Mover was an internal research project developed at Fraunhofer Portugal with the goal of encouraging older adults to be more active. It was an Android application that tracked users' activity as well as detecting falls.
I led the design of the future vision for Mover, considering how fall risk assessment, social features, and other new functionality might fit in.
The project did not have the resource allocation to allow direct formative user research like interviews or questionnaires, so I chose to start by conducting a literature review and a comparative analysis to better understand the needs and opportunities for Mover.
I defined the following research questions for my initial study:
What makes Mover different than other activity-sensing applications?
Why would a user choose Mover over applications that offer more functionality?
What is its niche?
How should fall detection be incorporated into the design? Fall risk assessment?
What is the best way to add a social and/or competitive facet to Mover?
I used a matrix to compare similar and related products based on the features offered and the target audience.
Other apps with fall detection or similar features
I first searched for applications with a fall detection feature or other focus on elderly users' health.
pMonitor: Fall detection, location tracking, panic alerts to emergency contacts
iFall: Fall detection with emergency contact
Age Care: Fall detection, location tracking, emergency contact, activity sensing
How's My Mom?: Tools for caretakers to report to loved ones
Rolling Ball: Tests fall risk – walk 15 meters keeping ball in center of target
Other activity tracking apps
Since there are many applications that use GPS to track fitness activities like running or biking, I chose two popular free applications from this category. I looked at Endomondo and Runkeeper, two popular activity sensing applications with various additional features including sharing achievements, playing music, and counting calories.
I focused my literature review on articles about other applications that sense activity or address health concerns of the elderly. The particularly salient articles I reviewed were the following:
Activity sensing in the wild: a field trial of UbiFit Garden (Consolvo et al., 2008)
UbiFit Garden was an activity-sensing application that aimed to encourage mainstream users to be more physically active. It used a graphical display of a garden on the user's mobile phone wallpaper to show activities completed and goals achieved. The authors of this work have published several articles on UbiFit Garden and related projects (including the next item on this list).
Design requirements for technologies that encourage physical activity (Consolvo, Everitt, Smith, & Landay, 2006)
The focus of this study is a mobile application prototype called Houston that logged step count data from a pedometer. It also allowed for sharing with "fitness buddies" of the user. As the title suggests, the article discusses findings that can be applied to similar applications. The app was aimed at mainstream users.
Digital family portraits: supporting peace of mind for extended family members (Mynatt, Rowan, Jacobs, & Craighill, 2001)
This article discusses a project in which digital picture frames visualized an elderly family member's daily activity in several categories (the activities were not actually sensed in the study, but were "Wizard of Oz" prototyped based on interviews with the user). The picture frame served as an ambient display meant to replicate "natural activities undertaken to keep an 'eye out for Mom' if she lived next door," with the aim of giving family members peace of mind.
Acceptance and usability of a relational agent interface by urban older adults (Bickmore, Caruso, & Clough-Gorr, 2005)
This article describes a project called Laura, in which a computerized system with an animated face and synthetic voice spoke to elderly users about their health and activity levels daily, offering responses for the user to reply.
Designing acceptable assisted living services for elderly users (Vastenburg, Visser, Vermaas, & Keyson, 2008)
This article asserts that assisted living design needs to change its focus from the technology to the users. In particular, the authors argue that assisted living applications should demonstrate a short-term benefit that outweighs the short-term cost, leverage the existing social network, and fit within known metaphors and patterns. Three prototypes are described, including one that aimed to motivate exercise.
Social networks can be encouraging, but can also cause tension
Sharing with a social network can encourage activity. When Houston was tested with and without social capabilities, only those with them maintained their activity levels. Moreover, elderly users love receiving personal messages (Vastenburg, Visser, Vermaas, & Keyson).
Social features can also make some users uncomfortable: users don't want to share when they have not done well, and they often find privacy settings to be too difficult.
➜ Make it social, but tailored to our users
➜ Conduct interviews to understand users' opinion on issues like privacy and competition
➜ Prototype different approaches to evaluate their reception
"Glance-able," persistent displays have the greatest impact on users' behavior
Persistent displays help people to conceive of their activity as long-term trends, to have a bigger influence on behavior. Seeing an overview often reminds users of their goals.
➜ Summarize activity in one place, like the home screen.
➜ Consider creating a live wallpaper for Mover, to display the glance-able overview.
➜ Test visualizations to find what is easy to understand.
Positive reinforcement really works. Offer more small rewards to encourage activity
One participant commented, "It was fun to go back and go, yes, there's my star for that day!" (Consolvo et al., 2006, p. 463). Users also remarked that they were surprised how much they cared about the flowers that represented their achievements (UbiFit Garden).
➜ Mover should do more to acknowledge and reward user's efforts: stars, coins, points, or something else besides only showing the current status.
➜ Enable encouragement via the social network in the form of likes or comments.
Use conversational prompts with pre-defined responses to keep users engaged
Elderly users responded positively to a human-like "relational agent" (Laura). They also found it easy to select responses from set choices.
➜ Update the tone of dialog to be more friendly and human-like.
➜ Prompt user for input when inactive for long periods or e.g. to test for fall risk.
Allow for manual input of activities
Users were frustrated when they did not receive credit for activities that had not been tracked (Houston, pedometer log). When given the option, users logged 26 different types of activity (UbiFit Garden). Moreover, a healthy, active life includes more than just moving (Digital Family Portraits).
➜ Consider allowing manual input, including different types of activity, through a checklist or prompt.
Good design is especially important for elderly users
Fisk et al. argued that older adults are not as capable of compensating for bad design as younger and more experienced users, while Vastenburg et al.'s results showed that a poorly matched metaphor made the application difficult to use.
