Individual thesis project | 8 months, 50% time | Grad school coursework | 2014


I delved into a 'mini wicked problem' for my Masters thesis project: what causes Grand Canyon National Park to become a perfect storm of visitor calamity, and what interventions be able to help decrease the high rate of rescues? 

I used design research to explore what makes Grand Canyon so dangerous for visitors, and to audit current visitor safety efforts. Then I used a variety of synthesis methods, as well as looking to behavioral science and risk management, to understand further what I had found. Then I explored how we might improve the situation by utilizing tools and best practices from information design, interaction design, service design and behavioral design to encourage better decision making, and nudge visitors towards safer choices.

Download the final thesis document from CMU's Thesis Repository, or read on for the (somewhat) condensed process story.



I spent 8 days at the Grand Canyon conducting field research, shadowing and interviewing Park Rangers, observing and talking with visitors and auditing the current safety and information touchpoints. I grouped my exploratory research around six different discovery focus areas:

In addition to my field research, I also conducted extensive secondary research around existing work on the subject matter and best practices from both design and other applicable fields. A breakdown of how research endeavors map the discovery topics can be viewed below:  

One interesting finding from my field research was around who hikes at the Grand Canyon. I decided to refer primarily to "visitors" in my thesis rather than "hikers" because most people hiking at the Grand Canyon aren't "hikers" at all - they're just normal people who hike when they come to a famous National Park because it's "the thing" to do there. Many are traveling in groups, and add a stop at the Canyon to a larger roadtrip or a vacation in Las Vegas. 


My research trip was extremely helpful, but it also created a massive amount of data and insights for one person to synthesize. Since I had sought out not only to understand the problem’s “as-is” state, but to attempt to get to the bottom of it, there was alot to process. 


I used frameworks, diagramming, ideation, and affinity clustering to try to boil down the key things I had learned:


As I was working to synthesize and stay true to my research findings, possible intervention ideas were constantly popping into my head. Sometimes I simply post-it noted them and added them to an ideas parking lot. But I found that playing with some of these ideas could be a helpful way to pick the forest out through the trees:

I clustered my ideas to see which problems I was innately draw towards solving, and why I thought that lever would work. This allowed me to work both forwards and backwards to reach a better understanding of the problems at hand, and how they interacted with one another. 

I also looped back to many of the theories and methods I had discovered in my exploratory research, drawing from risk communications, behavior science, and design. I was beginning to feel like I had a handle on what was going on, but it was still rather complex and interrelated, with a constellation of opportunities for improvement rather than a single unifying need. 


I found there were some key themes that bubbled up to the top level of the "why" tree of what led visitors to problems. I found that visitors lacked the information they needed, they didn't have the right stuff for their hike, they were going farther than it was safe for them to go, and they weren't really understanding what they were getting themselves into. 


Another key synthesis finding was that visitors usually do "get" that the Canyon is hot, difficult to hike, and that they're out of shape, didn't know what to expect, and didn't bring enough supplies -- AFTER they're well into their hike when it is too late to prepare or choose better:

PREFERRED ORDER of understandings (pink) to inform actions (purple). For instance, visitors should choose which trail to take based on understanding the challenges of the Canyon, and how difficult the hike being considered is, and what their group's fitness level is. Ideally each of the purple decisions are thoughtfully and actively made.

PREFERRED ORDER of understandings (pink) to inform actions (purple). For instance, visitors should choose which trail to take based on understanding the challenges of the Canyon, and how difficult the hike being considered is, and what their group's fitness level is. Ideally each of the purple decisions are thoughtfully and actively made.

OBSERVED ORDER of understandings (pink) coming after actions (purple). For instance, visitors often aren't even actively choosing a specific hike thoughtfully, and then they aren't realizing their group's ability level, or how hard their hike is until after they have, for reason, turned around and headed back uphill. This is obviously a bad time to realize you've choosen to hard of a hike because your only options at that point are to complete the hike back out of the canyon, or call for rescue.

OBSERVED ORDER of understandings (pink) coming after actions (purple). For instance, visitors often aren't even actively choosing a specific hike thoughtfully, and then they aren't realizing their group's ability level, or how hard their hike is until after they have, for reason, turned around and headed back uphill. This is obviously a bad time to realize you've choosen to hard of a hike because your only options at that point are to complete the hike back out of the canyon, or call for rescue.

