methods help us understand player experience as is, risk management tools complement them, by supporting teams in exploring uncertainties and risks before they materialise in player experience. The talk will introduce the foundations of both methods, discuss when they are most useful within the game development lifecycle, and illustrate them with practical examples. Each method will be supported with best practices, typical challenges and limitations, together with pitfalls to avoid.

Takeaways:
– Being better prepared to explore player experience risks before they surface in playtests.
– Adding two risk assessment tools to your GUR toolkit.
– Translating existing GUR skills into proactive risk assessment.
– Supporting your teams in making better decisions when the future is uncertain.

Presenter

The rollercoaster ride of being made redundant

Moderator: Emma Varjo

Roundtable (For In-person attendees only and NOT included in the live stream pass)

Losing a job is more than a professional setback: it’s a deeply emotional experience. Even when the signs are there, redundancy is a shock, leaving many to navigate feelings of grief, anger, and self-doubt before reaching clarity and acceptance. This roundtable will explore redundancy through the lens of the 7 stages of grief, reframing them into key phases many experience: the initial shock and denial, the anger and bargaining that follows, the guilt and self-doubt that often take hold, and ultimately, the reflection, testing, and eventual excitement for new opportunities.

Through shared stories and open discussion, we’ll explore questions such as:

• What emotional responses are common, and what kind of support do newly redundant employees need?
• How do you manage the emotional toll of rejection and self-doubt?
• How can you reframe redundancy as an opportunity for growth?

While we will acknowledge the emotional impact, the heart of this discussion will focus on the Testing & Reflection stage—the phase where people have the most agency. We’ll talk about practical ways to redefine career goals, network effectively, and regain confidence in the job search. Whether you’ve been through redundancy, are currently experiencing it, or want to support others, this roundtable provides a space to share, reflect, and gain new perspectives.

Moderator

Between Metrics and Meaning: What mobile and PC/console research can learn from one another

Presenter: Jennifer Klatt

What mobile and PC/console research can learn from one another

Mobile and PC/console studios treat player research as two different disciplines, and their different production cycles, business models, and player expectations have genuinely pushed each toward different strengths. Mobile teams are fast, data-rich, and decisive. PC/console teams go deeper on emotion, motivation, and the felt experience of play. Both are also blind in the other’s strong spot. Mobile teams can drown in telemetry data and still not know why players churn, while PC/console teams can run beautiful qual and still be slow to act.

This talk argues that the best player research isn’t metrics or meaning. It’s the deliberate combination of both, and each industry already holds half the answer. These differences can be observed in PlaytestClouds usage data. A deep dive into differences in playtest patterns is presented before the advantages of each approach are discussed.

We’ll look at three exchanges areas where PC and Console research tends to operate differently:
–  Understanding players. Why PC/console research often reaches motivation and appreciation that engagement metrics miss, and how mobile teams can get there without slowing down.
–  Behaviour meets experience. How mobile pairs more commonly playtesting with telemetry, retention, and progression, and where behavioural data alone falls short.
–  Speed and decisiveness. How mobile’s early concept testing and clear kill/go calls can de-risk traditional milestone pipelines.
–  Continuous vs milestone research. The trade-offs of embedded, ongoing research versus large one-off studies at key milestones like alpha or beta.

Attendees leave with practical, immediately usable ideas: introducing earlier, lower-cost research into traditional pipelines; combining qual insight with telemetry for stronger calls and moving past engagement metrics to real motivation. Whether you work in mobile, indie, or AAA, the future of player research is combining the best of both, to build games players don’t just come back to, but genuinely care about.”

Presenter

Cracking the Beta Insight Code: How to Capture Meaningful and Representative Player Feedback

Presenter: Michael Cai

Most game studios run closed‑beta and open‑beta programs, yet few have solved a core challenge critical to researchers: recruiting players who truly represent their core and opportunity target audiences and engaging them in meaningful research activities such as post‑play surveys, focus groups, and in‑depth interviews. Even studios that attempt to integrate insight work into beta periods often encounter extremely low response rates, inconsistent participation, and uncertainty about whether the feedback they do receive reflects the target player base. As a result, a major opportunity for learning is frequently lost. Teams walk away with mountains of behavioral telemetry—rich data about what players did—but very little understanding of the motivations, expectations, and decision-making processes that explain why players behaved the way they did. For a game that spent tens or hundreds of millions in development costs and hundreds of thousands in in-lab playtest, the inability to leverage beta programs for insights is very frustrating for curious researchers.

