Convoy
HCI Researcher & Product Designer
Convoy is a proposed animal rescue technology service with the goal of reducing the rates of euthanasia and saving animals' lives. Designed to support the coordination and facilitation of transporting animals from high-risk, high-kill environments to no-kill facilities where they can be provided adequate time to be successfully adopted.
Problem
Estimates vary, but at least 2 million animals are euthanized each year in the US because shelters simply run out of space before animals are successfully adopted.
Sadly, there is often room to board animals at alternative shelters, but the planning and coordination to understand capacities, and organize and facilitate transportation is fragmented and inefficient.
Software can help solve that.
Solution
Provide a technology service layer to dramatically increase the number of animals transported and ultimately adopted, reducing overall euthanasia rates.
Alert no-kill shelters within a specified radius that animals are in danger of being euthanized.
No-kill shelters with capacity agree to accept them.
Transport and communication is centralized without the need for phone calls, spreadsheets, and emails.
Animals in grave danger of being euthanized are provided a lifeline by these volunteer organizations who ultimately share the same goal.
Stakeholders
There are three stakeholder groups for this system currently existing within the animal rescue space:
Shelters designated as ‘high-kill’ because they are routinely over capacity and have to address the issue with euthanasia.
Shelters that rarely reach capacity and are classified as ‘no-kill’.
Volunteers and organizations who transport animals between shelters.
Research Approach
Semi-structured qualitative interviews were conducted with participants from animal rescue organizations, shelters, and transporters. Some interviews were recorded and transcribed verbatim. All interviews followed a standardized format yet deviated when participants explored something that expanded the understanding of the problem. Numerous informal conversations and email exchanges provided supplementary data.
Ethnography in the form of field observations were conducted at various facilities. Combining data collection methods such as interviews and ethnography helped ensure that conclusions drawn weren't biased by a specific methodology. Field notes were made throughout observation to capture things that may have become so routine to participants they failed to mention them or were not comfortable expressing. My positionality as a researcher was strengthened by my experience adopting and rescuing animals, creating a trustworthiness allowing me better access.
Users
Primary: The active users of the system are the animal rescue organizations saving animals from high-kill environments, and the animal shelters who have to manage capacity through euthanasia.
Secondary: The users who will receive outputs from the system are the animal transporters. When an agreement is finalized to transfer animals between organizations, they effectively act as an intermediary between the two primary user groups.
Tertiary: The users having no direct involvement in the system but who are impacted by it are animal owners surrendering their animals to shelters and animal adopters who adopt animals from rescue organizations and shelters.
Task Analysis
Currently, organizations rescue animals from high-kill environments by a time-consuming and cumbersome process of calling shelters to find out if there are any animals in danger, reviewing websites of high-kill shelters for animals ready to be adopted, visiting facilities lacking online profiles, and calling known volunteers to transport animals from high-kill environments.
Critical tasks:
Shelters alerting that there are animals at risk of being euthanized
Rescuers agreeing to rescue animals at risk of being euthanized
Rescuers requesting transportation assistance to transfer the animals
Transporters accepting the request to transfer the animals
Rich Picture
The introduction of an organizational technology can change work practices and power dynamics. It is, therefore, crucial to understand who does the work, who benefits from it, and what barriers can impact adoption. Utilizing both contextual inquiry and soft-systems modeling as requirements gathering methods, provides the view of the various organizations as the system, with people and technology as components of that system. A rich picture can help to better understand the tasks these organizations perform, as animals start their journey either from the street or being surrendering by their owner, through the various organizations dedicated to rescuing, transporting, and rehoming them, in an effort to ultimately save their lives and provide them with a loving, stable home.
Analysis
Government-run animal shelters are inundated with animals, both surrendered and stray. They lack the resources and time to update animal profiles or transfer at-risk animals. No one wants to euthanize animals, they literally just run out of space and time. Providing the opportunity for local volunteers to transport animals can expand the support. All groups create trusted partner networks of organizations they have worked with before which results in small networks to rely on. All organizations noted a need for more efficient communication between groups, broader networks, and more accessible technology.
Recommendations
Mobile first: Ensures that organizations can access the technology no matter where they are, or what they are doing. It allows for push notifications which are immediate.
Automated Group Chats: A group chat can be automatically created with all relevant parties so that communication is contained and transparent, not fragmented as it is currently.
Automated alerts: It would be preferable to integrate with animal shelter databases, along with their intake dates, to automate the creation of animal profiles, estimated euthanasia dates.
Verifiable Organizations: Organizations could verify each other and network effects would ensure a larger set of trusted partners to work with.
Volunteer Drivers: Volunteers could select a local radius they are willing to transport between and accept transportation requests in their area.
Wireframes
Evaluation Methods
Objectives
User satisfaction
Effectiveness
With the goal of answering the following questions:
Will users find the features easy to use and useful?
Will users be able to accomplish tasks, generating few interaction errors?
Data Gathering Approach
Descriptive research methods were applied for evaluation as experimental research is better at uncovering laws and principles rather than behavior in the real world, and would have required a controlled laboratory environment. Participants were given a high-level overview of the goals of the system and purpose of the evaluation. They were informed that the design was a prototype and not a complete system. To support the evaluation objectives participants were asked to perform the five tasks outlined below on a mobile device using a prototype.
Observational notes were taken throughout the procedure and where possible audio was recorded and transcribed and a mix of data was captured and analyzed. Taking a mixed-methods approach to evaluation by combining both qualitative discovery with quantitative analysis can strengthen the validity and reliability of results.