This meant that a key challenge was how to try to help visitors gain this understanding sooner - or be prepared to guide them more at the moments when they actually realize they need help.


At the halfway point of the project we have a poster presentation to read out what we've done so far, and where we are headed. It was a challenge to boil down everything I had learned so far into a 15 minute talk and a single poster, but the forcing function helped crystalize my key learnings.  

View the key diagrams below (click to enlarge) or view the full poster PDF in a new window.


In addition to sharing my findings about the As-Is state, my poster also included a few initial thoughts on possible intervention directions, flip through below:


I also made a fun postcard highlighting the key items that I'd found hikers should NOT do, because they are also kinda good advice for life in general:



As I moved from the understanding the problem and its context to considering possible interventions, one possible direction I wanted to explore was a hike planner tool. This could help close the knowledge gap, and aide visitors in matching their skills and interests to an appropriate activity. It was also the most interactive and technological of my intervention options, so it could be a good fit to spend more time iterating on.

I first asked participants to interact with existing information sources, and asked which hike they would choose based on this material. Then I speed-dated a few low-fi concepts that showed the kind of functionality an on-site activity choosing tool would have, and used a card sort to prompt a discussion and probe at people's expectations and needs were around hike planning. 

I was right that low-fi concepts were all I needed to learn if this approach could be fruitful - the feedback I received from testing these concepts was invaluable, and pushed me back away from a digital solution. Testers were offended by the idea of tech at the trailhead, without even being asked about it, and this was coming from CMU students!

I also found that planners will have already planned before reaching the park, and that most of my problem hikers won’t want to engage with a planning tool in this way. While I was interested in using this project to explore the digital form a bit more, I was also excited to pivot back a larger world of possible intervention approaches and artifacts. My findings from this probe also aligned with my "order of operations" hypothesis that hikers didn't yet know they need help when they're about to start their hike. 


After sharing my new understanding of the as-is, analyzing the current intervention approaches, ideating some possible solutions, and testing one particular approach I had further solidified my approach to designing some possible interventions. I still didn't know what the right answer was, but I had a good idea of what the goals of the interventions should be. 


I wanted to address some of the key issues than I had identified with current interventions when crafting my own. I had found that: 

  • Visitors are seeking hike planning and choosing information, not safety information

  • Visitors have difficulty comparing hikes to each other and choosing one

  • Information touchpoints often aren’t convenient for or relevant to visitors

By providing critical information in a visitor-accessible manner and at the right time and place, I hoped that visitors will be more likely to make the preferred, safer choice at the key decision points I identified: Choosing a hike, choosing which supplies to bring, and choosing to turnaround. So:

How might we enable informed choices?

  • Afford and guide comparison

  • Create distinct decision points

  • Co-locate information with decision points

  • Mimic visitors’ decision approach

As long as visitors are overconfident, they are likely to remain "deluded", dismissing warnings as intended for other people and making overzealous plans. I propose that by helping visitors be a little less optimistic, they can more accurately assess themselves and their hike, leading to smarter choices that decrease the visitor-hike mismatch. Countering optimism bias should also encourage them to realize more quickly when they are outmatched.

How might we counter optimism bias?

  • Redirect visitors’ attention to overlooked information

  • Use contrast to draw attention

  • Increase availability of overlooked criteria

  • Make overlooked criteria easy to grasp

  • Increase availability of poor outcomes and base rates

  • Create feedback loops

  • Highlight options and decision opportunities

  • Use information to empower less optimistic group members

I found visitors often do not have a strong idea of which activities they want to engage in in advance. Instead, they look to find out what is “the” thing to do at the Park. Redirecting these defaults away from dangerous hikes could yield major safety gains.

How might we boost alternatives?

