REI Backpack Finder
Helping novice backpackers find the perfect bag, without the hassle
Project overview
Team: Brianna Pritchett, Lindsay Kelly, Xue Zhao, Yanfeng Jin, Xinyi Chen
Challenge: Allow novice backpackers to surmount the mountain of knowledge required to pick a safe, comfortable bag.
Context: Semester project for Research Methods in Human-Computer Interaction
My roles
-
Wrote and submitted an IRB application
-
Responsible for creating online survey during needs-gathering stage
-
Researched REI's learning resources for new campers
-
Conducted an open-ended interview with an expert backpacker
-
Conducted literature review on camping behaviors
-
Conducted user feedback sessions on various stages of prototype
-
Organized materials for cognitive walkthroughs
-
Wrote and gave final project presentation
Introduction to Problem
Backpacking is a type of camping where you carry everything for the trip in your bag. Often, backpackers' bags are 20-25 pounds and they hike for 4-5 hours a day, meaning that if the bag doesn't fit properly, campers can end up in an uncomfortable, even dangerous situation with little access to healthcare.
Additionally, there's so much to know when picking a bag:
-
Is it the right size for your torso and hip measurements?
-
Is it the right capacity for the length of your trip?
-
Will it carry everything you need - without being dangerously heavy?
-
What features are right for the nature of your trip?
We were drawn to this problem because of its high-stakes, complex nature.
Product Context Comparison
During our first round of research, we were interested in getting a better idea of the product space and the purchase process (especially since none of us have recently gone backpacking and many on the team hadn't heard of REI). We conducted the following methods.
Observation
Observed people shopping in the tents, shoes, and backpack sections of REI Atlanta
Online survey
Asked about priorities and resources while shopping for backpacks; 24 responses
Interviews
8 semi-structured interviews with REI customers and sales staff
In-store task analysis
Walked through the process of finding a backpack in an REI store
Class participation
Attended a lightweight backpacking class offered by REI
Web & mobile task analysis
Walked through the process of purchasing a bag through REI's website and mobile app
Icons made by Freepik from www.flaticon.com, licensed by CC 3.0 BY
We analyzed as this data a group using affinity diagramming for the qualitative data. For the quantitative data (demographics), we used summary statistics.
This led us to two personas: Expert Ed, an REI salesman, and Adventurous Amy, a backpacker who is about to embark on the Appalachian Trail. At this point we were still deciding who we would build a product for - salespeople or customers. We eventually focused on customers because of logistical difficulties reaching REI salespeople.
We additionally created a storyboard for each context in which a bag could be bought - in-store, online, or through the app.
In-store context
App context
Website context
After considering each context, we initially decided on the app context. However, we found in later research that people were hesitant to make larger purchases on their phone - as such, we switched to focus on the website context.
User Needs Analysis
In our second round of research, we were more interested in gathering user needs within this context and discovering design opportunities for solutions. We conducted the following research methods.
Market research
Looked at brand-specific sites (e.g. Osprey) and general camping sites (e.g. Campmor)
Product Q&A
Categorized + counted questions asked on certain product pages on REI's website
Expert interview
Talked to an expert backpacker about what makes a good backpack
Online questionnaire
Sent out a detailed online questionnaire about what people expect from each context
Contextual inquiries
Conducted contextual inquiries with people purchasing bags through app and website
Literature review
Conducted a literature review on behavior trends in camping
Icons made by Freepik from www.flaticon.com, licensed by CC 3.0 BY
We analyzed all of this data with a brand new affinity diagram and modified our personas to match our new findings. Approximately 10% of our users are Connected Carrie, 45% are Best-of-both-worlds Ben, and 45% are In-store Irene. Our aim is to make Carrie happy with her purchase and allow Ben to be better informed going in-store.
Describe your image.
Design Requirements
After these two rounds of research and analysis, we narrowed our key findings down into a concrete set of design requirements. We detailed desired functionality for the app as well as general considerations about our user group.
Functionality:
1.
Allow users to see the difference between brands (price, quality, warranty)
2.
Determine correct bag size, given user's measurements and trip details
3.
Suggest features for the user based on their information
User Considerations:
1.
