11.1 OVERVIEW

My team was asked to address the following challenge: 

How might we leverage AI to create a better experience for outdoor runners? 

Project Details 

Sponsor: Pacer, a forthcoming AI assistant for outdoor runners
Duration: 11 weeks
Teammates: Eleanor Nesbit, Oliver Engel
Project roles: UX researcher
Project tasks: primary and secondary research, research synthesis, sketching, documentation

11.2 DISCOVERY

I started with secondary research and a competitive assessment of existing running apps to get a better understanding of the problem space, then primary research began, centering around interviews and mock contextual inquiries. 

Research Questions

1. How do runners use existing technology to enhance their running experience?
2. What are the biggest pain points people have while running outdoors?
3. What aspects of the outdoor running experience could an AI assistant improve?

Research Methods

• Competitive Assessment: We used and rated the leading running and exercise apps on the market.
  
• Interviews: Five (5) 30 minute interviews with regular runners or joggers to better understand running habits.
  
• Mock Contextual Inquiry: Five (5) modified contextual inquiry sessions with each interview participant where they traced a recent running route using an online mapping tool while thinking aloud to better understand feelings and preferences while running.

Participant 1 during the mock contextual inquiry

Participant 3’s mock contextual inquiry (via video call)

Competitive Assessment Takeaways

• Crowded market for running apps
• Popular routes exist in Strava and MapMyRun
• Google maps-style directions during runs exist in an app called RunGo
• No single app combines all these features and updates directions on the fly

Interview Takeaways

• User interviews were organized around a set of four areas of inquiry: participants’ running habits, motivations for running, digital product usage, and route selection
• Interviewees didn’t seem used to talking about their running habits, possibly since it’s a solitary activity for the most part
• Once the interviews got going our participants seemed to warm up and remember more details about their habits and motivations

Mock Contextual Inquiry Takeaways

Following each interview, we also tried to imitate the experience of going on a typical run with our users. Using onthegomap.com, participants traced the route of their last run while narrating what they encountered and felt along the way. This proved to be a particularly rich method for us. It revealed a number of feelings and preferences (preferred running surfaces, best intersections to cross) that didn’t come up in the user interviews held immediately beforehand.

Participant 1’s recent running route in Seattle

Participant 3’s recent running route in  Chicago

I listened to transcripts of the user interviews creating post-its of all the notable details. Then using grounded theory, I formed groups of themes from the data.

Key Insights

• Runners shift between two modes: casual and training
  

• Route planning is complex and involves trade-offs 
• Runners want their assistive technology to be useful yet unobtrusive
  
• Runners desire a balance between planned routes and spontaneous discovery and change

Runners with an upcoming race are more concerned about their distance and times, and they run progressively longer routes to build up to their race’s distance.

Recommended features:

• Closely match an upcoming race’s elevation gains/losses
• Personalized time and distance goals based on past results

A training assistance sketch

Route recommendations should be accommodating to the many different considerations that runners take into account while choosing a run. These recommendations should be taken as pieces that fit together in a cohesive experience.

Recommended features:

• Prioritize areas popular with runners
• Prioritize running surfaces like dirt, grass, and synthetic track over asphalt and cement, unless there are weather considerations
• Avoid streets with long traffic light signals
• Generate routes based on a desired mileage
• Provide several route options for the runner to choose from

A smart route recommendations sketch

Audio feedback is a potentially valuable feature that should be integrated seamlessly into the running experience. It is the best medium to provide real-time feedback without requiring the runner to stop and look at their device.

Recommended features:

• Audio features should be thoughtfully integrated with music and podcast listening, which are common activities that accompany running
• The type of audio feedback provided should be customizable

An audio feedback sketch

Features should be designed for serendipity and flexibility and account for the unplanned–keeping a balance between a runner’s desire for consistent routes and their interest in trying new things to avoid boredom and make the act of running more exciting.

Recommended features:

• Allow the runner to deviate from the planned route and update the route accordingly, ensuring that the run distance is unchanged
  
• Add extra distance to the end of a run, but make sure the new route ends at the same spot

A planning for the unplanned sketch

Next Steps

• Meet with stakeholders and users to prioritize concepts and develop an initial roadmap.
• Begin prototyping the core set of features, using Wizard of Oz and/or Adobe XD’s voice prototyping tool.
• Perform real contextual analysis alongside runners out in the world.

Personal Takeaways

• Create your own research methods: Don’t feel limited by the standard list of design methods. Sometimes the best tool for your specific job doesn’t exist yet. 
• Everything hasn’t been invented already: During competitive assessment, we were unsure what new insights we would be able to contribute, but our research turned up plenty of pain points with existing products and a number of opportunities to design something better.