ALEXA SKILL: PET PANTRY 

 

This project is from my time in Interactive Media III, a course within the Digital Design program at the University of Colorado, Denver. We were tasked with modifying an existing Alexa skill template in order to create a new skill for the Amazon Echo. This could be on any topic of the student's choosing. The only requirements to this assignment was to build the skill to completion by the due date, in a state to which it could be demoed in front of the class flawlessly, and conducting user tests within the process of its creation. 

Being a dog owner and animal lover, I chose to create a skill that would be useful to myself as well as pet owners everywhere. Finding information on food that is safe, or unsafe, for dogs and cats is time consuming and quite unreliable at times. Having access to this information via the voice user interface of the Amazon Echo would save time and worry, especially if this information is backed by veterinarians (which is a requirement of mine if this skill were to be published to the Alexa app for public use). 

Project Timeframe: 6 weeks

MY ROLE 

 

I was the only individual working on this project and was responsible for the experience strategy and design, research, and development of this Alexa skill. Although this was my first opportunity working with voice user interface design, my background in connected play, from my time working at Sphero, proved vital in the project's success. Creating a connected UX environment requires fundamental human-centered design approaches, whether that be in connected play or VUI design. The user has not one, but multiple devices to interact with and each one demands time and attention in its production.

 

In the case of the Amazon Echo, the user's focus is regularly on the VUI. However, the Alexa app component needs just as much design attention, as I realized during my research. Both the Echo and companion app must seamlessly guide the user towards their goal at every interaction point. 

SKILLS USED

  • UX & VUI Design 

  • Human-Centered Design

  • Graphic Design & Brand Identity

  • UX Research

  • Competitor Research

  • Alexa Skill Development

PROJECT DELIVERABLES

  • UX Flows 

  • Sitemaps

  • Branding

  • Demo Videos 

  • Skill Description 

  • UXR Documentation

TOOLS USED

  • LucidChart

  • Adobe Illustrator

  • Sublime Text

  • AWS Lambda

  • Amazon Development Portal 

  • Amazon Echo & Simulator

PHASE ONE

RESEARCH & DEVELOPMENT

NEEDFINDING 

 

I began the process of developing my Alexa skill concept by brainstorming various ideas I considered to be useful via voice user interface and the accessibility of the Amazon Echo. These thoughts were shared with classmates, family, and friends. The outcome: a unanimous vote for a painless way to know of food items that are safe to feed their pets. Admittedly, I was not aware that cat owners would necessarily fancy this type of product, nevertheless, I received requests for cats to be included within this project. This gave the skill a unique spin, one that would surprise as well as delight the unfamiliar niche that is human-food eating felines and their owners. 

Once set on an idea, I found a similar skill in the Alexa app called Dr. Dog. Unfortunately, Dr. Dog received many one star reviews due its limit of food item information available, among other baggage. Testing this skill for myself, I found it included only basic, widely known foods with no additional information except whether the item was safe for a dog to eat or not. This validated the need for a new skill while giving concrete actions to further improve the user experience when building out my concept. 

BLANK CARD SORTING ACTIVITY

 

After conducting a blank card sorting activity with a few individuals in my target demographic, I found there were user goals that I was incapable of implementing due to technical challenges: limitation within the templates provided (in addition to my most basic knowledge of computer programming) as well as the need for VUI advancements that are not currently available but are being heavily explored, specifically the evolution of linguistic taxonomy recognition in AI programs. 

Left: This individual requested additional information that was specific to their prior inquiry, though there is limitation to Alexa's ability to enter or remain in reception mode after she provides information to the user. Additionally, the data input required for such a fluid and detailed conversation is beyond my capabilities. 

I was unsuccessful in finding a template that would perform as I intended on its own. Thus, I created two versions of an Alexa skill that could meet the primary needs of my users and proceeded with A/B testing in order to gather further insight into user goals.

ITERATION ONE

HEALTHY PET

       

ITERATION TWO

PET HEALTH

       

PHASE TWO

A/B TESTING

TEST PLAN 

  

OBJECTIVES

  • Establish desired skill functionality

  • Indicate pain points within each process

  • Determine appropriate invocation name

KEY MEASUREMENTS

  • Qualitative

    • Determine dissatisfaction within each skill variation

    • Requirement to invoke the skill every-time (Healthy Pet)

    • Requirement to indicate specific pet during repeated use (Pet Health)

    • Evaluate interpretation of skill functionality

    • Determine appropriate invocation name

  • Quantitative

    • Determine success delta between skill variations

    • Gather sample utterances and food item inquiries

 

PROSPECTIVE TEST SUBJECTS 

  • Pet owners

    • Experienced Echo users

    • Non-experienced Echo users

  

DEPENDENCIES

  • Hardware

    • Echo Dot

    • iPhone/iPad

  • Software

    • Skills in AWS

    • Alexa App

    • Echo Simulator

    • Quicktime

 

LOCATION

  • Remote

  • Available locations in Denver/Boulder

 

LEGAL

  • Informed consent form

  • No NDA required

TEST DESIGN 

  

30 MINUTE SESSIONS

  • Switch skill "A" and "B" per test

INTRO

  • Gender

  • Age

  • ASK: What type of pets do you own?

