Optimizing Smart TV On-Screen Keyboards

Enhancing Usability for Seamless Typing Experience

OVERVIEW

Text entry using the traditional D-pad remote is time-consuming and cumbersome. This project aimed to create a more efficient and user-friendly text entry experience that would enhance the usability and accessibility of Smart TVs.

Working alongside Professor Debayan Dhar, PhD student Suvodeep Misra, and colleague Abhinav Das, this project represents a research initiative aimed at making a contribution to scholarly work.

Design Process

ROLE

UX Researcher

TOOLS

Research

We began by exploring existing studies on text input methods for Smart TVs. Research showed a gap between measurable performance data and users’ personal satisfaction, highlighting the importance of designing for user satisfaction as well as efficiency.

These studies also provided valuable methodologies for optimizing keyboard layouts and assessing user experiences, which helped shape our research approach.

Ideation

In this stage, we focused on identifying common pain points, such as slow navigation and user frustration with traditional smart TV keyboards. We brainstormed over 50 keyboard layout concepts, exploring variations in key placement, input speed, and overall user experience to create more efficient and user-friendly designs.

Few Ideations:

KLM-GOMS Analysis and Prototyping

Using the Keystroke-Level Model GOMS (KLM-GOMS), we quantitatively analyzed each keyboard layout to predict how long it would take expert users to complete specific tasks. This model helped us rigorously evaluate the efficiency of various designs, guiding our selection of the most promising prototypes for further testing.

We developed these prototypes with HTML, CSS, and JavaScript, resulting in functional keyboard interfaces that could be accessed on smart TVs via remote control. The final prototype selection was based on both KLM-GOMS performance and each layout's potential to address key user pain points.

We used the following interactive prototypes for testing:

Usability Testing

To evaluate each prototype, we employed the Think Aloud Protocol Analysis (TAPA), allowing participants to verbalize their thoughts and experiences while interacting with the keyboard layouts. Each user entered test phrases such as "Multikey," "YouTube," and "Dance with joy," chosen for their variability and familiarity to users. Five prototypes were tested per participant, with the order of prototypes randomized to reduce bias in results.

Additionally, a Questionnaire was administered to gather ethnographic data and gain deeper insights into user backgrounds and preferences.

Questionnaire for Collecting Ethnographic Data:

  • Date of Birth

  • Where are you from?

  • Occupation/Background

  • Have you used a smart TV before?

  • If yes, how frequently do you use it?

  • Which apps do you use on your smart TV?

  • Do you use the on-screen keyboards on these apps for search?

  • If yes, how comfortable and efficient are you using a smart TV keyboard?

  • Is there anything you particularly like or dislike about smart TV keyboards?

After each prototype test, participants completed the Task Load Index (TLX) Questionnaire on a digital form to quantitatively gauge cognitive load across several dimensions:

Task Load Index Questionnaire:

  • Mental Demand: How much mental and cognitive effort was required to complete the task?

  • Physical Demand: How physically demanding was the task?

  • Temporal Demand: How much time pressure did you feel while performing the task?

  • Performance: How successful do you believe you were in accomplishing the task's goals?

  • Effort: How hard did you have to work to complete the task?

  • Frustration Level: How frustrated or stressed did you feel while completing the task?

Data Analysis

We recorded the time taken by each participant to enter each phrase with each prototype. TLX data was collected after testing each layout to assess cognitive load consistently.

The feedback for each layout was meticulously documented, with a particular focus on user criticisms, preferences, and observations. Through this approach, we derived quantitative data on task efficiency and cognitive load alongside qualitative insights, resulting in a comprehensive evaluation of each layout's performance.

Results and Conclusions

The project's culmination led to valuable findings. Certain layouts were better received than others, with users expressing preferences based on familiarity, ease of navigation, and placement of essential keys like space and backspace.