➜ Evaluate designs against established heuristics: compatibility with mental models, consistency, error recovery, feedback, etc.
➜ Conduct usability testing.
The obvious shortcoming was that I did not actually speak to users. To offset this, I incorporated user interviews into usability testing of prototypes in the next phase.
Some findings from the literature applied to mainstream users rather than older adults.
Ad hoc personas and scenarios
After presenting my results to the team, we discussed the future for Mover. Based on our ideas I developed a set of four ad hoc personas and scenarios to guide the design. To emphasize the fact that they were ad hoc personas -- lacking a solid basis in user research -- I used sketches to represent the personas.
I began by prototyping on paper, working with the help of Francisco Nunes and Paula Alexandra Silva during this phase. Once we reached a somewhat stable idea of the new design, I created an interactive prototype.
To get feedback on the prototype before entering development, I planned a usability test that I administered to six older adults with Francisco's assistance.
The overall objectives of the usability testing study were to determine how well Mover's features met the users' goals, and to evaluate how easily target groups could perform the primary tasks. I defined the following research questions:
Do users understand the purpose and functionality of Mover?
How easily are users able to register an account? Do they find the process to be acceptable?
Do the methods of representing activity make sense to the users? Are they able to accurately interpret them?
Do users understand the concept of "nudging" a friend?
Do users notice and understand the share buttons?
Does the organization and flow of the application match users' expectations? Where do they get stuck?
Do users think the application is useful for them?
I used Adobe Fireworks to create a set of linked HTML files that could be run in a browser on a smartphone to imitate the behavior of a native application. This method enabled participants to move throughout the application more easily than a paper prototype would allow.
Participants were recruited via friends and families of colleagues. We tested with six participants similar to our targeted user groups, as defined by our primary personas:
3 adults aged 20-45, concerned about the health of loved ones, moderate to advanced technical skills
3 adults aged 45-65, desiring to maintain an active lifestyle, moderate to advanced technical skills
Sessions were conducted in a group meeting space at Fraunhofer Portugal with one moderator and one observer/debriefer. Participants were asked to think aloud while they used the prototype, which was run on a Samsung Galaxy S. Each session lasted around 35 minutes and included a brief introduction, three tasks, and a debriefing interview.
1. Sign up
A friend of yours has been trying to get in better shape and recently recommended Mover to you. You just downloaded the application and want to set up an account so you can follow your friend's progress and get more exercise yourself. Sign up for Mover using your Facebook account. Since this is an early prototype, you will not actually be able to input any data into the forms, but tapping empty fields will fill them.
2. Encourage a friend
Since you signed up with your Facebook account, Mover already has a list of your friends who are using the application. Have a look at your friends. Find someone who is not meeting their goals and give them some encouragement.
3. Review and share your activity
Now imagine you have been using the application for a few weeks. You are having so much fun using Mover that you want to tell more people about it. Post a message to Facebook informing your friends about one of your recent successes.
During the test sessions, Francisco and I both took notes about the users' actions and comments. After testing was completed, we met to discuss our notes and general impressions. We created a spreadsheet to record and compare which users experienced which issues during the test. From our discussion and analysis I arrived at the key findings and recommendations presented here.
Visualizations of status and achievements were not clear to users
No participants demonstrated understanding that stars represented attaining Mover status. Only 2/6 associated stars with the nudge icon, and 2/6 were unable to determine which friends had moved the least.
➜ Simplify the rewards system and its representation so that it is not duplicating or overlapping.
➜ Design an initial walk-through to acquaint new users with key features.
Users overlooked or could not locate the social features
3/6 users struggled to find a way to make a post to Facebook. One user commented that the Facebook button appeared to be a part of the person figure. Another user was unfamiliar with social media and did not associate the icon with Facebook.
Users often went through numerous screens searching for a way to make a post. Only one user saw the nudge button and chose this action without receiving a tip. One user stated that she had no idea what the button meant and only clicked it because of its location.
➜ Social buttons need to be more easily identifiable. A more recognizable icon should be used. Change the term "nudge" to something more widely understood. Make sure the Facebook button looks like a button, perhaps with an action verb accompanying it.
➜ Both the nudge and Facebook buttons should appear anywhere the user expects to find them.
After the first test, the registration process was changed so that one field was displayed per screen. Subsequently, users navigated the process more smoothly than the first participant, who was overwhelmed by reading and answering several form elements on one screen.
Several of the Portuguese-speaking participants were confused by the wording of the registration fields, specifically for the display name. We planned to change the text from "What would you like to go by?" to something less colloquial, like, "What's your name?"
At least 2 users were confused by a question in which "No" was already selected as the default answer. We planned to change this to a check box or a toggle button.
The user may not receive sufficient feedback after "nudging" a friend. One participant disliked the action because she did not see any results of selecting the button. We planned to add a brief description of what happens after nudging a friend (i.e. "Mom will be notified next time she opens Mover").
One drawback of this study is the low number of participants, but testing with six users revealed the major problems. We also didn't test with users from the 65+ group, but that persona did not interact with the app much, rather carried it on their person.
Another factor that affected the test is language. The test introduction, instructions, and debriefing interview, as well as the text in the prototype, were written in English. Only 1/6 participants spoke English as their first language. Some participants received some explanations and made some “think aloud” comments in Portuguese. However, this can provide an unintended benefit, helping us make the app more accessible to users with any level of English.
Following usability testing, I reported my findings to the team and revised the prototype accordingly, documenting the interaction map of the entire system for the graphic designers and developers to reference.
However, the project was put on hold to focus on Dance! Don't Fall, and soon afterwards I left Fraunhofer.