  • ASK: Do you own an Echo or Echo dot?

PET HEALTH/HEALTY PET

  • Show first skill in the Alexa app

  • OBSERVATION: Are they able to invoke the skill?

    • Are there any pain points?

    • If so, where?

  • OBSERVATION: Do they reference the app??

    • If so, why?

    • How often?

  • OBSERVATION: What is their frustration level?

  • OBSERVATION: What were some moments of delight?

  • OBSERVATION: What do they ask about?

  • Additional feedback/observations

A/B USER TESTING

6 PARTICIPANTS

EXECUTIVE SUMMARY

  • Participants were unaware that the skill variations were solely about healthy food items for their pet

  • Participants were unaware of the necessity to re-invoke the ‘Healthy Pet’ skill in order to request additional information

  • Participants would rather hear more information than intended, rather than having to specify their pet type when requesting additional information in ‘Pet Health’

PARTICIPANTS

  • 6 total participants

    • 83.3% | 5 participants were female

    • 16.7% | 1 participant was male

    • 33.3% | 2 participants were 18-24 years old

    • 33.3% | 2 participants were 24-30 years old

    • 33.3% | 2 participants were 30+ years old

    • 100% | 6 participants own dogs

    • 16.7% | 1 participant owns both cats and dogs

    • 50% | 3 participants own an Echo or Echo dot

  

PET HEALTH/HEALTY PET

  • Show second skill in the Alexa app

  • Repeat questions and observations from above

 

DEBRIEF

  • ASK: Where might you use this type of skill?

  • ASK: Did you prefer the first or second skill?

    • Record which skill was shown first

      • Pet Health

      • Healthy Pet

        • ASK: Why?

  • ASK: How did you feel about Alexa asking about your pet type again while using the Pet Health skill a second time?

  • ASK: How can we improve this experience?

  • ASK: What else should we include?

  • ASK: Should we take anything out?

  • ASK: What would you call this skill?

  • Additional feedback/observations

ANALYSIS

  • Pet Health

    • 100% | 6 participants were able to invoke the skill

    • 66.7% | 4 participants had trouble with the skill

      • ‘Ask something else’

      • Said ‘Yes’ or ‘No’ instead of ‘Dog’ or ‘Cat’

    • 50% | 3 participants referenced the app during this time

  • Healthy Pet

    • 100% | 6 participants were able to invoke the skill

    • 83.3% | 5 participants had trouble with the skill

      • Invoking the skill for further questions

      • Knowing the information in the skill only held food items

    • 50% | 3 participants referenced the app during this time

ACTION ITEMS

  • App

    • Modify skill icon

    • Revise skill and invocation names

    • Revise skill descriptions and sample phrases

    • Include more legal copy in the skill description

    • Reconstruct information provided in the Alexa app after using the skill

 

  • Debrief

    • 50% | 3 participants preferred Pet Health over Healthy Pet

    • 33.3% | 2 participants preferred Healthy Pet over Pet Health

    • 16.7% | 1 participant did not have a preference

    • 50% | 3 participants explicitly said they preferred the name Healthy Pet

      • “Easier to say,” “Stuck in my head better,” “Easier to remember”

    • Other suggested names

      • “Food for Dogs,” “Dog Food Q&A,” “Common Foods for Animals” 

      • “Healthy Diet,” or something that includes food

      • “Can my pet eat it,” “Eat or Not,” "Pet Can Eat," “Can Pet Eat,” “Pet Food Check” 

      • “What can my dog eat,” “Can my pet eat this,” “Pet Food Okay,” “Pet Chef” 

      • “Food for Healthy Pets,” “People Food for Healthy Pets”

 

  • UX

    • Modify skill flow as follows:

      • Open skill > Request food item > Receive information for both cats and dogs > Choose to ask something else or end the skill

    • Include lists of common can and cannot eat food items 

    • Include ‘Cancel’ and ‘Quit’ utterances to the stop intent

    • Include more information on why pets cannot have certain foods

    • Fix bugs

ITERATION THREE

PET PANTRY