Design Workshop:

Speaking and Pronunciation Training

How might we help English learners in Japan improve their verbal fluency?

2021, Japan

Background

User research revealed that English learners in Japan tend to struggle the most with speaking skills, compared to grammar, vocabulary, listening, and reading. In response, this project aimed to integrate a third-party voice recognition API to help students improve their pronunciation and speaking abilities.

The project was part of an effort to strengthen the market competitiveness of Reallyenglish, an English e-learning service in Japan that offers a variety of online courses for learners at different proficiency levels. Initiated in June 2021, the design phase lasted for two months, during which I served as the sole designer.

I began by researching both the specific challenges Japanese learners face in English-speaking and the technical capabilities of the voice recognition API. I then facilitated a design workshop to generate innovative ideas for supporting speaking skill development. Following that, I created multiple mockups and worked with the team to narrow them down into a viable short-term solution, along with proposals for potential long-term improvements.

Due to internal engineering resource constraints, development began in late 2022, with the product scheduled for release in March 2023.

This case study focuses on the research and the design workshop I conducted in August 2021.

User Research

The goal of this research was to understand our target users, English learners in Japan, including their motivations, learning methods, and unmet needs in the area of spoken English and pronunciation.

1. One-on-One Interviews with Internal Sales Team

In 2021, I conducted 1:1 casual interviews with members of our internal sales team. These conversations helped me:

Build trust with stakeholders
Gather firsthand insights about customer feedback and competitor offerings
Collect ideas and feature requests that could address real customer needs

Key user needs identified:

Improve TOEIC Listening & Reading scores
Target: university students and corporate employees
Practical English for daily life, workplace communication, and cultural understanding
(実生活で役立たせるための教育、外国の文化を教養、一人で会話の学習)
Target: university students and working professionals
Specialized English for academic or professional domains
(専門領域強化 eラーニング)
Target: university students in specific fields

2. Review of Past Client Case Studies

Our sales team had documented why various universities and companies chose our English learning services. By analyzing these case studies, I uncovered:

The specific challenges faced by different client types
Features and outcomes that clients valued most
Areas where users expected improvement or additional support

Key feedback about things that customers think are important

3. Secondary Research via Google

I conducted desk research, reviewing blog articles, app reviews, and forum posts in Japanese. This allowed me to:

Hear directly from English learners about their personal experiences
Understand frustrations with existing tools and teaching methods
Identify emerging self-study trends and preferred learning formats

The blogger (link) thought that it's important to practice speaking.

Key Insights

A recurring challenge among Japanese learners is the difficulty in speaking with native English speakers. The root of this issue often lies in over-reliance on Katakana-based pronunciation, which differs significantly from native English phonetics.

We identified three user segments:

Learners unaware of the Katakana gap: They don’t realize their pronunciation deviates from native English norms.
Learners aware, but lacking confidence: They know their pronunciation needs work but are unsure how to improve.
Motivated learners looking for practice: They actively seek feedback and tools to refine their speaking skills.

Design Challenges

Based on the insights, the main design challenges became clear:

Raise awareness: Help users recognize the difference between Katakana pronunciation and native English pronunciation.
Provide assessment and guidance: Offer clear, personalized feedback and actionable improvement suggestions that build confidence and support continuous learning.

This research laid the foundation for a collaborative design workshop, where we ideated solutions to support verbal fluency and pronunciation training using voice recognition technology.

Design Workshop: A Simplified Design Sprint

How might we help English learners in Japan improve their verbal fluency?

To align the team and generate a wide range of fresh product ideas, I facilitated a half-day simplified design sprint workshop, from 11:00 to 16:00 with a one-hour lunch break. Although I was the only designer in the company, the session included engineers and project managers to bring diverse perspectives.

The slides I prepared for the design workshop introduced the design sprint steps.

Pre-Workshop Preparation

Before the day of the workshop, I prepared the following:

Presentation slides to guide participants through each step
Workshop background and goals, based on user research insights
A shared “lightning demo” document where each member uploaded screenshots of relevant apps as design inspiration
Calendar invites to book everyone's time
Workshop materials: sticky notes, A4 paper, sticky dots, pens, and a timer (I used my phone)

Participants:
1 designer (me), 2 engineers, and 2 project managers — 5 in total

Workshop Agenda

1. Background & HMW Generation (20 min)

I introduced the workshop goals, summarized research findings, and asked our Japanese-native member to share her personal experience learning English.

Key takeaways from the discussion:

To achieve verbal fluency, learners need:
→ Confidence
→ The ability to speak full sentences naturally
→ A relaxed mindset about making mistakes
→ Better pronunciation
Many learners are unaware of the gap between native English pronunciation and Katakana English.