Interpret set out to address this challenge after hearing it repeatedly from key clients across multiple genres and platforms. We created the Early Access methodology a few years ago, which has helped quite a few AAA and AA studios transform their beta programs into powerful insight engines. By capturing representative player feedback, uncovering underlying motivations, and contextualizing behavioral data with qualitative depth, we provide executives and development teams with the evidence they need to refine and polish their games before launch. This approach has proven effective across different game types, audience segments, and development timelines.

In this session, we will share how this methodology works in practice, including detailed design principles, operational best practices, and key lessons learned from real-world implementations. We will also present multiple case studies demonstrating the flexibility and impact of this approach, including one where our team produced a comprehensive report based on telemetry data, survey data, and focus group insights, within just one week of the closed‑beta program’s conclusion. Ultimately, we hope this talk inspires other researchers to explore more effective ways of leveraging beta programs for player insights and sparks meaningful discussions that help advance our collective methods and processes.

Presenter

Outnumbered but not outmatched: Lessons from scaling research in a busy studio

Presenters: Gaston Robert & Tristan Dufour

The Core Challenge :
In a 350+ person studio managing multiple concurrent productions, User Research (UR) often defaults to a bottleneck. When we, only two researchers, faces hundreds of developers, the traditional “test-and-report” model becomes unsustainable. This talk addresses the fundamental tension of how to keep player experience a daily priority rather than an occasional checkpoint when you are structurally outnumbered. We will present our transition from reactive testing to a proactive, embedded, and highly scalable research practice.

Expected Takeaways :
Attendees will gain a toolkit for immediate application, including:
•  Operational Frameworks: How to build research roadmaps that align with complex multi-production schedules.
•  Technical Credibility: Strategies for integrating UR deeper into the development pipeline, ensuring feedback is always actionable.
•  Cultural Advocacy: Proven methods for building a “player-first” mindset across a large studio, even when researchers lack formal organizational authority.
•  Optimized Workflows: Practical hacks for reducing reporting overhead and increasing the “per-hour” impact of small, lean research teams.

Presenters

Impact Is Bad, Actually: Rethinking Success in Games User Research

Presenter: Michael Berthin

Games User Research often talks about impact as its ultimate goal. We want our work to help teams create better games and better player experiences. But what exactly do we mean by impact? And are the things we most easily recognize as “impact” always aligned with the purpose of our work? This talk explores the tension between visible impact and meaningful impact.

Attendees will leave with a broader perspective on evaluating research impact and value, and with questions they can apply to their own practice:

* Are we measuring the things that matter, or only the things that are easiest to see?
* How should we think about research contributions when teams start from very different places?
* How do we recognize research that changes understanding, not just immediate actions?

Rather than proposing another framework or metric, this talk invites the audience to reconsider what Games User Research is ultimately trying to accomplish—and whether our current ways of thinking about impact always help us get there.

Presenter

Being Somebody – A Framework for Understanding Player Customization Motivations

Presenter: Elizabeth Zelle

Systems that allow players to customize their characters in games are common, but often game teams struggle to understand WHY players engage in customization or HOW to prioritize development work related to customization. This talk is intended to share my learnings from working with multiple game teams over the last 15 years on how User Research can inform and guide our devs to better understand how to meet the needs of the most players with customization systems.

I first gave a talk on the topic of player customization as a speaker for IDGA Chicago in 2016 and have spent the last 10 years iterating on that to come up with a framework built on psychological motivations that helps game teams better understand why their players customize, breaks down why different demographics of players engage with customization differently, and gives game teams the language to discuss internally their goals around customization even when UR isn’t in the room. This isn’t an academic model, but instead a simplified breakdown of the existing research packaged specifically for the needs of game development teams.