Think-aloud Protocol
Error Counts
System Usability Scale (SUS)
Interviews
Participants
A purposeful sampling method was initiated to recruit participants for evaluation based on their role within the animal rescue space and the time they were willing to dedicate to the research. Two participants who had already been engaged in the initial discovery phase recommended other potential participants. Information on the age, gender, location, and rescue type of each participant is seen below.
P1 | Female, 45-54, urban - works full-time as a facilitator for a rescue organization.
P2 | Male, 55-64, suburban - volunteers for a rescue organization.
P3 | Female, 25-34, urban - works full-time as a vet technician for a rescue organization.
P4 | Female, 25-34, suburban - volunteers for an animal transport organization.
P5 | Male, 25-34, suburban - works full-time as a technologist for a rescue organization.
P6 | Male, 25-34, suburban - volunteers for an animal transport organization .
Tasks
Participants were asked to undertake four structured tasks (tasks 1-4), seen below, and then a fifth task that allowed them to freely navigate around the application. During tasks 1-4 error counts were recorded. Task 5 allowed for more qualitative data to be captured and to inform the SUS questionnaire.
Create Account
Create Alert
Create Request
View Convoy
Free Interaction
Success Metrics
SUS = 68+
<2 Errors/task
Testing Environment
Testing was undertaken in context at the respective organizations where participants P1, P3, P4, & P5 worked. They were seated at their desk undertaking tasks on their mobile devices, which were a mix of Android and iOS. Participants were able to perform the testing uninterrupted, however the noise of animals and colleagues could be heard in the background. It wasn’t safe to perform testing while driving, so all tests, including those for transporters were performed indoors. For participants P2 & P6 who volunteer infrequently, testing was undertaken remotely. The prototype was accessible on their mobile device and their reactions and feedback was gathered via a video chat. Remote testing enabled participants who were not local to contribute to the evaluation and research.
Results
Interview Findings
First Impressions
All participants were impressed with how clear and easy the system seemed to use.
Favorite part
The amount of time it will save was noted by all participants as the most useful aspect of the system.
Confusing or frustrating
Some participants were unsure if they could just choose the animals they wanted.
Three participants mentioned that although it was easy to open the menu navigation to navigate to the various sections of the app, it would be less frustrating if they were more readily available.
One participant wasn’t sure if you had to use a transporter for every Convoy.
Improvements
Two out of the six participants suggested that rather than having to fill in their organization’s details when creating an account it could be captured from their Facebook business page.
Three participants noted that they would like more control over their radius.
Other participants thought the pickup and dropoff range would be problematic knowing the nature of animal rescue workers.
Two of the six participants, who were older, mentioned that the size of the font could be larger to help them with readability.
Deal breakers
All participants identified that if the system was much more time consuming than the prototype then they may not use it.
Some were concerned about the verification system devolving into everyone getting verified and it being rendered useless.
All participants noted that rescue groups typically know what type of animals they wanted to rescue at any given time because of capacity and wanting to have a diversity for people to choose from, therefore it was imperative that they be able to select only the animals they wanted and not the entire list.
Usability Issues
By analyzing verbal think-aloud data, errors counted and interview data the usability issues described below were identified.
Facebook Connect | New Feature Request
Persistent Navigation | Minor Problem
Tapping on areas not supported by the prototype | Cosmetic Problem
More control over rescue radius | Minor Problem
More control over the pickup and delivery times | Minor Problem
Multiple verifiers for each member | New Feature Request
Ability to select only animals wanted not entire list | Cosmetic Problem
Font size larger or customizable | Minor Problem
More control over alerts beyond off/on | Minor Problem
SUS Scores
Error counts
The other measure analyzed was performance accuracy and the effectiveness of the UI in terms of error count per task for each participant. There were a total of 13 errors recorded across 24 attempts at the 4 tasks with 6 participants.
Discussion
Based on the quantitative measures the UI can be considered successful in terms of user satisfaction and effectiveness. All six participants scored higher than the industry standard of 68 on the SUS questionnaire with the range being 75 to 95. Additionally, no task generated more than two minor errors for each participant and each task was successfully completed by all participants. The tasks requiring users to create something generated four times the number of errors than the task requiring users to view part of the system, however, these errors were still within the acceptable range.
Participants P1 and P5 found the system the most useful and easy to use scoring it at 95 each. This is not surprising given that P5 is a technologist who understands software and its benefits and P1 feels the pain points that Convoy seeks to solve very tangibly every day. Participant P4 who volunteers at an animal transportation organization scored the lowest at 75. P4 is less involved with actual animal rescues and more on the administrative side, so the score is not unsurprising in that context. Participant P4 also generated the most number of errors, five across four tasks, which likely contributed to the lower satisfaction score as well.
Whilst the quantitative data revealed successful metrics, the qualitative data was able to uncover some usability concerns and feature requests. The number of fields required to create an account is necessary to provide user profiles with enough data to automate other aspects of the system and to ensure members are credible and trustworthy, but it could be cumbersome for the user. The idea to programmatically populate those fields from a Facebook business page is a good one and easy enough to implement.
It was unexpected for the menu navigation to be raised as inconvenient, from a heuristics perspective it was a valid approach, but for certain types of users who only use one or two sections of the app regularly, it makes sense that they would prefer to have the navigation persistent across the app to easily access those areas at all times. Some of the feedback pertained to more control over the settings which was really an artifact of the prototype scope. A fully functional system could easily provide deeper control over radii, alerts, and font display sizes.