  • Provide visitors with more medium options

  • Allow visitors to get a sense of accomplishment without

  • endangering themselves

  • Make safe activities be and/or seem more compelling

  • Emphasize experience over checklist accomplishments


1. To reach visitors, interventions should:

  • Be placed on-site at the Park

  • Be accessible and usable by a high volume of visitors

  • Be hard to avoid

  • Be readable and professional

2. To work within the park system, interventions should:

  • Work with the Park’s laissez faire policy

  • Be economical

  • Stand up to environmental challenges

  • Provide helpful guidance without requiring a person

3. To appeal to visitors, interventions should:

  • Provide the information visitors seek

  • Afford at-a-glance use

  • Afford interaction without relying on digital

  • Be group-friendly

4. To enlighten rather than confuse, interventions should:

  • Use hierarchy to make clear what’s most important

  • Minimize words where possible, while maintaining clarity

  • Provide varied levels of information

  • Be consistent across channels, touchpoints

  • Aid decision-making or provide how-to knowledge

  • Assume no prior knowledge

  • Use everyday language AND make measurements meaningful

  • Meet visitors where they areUse specifics to illustrate a larger point

  • Make it easier to do the right thing

  • Utilize the power of affect


Having learned much more about the nature of the problem through testing just one possible solution, I wanted to explore more of my possible interventions and get feedback on each to really test my insights and principles. Unfortunately my time was limited so I didn't have a chance to iterate much, or get formal user feedback on my intervention ideas. 

Instead, these concepts serve as "conversation starters" that help illustrate what I learned and some ways we might begin to tackle this issue. 


ON-TRAIL INFORMATIONAL SIGNAGE creates a designated decision point, precipitating group discussion. By framing the choice unambiguously as one of turn back now or continue on with a longer return hike, it challenges the tendency of visitors to mindlessly continue down the trail and encourages reluctant group members to speak up.


DESIGNATED HIKING LEVELS makes it easier for visitors to gauge the overall difficulty of hike sections, and pairing this system with the LEVEL FINDER helps visitors choose the appropriate hike.



AN ELEVATION-BASED MAP is designed to highlight elevation and effort and combat optimism bias by encouraging visitors to think about the way back uphill right from the start. THE TRAIL GUIDE pairs the elevation guide with important safety and hazard information to empower visitors and remind them that things can and do go wrong.

AN OVERVIEW MAP, styled similar to a ski resort trail map, highlights varying Hike Levels and allows visitors to visually compare trail length and steepness. A hawk’s eye view makes elevation change much more apparent and provides a sense of scale to help visitors begin to grasp the immensity of the canyon.


Helping visitors "get" the challenges of the canyon better before they begin a hike would help them make smarter choicer and be better prepared, but it's also a big challenge, because behavioral science tells us it's extremely hard to help people understand something experiential before they actually experienced it.

These concepts attempt to put visitors into an experiential mindset, to help increase the availability of poor outcomes, and to frame the hike as a challenge, rather than just a fun day out:

The UPHILL CHALLENGE creates a fun, interactive way to close the hot-cold empathy and perception gaps by letting visitors experience the challenges of uphill hiking at the Grand Canyon before they actually head out to hike. 


#FAIL PHOTO OP redirects the existing visitor behavior of mocking safety signage to a more productive dibiasing end, increasing availabilty of rescue outcomes, and prompting visitors to figuratively (and literally) put themselves in the place of someone needing rescue because of their own “stupidity.”



ON-TRAIL DE-BIASING SIGNAGE  relies on proven de-biasing techniques to encourage visitors to be more realistic and informed when deciding on a hike. Rather than focusing on issuing a warning which visitors will assume is “not for them,” this signage draws attention to overlooked considerations, and emphasizes base rates, intending to help visitors make a critical reality-check:

REBRANDING TRAILS makes safer trail options more appealing to visitors, and draws attention away from less safe options. By rewording trail descriptions and choosing images associated with hikes and destinations carefully, the Park can help steer visitors towards better summer day hike choices. 


One example of this is rebranding the full length of the Rim Trail as a destination hike. The Rim Trail is 16 miles long, giving it great potential as a landmark hike for visitors, but current trail descriptions and branding make the Rim Trail unattractive to fit visitors. By creating a fun name - The Full Enchilada - and a brand, this hike can entice visitors away from long inner canyon treks that are much more dangerous.

A different framework for organizing the concepts

A different framework for organizing the concepts


In the last stage of the process, I created my thesis book that goes through the entire process, from secondary findings to ideation and final strategies and concepts. Please download the official document PDF from CMU's Thesis Repository


I had hoped to return to the Grand Canyon to work with the team there and share my findings, but unfortunately a family medical emergency interfered with those plans. However, it remains at the top of my side project to do list, and I would love to even file to be an official study to see how people respond to some of these approaches in context...



I get asked this alot. Well, rather than starting from a particular area of design for my thesis, I wanted to start from a problem, and see how I could help tackle it through design. I was looking for a mini-wicked problem, if you will.