This talk will share that framework, with the intent to 1) provide the community with a recorded talk that both researchers and stakeholders can use for reference in the future and 2) provide other user researchers a foundation to build off of and adapt to the unique needs of their own game teams. This talk will also tackle one of the most common pushbacks to advocating for robust customization systems, which is that our “core audience” doesn’t care about customization and thus making the ROI low.

Presenter

Research That Does Not Stop at Delivery: collaboration lessons from product research in the World of Warships Franchise

Presenter: Nadezhda Ivanova

In live game development, research impact rarely depends only on study quality. It also depends on how well the research team is embedded into workflows, how independently it can operate, how clearly it collaborates with stakeholders, and how consistently findings are supported after delivery. This talk shares teamwork and collaboration lessons from leading the Product Research team within the World of Warships franchise. Rather than focusing on a single study, it explores an operating model that helps research move faster, reduce bottlenecks, and become a stronger partner for development, live ops, publishing, and product teams.

The core idea is that research should not function as a service desk that receives requests, runs studies, delivers reports, and moves on. In live games, where decisions are continuous, cross-functional, and time-sensitive, research needs to work more like product infrastructure: connected to roadmaps, able to move independently, and designed to keep insights usable. Using examples from World of Warships, the talk will focus on five collaboration lessons that help research become part of the product system rather than an external support function.

The broader takeaway is that impactful game user research is not only about running good studies. It is about building an infrastructure that allows research to be trusted, used, and acted on faster.

Presenter

The Research Your AI Isn’t Using: Building a Repository People and Machines Can Read

Presenter: Sohvi Silius

Whatever people make with AI, whether a plan, a spec, or a design decision, is only as good as the context behind it. User research is one of the most valuable kinds of context a company owns. It’s also one of the easiest to waste: scattered across old decks, hard to search, or with no central home at all. At Yousician, a Helsinki-based company behind several gamified music-learning apps, our research used to live in an Airtable index while the reports themselves were scattered across Google Drive, Slides, and Figma. Over a few months, the two of us on research rebuilt it as an AI-readable repository in Confluence and added a company-wide Claude skill, find-user-research, that anyone can use to ask about our users. Instead of pinging a researcher or digging through decks, stakeholders get an answer grounded in real studies. I’ll cover why a searchable repository is worth having in the first place, how we built and structured ours, and how people across the company use it now.

Attendees will learn:
– Why a research repository is worth building, and how to structure one so both people and AI can read it.
– The specific mistakes we made, so they can skip them.
– A sharper answer to a question the field is facing: what is research for when AI is becoming part of how companies operate?

If people use AI to ideate, write, and plan, existing research should be in that context. That doesn’t replace new research, but getting the right insight into the decision may be one of the highest-leverage things research can do now.

Presenter

No One Is Coming to Plan Your Career. So Let’s Do It Together.

Moderator: Emma Varjo

Roundtable (For In-person attendees only and NOT included in the live stream pass)

In game development, career growth is often informal, inconsistent, and rarely explained. Many developers do great work yet feel stuck, unsure how to grow, specialize, or change direction — especially when expectations are unclear or support is limited. This roundtable is a peer discussion about taking ownership of career growth in the game industry. We’ll share experiences around skill development, visibility, feedback, lateral moves, and saying “yes” or “no” to opportunities. The goal is not to offer a single path, but to surface practical insights and mental models that help developers make more intentional career choices.

Presenter

Is it interesting, or is it useful? The experimentation mindset is rooted in good research questions

Presenter: Daniele Borghini and John Dickens

In a development climate defined by efficiency demands, AI acceleration, and shrinking resources, teams struggle to take creative risks. This talk argues that an experimentation mindset, grounded in well crafted Research Questions (RQs), is the most reliable way to de risk innovation and keep player insight at the center of decision making. We move beyond vague testing (the dreaded “is it fun?”) toward useful data that directly guides design choices, for example “does this iteration better embody our creative pillars?”