While the discussion happened, each member wrote HMW (How Might We...) questions on sticky notes to capture opportunities or challenges. We then collected all the HMW notes on a whiteboard.

2. HMW Voting & User Flow Mapping (30 min)

Each member voted for their favorite HMW notes using 3 sticky dots.

Then we co-created a user journey map on the whiteboard:

We started by defining the end goal: improved verbal fluency.
Then we mapped out rough steps a user might go through to reach that goal.
It wasn’t about being precise—just about surfacing ideas.

Next, we placed the HMW notes along the flow. This helped visualize where the most opportunity areas were and gave everyone a shared understanding of the journey.

3. Lightning Demos (20 min)

Each person introduced 1–2 existing apps or tools from around the world. These apps featured unique approaches to language learning.

Thanks to the diversity in our team, we uncovered features from English-learning markets outside Japan, many of which had not been widely adopted locally.

4. 4x4x4 Brainstorming & Card Sorting (60 min)

This was a silent brainstorming session, designed to avoid typical pitfalls, like dominant voices or off-topic conversations.

4x4x4 means:

4 ideas × 4 sticky notes × 4 rounds
With 5 members, this ideally generates 4 × 5 × 5 = 100 ideas

How it worked:

Each member wrote 4 ideas on sticky notes and placed them on an A4 sheet (1 idea per note)
After 5 minutes, the sheet was passed to the next person
The next person read the previous ideas, then added 4 more new ones
We repeated this for 5 rounds

By the end, every sheet had ~20 ideas, each built on ideas from different perspectives.

Post-brainstorming:

After lunch and coffee, we grouped similar sticky notes through card sorting
We gave each group a category name
Members voted for their favorite 5 ideas using sticky dots

5. Paper Sketching (60 min)

Each person picked one idea and turned it into a paper wireframe using A4 paper and pens.

Sketching was done individually and quietly. Everyone worked at their own pace and seat.

6. Idea Presentation (30 min)

Each participant pinned their paper wireframe to the wall and gave a short explanation of their concept.

Ground rules:

Q&A allowed
No critiques or judging, even if the idea wasn’t feasible

7. Heat Map Voting (20 min)

Each member received 6 sticky dots and was asked to vote not on the whole idea, but on elements or features they found most interesting.

For example, they might dot:

A useful feature
An intuitive layout
A clever copy or interaction

This created a heat map of popular or promising parts across all concepts.

Paper wireframes and voting

8. Closing & Takeaways

We ended with a quick thank-you and a group reflection session. Everyone shared their thoughts on the workshop experience.

Results of the day:

~100 raw ideas
5 visualized concept wireframes
Clear visual indication of popular features via heat maps
A shared understanding of the design opportunity space

Why This Worked

Gave non-designers a voice in the creative process
Avoided brainstorming pitfalls by enforcing silence and structure
Balanced divergent thinking (idea generation) and convergent thinking (voting, sketching)
Encouraged collaboration while respecting individual creativity

After the Design Workshop

After the workshop, I digitized all handwritten notes and wireframes to create a structured record of the outcomes. From there, I began curating and evaluating ideas for both short-term feasibility and long-term potential.

Idea Evaluation

To prioritize the outputs, I assessed the ideas using three key criteria:

Useful: Does it address a real user problem?
Interesting: Does it offer a fresh or engaging experience?
Feasible: Can our team realistically build it in the near term?

For the feasibility assessment, I collaborated with engineers to ensure that we were aligned with our current technical constraints.

Workshop Output

71 ideas were generated during the brainstorming session, though many were not feasible at this stage
5 wireframes were created by participants during the workshop
4 additional wireframes were created by me post-workshop to explore more promising directions
The short-term solution we chose was the most fundamental and feasible, based on the current capabilities of both our content and engineering teams

71 ideas were generated during the brainstorming session.

5 wireframes were created by participants during the workshop.

The short-term solution we chose was the most fundamental and feasible.

Reflection and Takeaways

While the final deliverable was a relatively simple and achievable solution, the process of broad exploration proved invaluable.

The combination of user research and a structured co-creation workshop helped expand our thinking, uncover overlooked needs, and generate ideas that may not be implemented immediately, but provide a roadmap for future development.

As a product designer, I believe it’s essential to go beyond expectations, not just those of our users and customers, but also of internal stakeholders.

Why spend so much time on ideas that might not be built?

Because if we only design based on narrow, specific requirements, we risk losing sight of the bigger picture.

We must look deeply into our users’ pain points and aspirations, think beyond what stakeholders explicitly ask for, and explore bold, even unconventional ideas.
Only then can we return to reality, narrow our scope, and deliver solutions that are not just feasible, but impactful and visionary.

This mindset not only improves the immediate product but also lays the groundwork for more meaningful, future-forward innovations.

Other ideas for verbal speaking practice in the future.