The research leading to Haven and its attendant solutions fit into three families of methodology:
Navigating a Vast Domain (Exploratory Research)
When beginning our research, we turned to four exploratory research methods to better understand how people behave in disasters and why:
Secondary Research
A comprehensive reading of literature, both academic and popular, on a host of topics related to disasters.
Guerrilla Research
Quick and impromptu interviews with people about their experiences with disasters, or what they think they would do in a disaster.
Expert Interviews
Detailed interviews, mostly over gchat or phone, with experts in the domain of disaster preparation and management such as members of Paragon, Rubicon, and The Red Cross, as well as a number of employees of hospitals where disasters have recently taken place, both natural and man made, (hurricanes Irene and Maria, the Boston Marathon bombing, etc.)
Surveys
Amazon MTurk surveys of how people handled and/or would handle disasters.
Key Findings
People often overestimate how prepared they are for disasters.
During a disaster people prefer to turn to their friends or family for safety rather than a shelter.
People with special medical needs and their caretakers are the most adversely impacted by disasters and the most likely to go to hospitals.
Citizens with special medical needs are often unaware of what services are available during disasters.
Designing with Ideas (Generative Research)
From exploratory research we had to generate actual solutions that could help people and institutions in disasters. For this, we turned to a number of generative research methods:
Stakeholder Map
A diagram of the relationships between the key players from the government and media, to hospitals and volunteers. Everything revolves around the citizen in this example. Teletracking provides efficient patient flow to hospitals who provide treatment and sometimes shelter to those who need it during disasters.
How Might We
A method of brainstorming ideas. Here we asked: "How might we encourage people NOT TO GO" to shelters and hospitals during storms if they don't need to?
Crazy 8s
Another brainstorming technique: Try to think of eight design answers to a problem in eight minutes. Each yellow card is a prompt surrounded by some of our responses.
Where We Stand
Also known as "The Polak Game" for its inventor, Frederik Lodewijk Polak. The goal is to discover how people imagine the future and their impact upon it. We used this game to gauge the opinions of several Teletracking employees during our first visit to their offices.
Where We Stand
People plot themselves, or in this case a post-it-note with their name, on a grid representing their agreement or disagreement with these statements:
“While things will go wrong in disaster preparedness from time to time, overall they will get better in the next five years.”
“While there are large processes that will shape disaster preparedness in the next five years, the largest factor will be how individuals prepare and respond.”
The arrows are how people's opinions changed after discussion.
Story-Boarding
For our second meeting at Teletracking we brought five scenarios which we had storyboarded, then we "Speed Dated" (tested) them with employees for feedback.
"Redirection," pictured here, proved very popular and became a key component of Haven.
Story-Boarding
"Volunteers," was also well received, but did not fit within the focus later narrowed upon: citizens with special medical needs in the wider community.
Story-Boarding
"Capacity Creation" for hospitals in disasters also eventually fell out of scope.
Story-Boarding
"Personal Preparedness" suggested a way for Haven to reach its potential user base.
Story-Boarding
"Tracking" patients through sensors is a service Teletracking already provides hospitals and it could be expanded to include other visitors in times of disaster. This figures into the "smart" networks system on phase three of the roadmap.
Reverse Assumptions
From the feedback gathered during the second in person visit to Teletracking, our ideas were unwound through an excercise where assumptions made for our proposed solutions are flipped. Then services are designed that facilitate these new paradigms.
Co-Design
Employees of a hospital in Naples, Floraida engage in a co-design activity: card sorting.
Partipants organized a wide array of concepts such as "Capacity Management," "Camera," "Water Level," and "Injury" into groups by perceived relatedness. As they did this, they were encouraged to think their thought process out aloud.
Key Findings
The solution should be focused outward (towards citizens in the community at large) rather than inward (hospitals), and should have a product, such as an app, to go along with any service it provides.
Pitching the solution toward the caregiver of a person with special medical needs would be optimal, as they are usually more comfortable with newer technology.
Will It Help? (Evaluative Research)
User testing was conducted both on-site and remotely in Florida, as well as in Pittsburgh. People with special medical needs, as well as caregivers for people with special medical needs comprised the vast majority of our test subjects. Between tests, iterations upon the design of Haven were made based on user feedback.
User Testing: Florida
User testing was conducted in Florida, both in person and remotely through video chat. Several ciizens with special medical needs as well as their caretakers were interviewed at senior housing and community centers in the greater Tampa area. User testing was also conducted with officials involved in county level disaster planning throughout Florida.
User Testing: Pittsburgh
To continue iterating Haven's design, it was also necessary to conduct user testing in Pittsburgh. In these tests, a scenario involving a snow storm replaced the usual hurricane scenario. Aside from improving Haven's overall usability for the elderly and people with medical needs, the winter storm scenario demonstrated the viability of expanding Haven's coverage toward a wider array of disasters.
Key Findings
Potential users were often wary about having an app extract their data itself.
Users were more willing to trust an app offered in partnership with a well known and respected organization, such as the American Red Cross.
Instructions and functions should to be clear and precise for our target users.
"Scores," "countdowns," and other game-adjacent experiences would aid our users in preparation.