I decided to focus on something that had bugged me for sometime. It’s something we hear about every summer: visitors to National Parks doing seemingly incredibly stupid things that get themselves killed or seriously injured. While experts quoted in the media dismiss the cause of these incidents as sheer stupidity, I suspected there was something more interesting going on, and I wanted to get to the bottom of it.

My initial research confirmed that this was an issue of real concern for the NPS, and they were continuing to try new ways to address the problem. I thought by taking a fresh look at the issue through a design lens, I might be able to diagnose and treat the problem in new ways and help.

Here's the full introduction from my Thesis Proposal:

Each year people make about a billion recreational visit to US National and State Parks: over 280 million recreational visits to the US National Parks. For the overwhelming majority of these visitors a trip to the park ends without injury or death. Yet, despite numerous preventative measures, people do continue to require rescue, and even die, while visiting parks.

What is remarkable about these episodes is not so much their numbers; the vast majority of park visits occur without incident. It is that when incidents do occur, the surrounding circumstances can make the victims’ actions seem not only ill-advised, but downright foolish.

One recent high-profile instance of this occurred in 2011 in Yosemite National Park. A group of visitors climbed over a guardrail alongside the Merced River just 25 feet above the edge of 317-foot Vernal Falls. They stood in the cold, rushing water to play around and take photos. Bystanders urged them to return to safety. Suddenly, one person lost their footing, and fell into the raging river. A second person tried to rescue the first victim, and fell in too. Then a third person went into the water as they tried to help the second. All three perished after washing over the falls.

A Chicago Tribune feature article after the Vernal Falls deaths declared in its headline: “How do you explain park deaths? You can’t. Expert in Yosemite, Grand Canyon fatalities at a loss to explain ‘stupid’ behavior.”

In the piece, Michael Ghiglieri, an experienced outdoor guide and co-author of books detailing the history of deaths in Grand Canyon and Yosemite National Parks, was asked about the recent Vernal Falls incident and similar events where people seem to have thrown commonsense to the wind and found themselves in deep trouble. His expert opinion? “It’s so stupid it’s beyond belief.”

Mr. Ghiglieri’s opinion echoes the collective reaction the public and press often has to these high-profile events: “How could people be so foolish? Clearly, they have only themselves to blame.”

And park staff tends to agree that “human error” is a major contributing factor to injuries and deaths in the parks. In fact, it was cited by 100% of park staff surveyed from 30 National Parks as being a medium or high factor contributing to visitor accidents, making it the highest rated factor in the survey. Human Error was followed by behavioral factors (such as playing, running), age, level of visitor preparedness, and level of visitor experience in activity.

But is this type of behavior really beyond belief? Should we still be surprised, with all we know about human beings that people sometimes do things that don’t seem to make a lot of sense at first glance?

As stated in the recommendations of the NPS’ Report: “Park staff responding to the inventory questionnaire often identified visitor characteristics as significant risk conditions. Staff rarely rated communication or infrastructural hazards as important conditions contributing to visitor accidents. Some factors perceived as problems related to visitor judgments and behaviors, however, could also be understood as failure in communicating relevant information successful to visitors.”

In other words, is attempting to reduce human error not a key responsibility of safety communications, rather than something that falls outside of it?

The Report goes on to state that there is a real gap between this staff perception of bumbling visitors, and the perception visitors have of themselves: “…park staff members believe that visitor preparedness and level of experience in a given activity are important contributors to visitor accidents. Most visitors, however, considered themselves experienced in their chosen activity and many indicated that they were prepared with appropriate shoes, clothing, water, etc.”

This difference could be a result of self-reporting bias, but it could be something much more troubling. Do visitors think that they are heeding warnings, being prepared for activities, when in fact they are mistaken? And are park staff members dismissing this mismatch as human error, rather than recognizing a critical communications failure. Whether it is due to miscommunications, or “human-error,” if parks know that visitors will err, should they not design with that in mind?


No one actually asks this, but that process is a pain, so I'd like to give myself credit for this. Our thesis projects were required to go through the IRB approval process, and while I did pass as a IRB-exempt study, that still added additional constraints especially on recruiting for and conducting user research.

Becoming an official study at a National Park also involves a large waiting period and alot of red-tape, so I worked at the park with permission from the SAR Ranger team, but without official approval as a research study. This meant my interaction with visitors needed to be limited to observation and informal contact that didn't draw too much attention. I was expressly prohibited from probing visitors more formally due to a lack of official study approval from NPS.