Our users are novices who aren't familiar with jargon
2.
We want to invite users into the existing backpacking community
3.
We want to be cognizant of how nervous the users already are and simplify the process
Design Iteration
With this in mind, we split up and separately sketched out ideas, in order to encourage divergent and creative thinking. Parts of these sketches are shown below, one per team member - mine is on the far left.
Describe your image.
We then came together and discussed our ideas, converging into three main concepts.
Concept 1 is a survey with a "sticky" recommendations card - users fill out a survey with their details and they receive a card that follows them to each bag and indicates how well that bag fits their recommended criteria.
Concept 3 is an expert advice tool that simply puts you in contact with REI and community experts.
We conducted 7 user feedback sessions with novice backpackers to see what they would prefer to use and why. We find that:
-
71.4% of users prefer the survey to the interactive video
-
Users estimate that they'd spend around 7 minutes filling out a survey
-
71.4% of users say they see themselves using the expert advice tool, the rest say they might
Thus, we decided to go with a short survey that users would fill out with their details - this survey would then narrow their options to a few top recommended bags. At any point in the survey, they would also have the ability to talk to online experts.
In order to evaluate this system, we sat down with 7 users and asked them to buy a bag for a 5-day trip to the Smokey Mountains near Atlanta. Below is a page-by-page rundown of what we learned and how we fixed the issues we found.
Call to Action: Many missed this because it looked like an ad, so we made it less wordy and more prominent.
Measurements: This turned out to be a troublesome page, as many users don't know their measurements and don't want to take the time to find them, despite their importance in the fit of a bag. Lacking a good proxy, we made measurement-taking clear (and optional).
Secondary use: This was modeled after a function we found in market research, but was more confusing to users than it was useful. Thus, we removed it from the updated prototype.
Recommendations: Users liked having bags narrowed down, but wanted a clearer indication of why those bags were recommended, seen here.
Trip length: Users were unclear if "1 day" was overnight, so we updated the wording to reflect REI's expert advice terminology
Weight estimate: Users were generally confused by the utility of this page, so we made this more clear and modified the UI so that bag weight estimate is more distinct from what they're bringing.
Features: While users found this useful, users preferred having features recommended to them, so we folded this into the recommendations page.
Expert chat: Users indicate multiple hesitancies. First, they're not sure Ned is real, so we make it clear multiple experts are online. They're afraid of sounding dumb, so we suggest questions. Finally, we removed the option to email, because users wanted quick responses.
Evaluation
In order to evaluate the updated prototype (the "after" pictures above), we conducted 5 think-aloud usability tests with novice backpackers and 4 cognitive walkthroughs with usability experts. For users and experts, the task was the same as in the earlier prototype: find a backpack for a 5-day trip to the Smoky Mountains with your significant other. Experts were asked to imagine themselves as users. We find:
-
20% of users would use this instead of going in-store, which is larger than the amount that would buy entirely online without our tool (10%)
-
Average SUS score of 66, although many qualified that they wouldn't use this tool often, as this is a large purchase
-
More page-specific findings, detailed below.
Call to action: We removed the link to "Backpack Finder", as it was unnecessary to have multiple starting points.
Measurements: We modified the page to update recommendations as choices were made. Additionally, we labeled the videos more clearly.
Recommendations: Many users explicitly stated that they would try the bags if they were easy to return, so we propose a "Rent the Runway" business model, where people can have 3 bags sent to them, try them with their own gear, and return the other 2 for free. Users are additionally allowed to compare different bags.
Trip length: Besides making options mutually exclusive, we also moved the page description closer to the question (before, it was closer to Ned and no one noticed it was changing per-page). Additionally, we made the navigation more visually prominent and distinct.
Weight estimate: We updated the interface to more clearly indicate what the estimate is and whether this is too heavy, as many users were unsure how to interpret weight. We also added the ability to carry half a tent, as many forums indicate this is common.
Expert chat: The only change we made was to move the chat to the right, as an expert indicated this is more consistent with expectations.
Conclusion
-
I'm grateful to have been part of such a well-rounded, effective team
-
Moving forward, we'd like to send our final product to REI for potential implementation!