A speech prep tool to help you become a confident speaker.


The Project

Podium is a web application that helps students and professionals prepare for presentations and learn oral communication skills.

Project Details

  • Type: Graduate School Capstone Project
  • Duration: 6 months
  • Teammates: Shelley Xia & Wei-Hung Hsieh, Olen Ronning (Sponsor from 10K’)
  • Deliverables: interactive prototype, design specifications, poster, video prototype and final presentation in front of UW students, faculty and industry guests

The Challenge

Only 28% of U.S. employers think that new recruits are well-prepared in oral communication.

— Association of American Colleges & Universities

Oral communication skills can be difficult to learn, especially for students who are not enrolled in communications classes but are faced with situations where they must present an idea, share their research or participate in a class discussion. Even fewer options exist outside of the formal education system.

My Role

I led interaction design and research for the podium application. Some of my key contributions include:

Investigative Research

I facilitated semi-structured interviews, participated in a field study and conducted background research with my team in order to establish our project direction.

Insights & Ideation

I helped collect and synthesize a large volume of qualitative data in order to uncover insights about our target audience and inspire concepts.

Strategy & Process

I created frameworks and principles to help the team communicate and make design decisions. I also led content strategy and writing.

Interaction Design

I produced wireframes, interaction flows, sitemaps and design specs in order to translate our ideas into actionable product requirements.

Evaluation & Validation

I wrote evaluation guides and participated in usability studies. I worked through design refinements and communicated ideas quickly using sketches and flows.

Introducing Podium

How it Works

Users prepare speeches with recommended outlines, prompts and curated examples. Next, they record a practice speech and receive feedback in the form of voice quality data (e.g., pacing) to help them spot issues and make adjustments. Finally, they can share the video with peers, friends or family to collect meaningful human feedback on aspects such as body language, argument coherence or emotional impact.

Build an Outline

Add and edit section cards. Each section card has corresponding help information that’s customized based on two user inputs: speech type and audience characteristics.


Arrange Your Thoughts

Section cards are like digital notecards. They can be arranged by dragging and dropping.


Practice & Refine

Users practice their speech by video recording. Podium will generate a transcript and provide data on pitch, pacing, clarity and filler word use. Users can scrub the data or click on the transcript to find moments of interest.


Get Feedback

Users share their video with others by using podium’s customizable feedback form. Incoming feedback is available both inline and below the video.

The Journey

Uncovering Insights

We discovered that effective preparation is key to learning speech skills and gaining confidence with oral communication. To get here, we conducted user and market research to uncover the insights that would drive the rest of the process. We interviewed 10 students, 2 professors and the president of a local Toastmasters chapter at UW.

The following insights drove ideation and the concepts resulting from it:

Speaking is Personal

Our participants want to feel confident and sound competent when speaking. Inadequate preparation can cause feelings of apprehension and self-doubt. We knew that no matter what, our design solution needed to address this overarching reality.

Quality Feedback is Limited

Students often practice their speeches in front of family. These audiences don’t provide the best feedback due to lack of experience or emotional ties to the speaker. Quality feedback in general is scarce.

Practice is Paramount

Effective practice involves repetition, reflection and feedback. Our participants said it’s tough to incorporate all three elements without a structured plan of action.

Self-Recording is Awkward

Video recording practice speeches is a great method for learning but many hesitate to do it. It can be awkward to watch oneself speak, especially without an understanding of how to evaluate the recording.

Novices Struggle with Structure

Participants expressed concern that they weren’t engaging their audiences well because their speeches lacked compelling structures and stories.

Why Speech Prep?

From interviews we learned students struggled with presentations for one of three main reasons: speech apprehension, limited understanding of the subject matter or inadequate preparation. Inadequate preparation was the most feasible problem to address.

Learning oral communication skills is complicated to do in practice. You can’t read a textbook and suddenly feel confident. After consulting with subject-matter experts, we identified two learning goals:

  1. Speech prep has a unique workflow. A speech is not an essay or report.
  2. Speaking is social. A speaker and audience must be present in the learning process.

About that Workflow

Our design targets the three steps that most closely mapped to the pain points we identified during research: challenges with planning and structure, ineffective practice and lack of meaningful feedback.

speaking workflow@2x

Scoping the Opportunity

Our survey of the competitive landscape revealed two major problems with existing solutions:

  1. Comprehensive, high-quality learning tools (e.g., in-person classes or professional coaches) are inaccessible to a broad population.
  2. More accessible solutions offer content that’s difficult to apply in practice (e.g., free textbooks) or too narrow in focus (e.g., timer and teleprompter apps.)

Our vision was to go beyond the status quo and design a tool that’s approachable, comprehensive and personalized to maximize the potential for learning and growth.