Key information we will share:

– Why RQs are the secret sauce: how they link research to the most pressing decisions, prevent scope creep, surface misaligned stakeholder assumptions, and help pick the right methodology.
– Real world examples from Dune: Awakening and an unreleased Tencent title, where precise, pillar anchored UX metrics turned design intent into measurable, trackable signals (e.g. “do players know when a failed hit is their mistake, not the game’s?”).
– A practical four step process to get started: intake, dialogue, RQ craft, and results, replacing the broken “survey dropped on a desk” habit with a true strategic partnership.

Expected takeaways:
Audiences will leave able to decode conflicting stakeholder needs into prioritised RQs, anchor research to their game’s design pillars, and activate an experimentation mindset where upcoming decisions are tied to relevant user feedback. UXR stops being a moment of scrutiny and becomes a trusted strategic partner. We will sell you on a simple shift that makes research not just interesting, but genuinely useful.

Presenters

What Makes a Good Battle Pass? Evidence from World of Warships Players

Presenter: Dmitrii Sokolov

Battle passes are now a standard live-ops mechanic, but teams often evaluate them through business outcomes first: uptake, completion, revenue, and retention. This talk starts from a different question: what makes a battle pass feel good to players?
Using survey data from World of Warships players across three update cycles, I will show how players evaluated four event passes across recurring experience dimensions: free-track rewards, paid-track rewards, visual presentation, and progression pace. Across more than 12,000 complete player evaluations, these components explain a large share of overall battle pass satisfaction, but they do not contribute equally.
Attendees will leave with a concrete example of how to turn recurring live-ops survey blocks into a reusable measurement model, how to identify which parts of a battle pass most influence player evaluation, and how player research can support healthier monetization design decisions. The broader takeaway is that a “good” battle pass is not just one that keeps players busy. It is one where rewards, effort, and presentation combine into a value proposition players can recognize as fair.

Presenter

Leveling Up: From User Research to User Science

Presenter: Navot Naor

Games user research needs to grow up. My argument in this talk is simple and, I think, overdue: we should stop thinking of ourselves as user researchers and start becoming user scientists. That is not a change of title, it is a change of standard. Research solves a single problem and gets archived. Science builds on itself, holds up to scrutiny, and predicts. As long as we run one-off studies that inform one decision and then disappear, we will stay a service function. The moment our knowledge starts to accumulate, we become a discipline. The thing standing between us and that shift is rigor. Knowledge only compounds when every finding can bear the weight of the next one, so rigor is not academic overhead, it is the precondition for accumulation. Sloppy findings do not add up to knowledge, they add up to confident nonsense.

We are not the first field to face this. Neuroscience spent the early 2000s producing results that often failed to replicate, until a public reckoning with statistical power forced it to raise its standards, and its knowledge finally began to hold and accumulate. The lesson worth borrowing is not bigger samples. It is the willingness to confront our own rigor in the open. So I will name where our rigor is soft, then focus on the two failures that do the most damage precisely because they are invisible. Both are ways we mistake an artifact for a finding. The first is measurement validity: we invent a “how fun, one to five” item and then track it across builds as if it were a thermometer. Drawing on the study of happiness, a construct that fractures the moment you examine it, I will show that the fix is not a perfect definition up front but triangulation that sharpens the construct over time, which is itself the scientific act. The second is researcher degrees of freedom: the readout that quietly leans toward what the team had already decided to ship. I will also meet the obvious objection head on. We ship games, we do not have time for academic rigor. But a research function that restarts every sprint is slower than a science function that carries priors, and the two failures I focus on are nearly free to fix: an analysis plan written before the session, a reused validated measure, a second coder on a single study. This is process, not delay. The call to action follows from all of it.

We do not compete on research. We compete on games and on winning players, and rigor is how we do that. The method is not the moat, the game is. So practice the rigor on your own work on Monday, build shared standards and instruments together as a field, and earn the word “science” the way every science earned it: by submitting our methods to one another and building knowledge that outlives a single study.

Presenter