AWS Game Experimentation

Reimagining A/B testing to optimize player engagement, retention, and monetization

Demo / Prototype
View final designs
Role
Solo UX Designer & strategist for the 0→1 product
Duration
6 months
Outcome
Adopted by 4 AAA games, with ~74k players participating
AWS Game Experimentation

Driving a cross-team launch of a unified experimentation hub

I reviewed and presented designs to the Director at AWS Game Tech among other key stakeholders.

Collaboration

Testing game configurations across player cohorts was challenging

Icon 1

Need to constantly tune player experiences

Game studios require ongoing optimization to improve engagement and monetization

Icon 2

Testing custom configurations is complex

Hard-coding game configurations for each player cohort is inconsistent and error prone

Icon 3

Analytics implementation is time-intensive

Capturing meaningful performance data and identifying optimal configurations is tedious

Experimentation differs for product managers and developers

Product Manager

Primary Persona

Product Manager Persona

Runs experiments on targeted player segments to assess how new features impact key game KPIs.

Game Developer

Secondary Persona

Game Developer Persona

Mainly focuses on coding and development but also runs experiments when a PM is unavailable.

Customer workshops revealed key insights

I facilitated 4 1:1 customer workshops to research and document user needs and behaviors - what makes them start an experiment and what are their goals, how do they monitor an experiment while it's running, and how do they analyze the results after an experiment is completed.

Customer Workshops
Key Findings

Mapping the customer journey to communicate value proposition

I built a customer journey map to clarify player needs and communicate the service's value during the Amazon CXBR review. It was grounded in a New World use case where a PM tested which start area produced the longest session length.

Customer Journey

Moving from research to wireframes with customer validation

I generally begin my process with understanding the business requirements, identifying the customer needs and then iterating on the designs while finding ways to validate them with customers throughout the process.

Design Process

Low-fidelity wireframes informed by qualitative research

The qualitative data gathered during research informed the workflow. I also developed concepts and low fidelity wireframes using Balsamiq, and reviewed them with the Director and Product Managers.

Workflow Connectors
Wireframes 1
Wireframes 2

Usability testing validated workflows and surfaced improvements

We validated the wireframes with 6 customers (3 Product Managers and 3 Game Developers) through a usability study that I organized and facilitated. I wrote the study plan for a task-based analysis to assess the effectiveness of the workflows.

Usability Insights

Finalized designs using the AWS Design System

I collaborated with the AWS design system team to understand existing patterns and identify gaps. I also worked with the PMs to prioritize features, starting with single experiments and planning multiple experiments based on user feedback.

Create experiment

Customers can set up an experiment and configure the parameters they want to test.

Create Experiment

Define variants

They specify different variants for player segments. I explored layouts that make comparisons intuitive, responsive, and clear across all screen sizes.

Define Variants
Responsive Variants

Choose participants

Customers select player segments for the experiment. Based on usability feedback, we added guidance on experiment duration, factoring in the number of variants and participants.

Choose Participants
Experiment Duration

View pending experiments

Experiment creation is separate from starting/stopping. For the MLP, customers can manually start/stop experiments, with future plans for pre-scheduled runs.

Pending Experiments

Monitor experiment

Customers track high-level metrics for running experiments. Once complete, the system identifies the winning variant. Key KPIs—acquisition, retention, and monetization—are highlighted to enable holistic assessment.

Running Experiment
Experiment Metrics

Contributing a counter visualization pattern to the AWS Design System

The existing design system did not have a counter visualization pattern to highlight an important number. I designed a couple of pattern variations and worked with the design system team to develop specifications and document scenarios for usage. The pattern was borrowed by designers working for other services like AWS EC2 and Pinpoint.

Design System Pattern

Establishing design processes and facilitating critiques across AWS UX

To strengthen collaboration among designers who owned different products at AWS, I set up a weekly critique process with a sign-up sheet, defined roles, clear guidelines, and note-taking templates. This enabled 15 critique sessions in four months and boosted participation and cross-team collaboration.

Design Critiques

Successful launch adopted by multiple games

The teams at Amazon Game Studios adopted the service and used it for conducting AB tests on games like New World, Crucible and The Grand Tour Game.

4
teams from Amazon Game Studios adopted the service
93
experiments created in the first 6 months
78%
of experiments initiated were completed successfully
74,211
players participated in experiments
Outcome Metrics

Iterating based on feature requests from successful adoption

Based on the success of the experimentation feature, internal teams requested the following features that I iterated on: a feature to conduct promotional campaigns and a feature to ramp-up new player experiences.

Feature Iterations

Reflections

This project taught me AWS core fundamentals and how to work through technical complexity. I dove into the tech, learned how the APIs and SDKs worked, and ensured parity between the UI and CLI. With a new, ambiguous service and as the sole designer, I learned to break big problems into manageable pieces, design simple yet scalable solutions, and communicate effectively under tight deadlines.

Other Case Studies

Let's Connect

I'm always open to discussing new opportunities, collaborations, or just chatting about design. Feel free to reach out!

divs.hariharan@gmail.com