Imagining Concepts

Prior to ideation, I worked on reframing our problem area.

How might we enhance the efficacy of speech preparation through an interactive and personalized approach?

We conducted several ideation activities. I facilitated the Crazy Eights activity for a group workshop involving eight people who were relatively unfamiliar with our project area. Amusing concepts like “Drunk Speeches” reinforced the notion that speaking is social. With Worst Ideas, the AI Speech Coach concept was dismissed quickly.

Feedback & Moving Forward

Managing feedback was difficult because it required the courage to abandon ideas we were excited about or the opposite: to follow our instincts in face of opposing voices, being confident in our knowledge of the topic.

Using insights drawn from research, I created six principles that aligned the team, facilitated design discussions and informed the playful product identity.


Testing Directions

We wanted to get our ideas in front of users as quickly as possible in order to narrow our direction. All three designs assumed audio would be primary user input for creating and analyzing a speech. We wanted to avoid awkward video recordings and focus the user’s attention on speaking out loud.

Three Options

  • Option 1: Clean interface, focused on speech-building task with an exportable script.
  • Option 2: Granular audio-recording and clip-comparison features. Focus on analytics.
  • Option 3: A step-by-step approach with tips along the way. Focus on learning and slower absorption of information.

User Reactions

We were fortunate to have been able to test with some of the same people we interviewed earlier, including a subject-matter expert. We learned some important things:

  • Where’s Video? — We thought audio recording would be approachable for beginners and sufficient for experienced speakers’ needs. This turned out to not be the case. Experienced speakers stressed the necessity of video. We figured beginners would want video eventually as they improved and gained confidence. We went with video but made it easy to hide the player.
  • Clean > Complex — Participants were overwhelmed by Option 2 but commented favorably about the voice analysis and transcript. They also wanted explanations for the data being shown. We decided to keep voice analysis features and simplify their presentation to make the information more digestible and actionable.
  • I Already Know This — Experienced speakers couldn’t see themselves using this tool, but students that they tutor or coach. This got us to think about other ways our design might alienate experienced speakers who could benefit from some features and annoyed by others. We concluded that learning-focused features would be optional and novices would be encouraged to take advantage of them.

“I don’t know what 83% means in this context.”
— MBA student

“This looks great for beginners…it kind of reminds me of Mavis Beacon Teaches Typing.”
— Undergraduate tutor

The Outline

Too much hand-holding in Outline Mode would annoy experienced speakers. Not enough and newbies could get lost. This led to the conclusion that the Outline would have dual identities: it would be a robust learning tool for those who need it and a blank canvas for those who know what they’re doing.

The Planning Wizard

Users can choose to start their speech by enlisting the Planning Wizard: a simple starting point to help novices get comfortable with structure so that they can feel confident filling out their outline.

The Reviewer Experience

During critique, we were asked “why would someone want to leave feedback on a friend’s random practice speech?” While this is a valid question, we’d learned from our participants that motivation to leave feedback wasn’t the issue — it was the quality of the feedback left. Students were already practicing speeches in front of friends and family, so we wanted the feedback feature to enhance those sessions.

However, we also considered a future application where users could share speeches with other users (strangers) on podium, creating an online community for feedback.

“I’m not going to waste time dwelling on things I am good at. That makes me feel like I’m wasting time in my self-assessment.”

Reviewer Hero Flow
Hero Flow: Reviewer Mobile Web


Future Directions

We saw several ways for the podium concept to grow beyond our initial design:

  • Content: With any learning application, credible content is key. We would seek guidance from experts regarding the creation of new content. For example, quick challenges to help users practice even when they don’t have an upcoming speech.
  • Integration: Modern presentations are visual. Podium could integrate with presentation software such as Keynote, Powerpoint or Prezi.
  • Tools: The ability to import any speech URL and have it run through the voice quality analysis system. This is a fun tool that empowers users to learn about speakers they admire.

What I Learned

Experience outweighs Features

We had to make many tough calls about which features were critical to delivering the experience we envisioned. We cared about the quality of what we could deliver in the time we had and this led us to prioritize experience over features. I learned that it’s OK to ship products that may be limited in scope if they excel in quality.

Ideas aren’t Personal

A great idea represents a shared vision among people. Podium taught me the power of combining and reshaping ideas to become better versions of themselves. This can only happen when team members let go of their personal opinions for the sake of something bigger.

Prototype Even earlier

Prototyping revealed flaws with our design and gave us a chance to address them. In the future I’d prototype even earlier and with lower fidelity because this reduces the likelihood of designing with certain assumptions too soon. For us it could have been choosing web as our platform. While our research findings informed this decision, we could have challenged it more aggressively.