Designer & Researcher
Beating the Bots
UX Graduate Capstone Project
January 25, 2021 to April 21, 2021
User Research + Service Design
Project Overview
My final semester at Purdue University was focused on a capstone service design project.
This study centered around an
in-depth exploration of particular communities referred to as sneaker cook groups. These groups comprise of collectors and resellers who utilize their exclusive resources to gain a competitive edge in acquiring limited clothing, footwear, art pieces, and various collectibles.
Context
Let's set the stage with a scenario. Picture yourself filled with anticipation for an upcoming sneaker release. You're so eager that you're ready to wake up early and join a virtual queue on Footlocker.com. You have a strong intuition that these sneakers will be in high demand and sell out quickly.
As the clock strikes 8:00 am, you attempt to load the webpage, only to find that the sneakers are already sold out in every size.
So, what went wrong here?
Problem Space
Here is the source of the problem: sneaker bots. These specialized mobile and web applications are developed with the purpose of swiftly checking out multiple pairs of sneakers in the shortest possible time. As a consequence, they overwhelm the website, rendering it virtually impossible for non-bot users to make a purchase.
Sneaker enthusiasts who engage in buying and reselling often rely on these costly bots to fuel their reselling business. Furthermore, individuals who possess these bots are frequently affiliated with sneaker groups, granting them insider information about upcoming releases.
Several barriers place ordinary consumers without bots at a significant disadvantage, exacerbating the challenges they face in acquiring desired clothing and collectibles.
Opportunity
The project goal is to create a new set of service strategies specifically tailored for small or medium-sized retailers. These strategies aim to enable them to successfully release highly sought-after sneakers to a wide range of potential buyers, while effectively circumventing the challenges posed by sneaker bots and avoiding manipulation by sneaker groups.
By implementing these strategies, retailers can guarantee an equitable purchasing experience, fostering a more inclusive environment for all customers.
User Research
Heuristic Evaluation
Goals
To identify the challenges and mistakes that companies currently experience when releasing high-demand sneakers to begin designing solutions that might avoid these mistakes
​
Explore existing possibilities that companies are already using and use this as inspiration to explore new possibilities through design
Digital
Ethnography
Goals
Understand how bots and sneaker groups are able to secure multiple purchases
Identify issues with the current anti-bot methods
​
Identify how people without bots are able to compete against bots
Research Through Design
Goals
Ideate and explore alternative service strategies for sneaker releases
​
Further explore early designed concepts to better understand what would be needed for an effective service strategy
Users without bots should be able to fully participate in the release
Fair Participation
01
Heuristic Evaluation
Given the lack of established heuristics specifically tailored to the sneaker release industry, I took the initiative to create my own set of heuristics. These heuristics were derived from insightful conversations I engaged in during the course of conducting my digital ethnography.
Users without bots should be approximately as (or more) successful as users who own bots
Success Rate
02
The number of participants should be relatively equal to the number of actual users
Real Participants
03
Bots shouldn’t be able to overwhelm the network (which would make it more difficult for users without bots to checkout)
Accessible Network
04
Users without access to large sums of money should have the same opportunity to purchase as users with access to large sums of money
Financial Access
05
Sneaker releases shouldn’t require buyers to have any certain knowledge or skill to participate
Prior Knowledge & Skill
06
The service strategy should provide buyers with a sense of exclusivity
Feels Exclusive
07
The service strategy shouldn’t create additional barriers for manual users
Fewer Barriers
08
There shouldn’t be any financial or technological barriers that prevent or reduce the ability of users to participate
Technology Barriers
09
Example of Evaluation Findings
Summary
Once you add your item to the cart, you are guided through a process before you can place your order. Online retailers use this process as a security method to reduce the overall number of bots making purchases.
​
The standard security process for online sneaker retailers can be seen below:
-
Captcha Checkpoint
-
Checkout Queue
-
Monitor IP Address
-
Monitor Billing Address
Findings & Takeaways
-
If you have a 0.9 Gmail rating, you might be able to skip the Captcha Checkpoint. Botters generally know about this method and prep their Gmails for this, whereas manual users are less likely to know this information
-
Antibot security will recognize if multiple orders are placed with the same billing or shipping information, but using virtual credit cards and manipulating your address will bypass the security
-
The four-step antibot security process simply isn't enough to protect against bots
Comparison to Heuristics
1
2
3
4
5
6
7
8
9
Next Steps
Now that I have a better understanding of some of the issues that are experienced in various sneaker release methods, I designed a digital ethnography that would provide me with the first-hand experience as a member of a sneaker cook group.
Digital Ethnography
Method
Join a sneaker group and use their information to participate in a variety of hyped sneaker releases
​
Interview users without bots to better understand what they need to prepare for a sneaker release
​
Interview users with bots (botters) to better understand what they need to prepare for a sneaker release
​
Interview managers and employees of a local sneaker retailer to better understand how they defend against sneaker groups and bots
Summary
I was notified within the sneaker group Discord that there would be a shock
drop/raffle hosted on a custom designer’s website (Fugazi). In order to participate in this raffle, you needed to reply to the SMS
confirmation. After entering the raffle, you had to reply ‘YES’ via text within a certain
amount of time (undisclosed) if you wanted to raffle entry to be valid.
​
This raffle was unique because it was a combination of a raffle and a shock drop. I was notified by the sneaker group that this raffle would end soon and to submit my entry immediately since there were so few pairs available.
Findings & Takeaways
-
During this raffle, I realized how broken local raffles are. I came across a thread that pushed members to enter local raffles. I have seen this before, but I came across something new. For high-value and high-demand sneakers, cook group members will enter raffles in all 50 states. If they win in a state that they don’t live in, then they will sell the account or winning email (depends on the raffle site) to
another cook group member who is located in that state -
Members will use ‘raffle bots’ to submit thousands of entries, whereas 'manual' buyers will only be able to submit one entry (if they are even aware that the raffle exists).
Next Steps
With a better understanding of how sneaker releases typically operate and how people with various resources attempt to place a purchase, I decided to turn some of my work into a physical artifact that could be referenced by a manager or owner of a sneaker retailer. My goal was to provide these stakeholders with the information that I've collected so they can create their own anti-bot service strategies that works best for their context.
Research Through Design
Method
Iteratively sketch and storyboard creative service strategies throughout the design process
Summary
At times, it felt like I was forcing myself to sketch during a period of heavy research. However, I found these sketches were useful in helping me think about what a successful sneaker release strategy would look like. When I started this project, I felt like there was no way to actually combat bots and sneaker groups. As I continued to ideate, I felt like there is certainly room for improvement with the current service strategies. This motivated me to further explore some of these ideas through storyboarding. I selected storyboarding
as a method because I thought it would be a good medium to explore different aspects of these experiences, such as various roles and touchpoints.
Service Strategy Toolkit
Strategy Example #2
Summary
Here is an example of how one of the service strategies that I designed would be presented to stakeholders. I initially considered creating a physical book or manual but I eventually decided to host the strategies within a series of PDF pages. I felt like this would make the service strategy toolkit easy to share and access for stakeholders across the United States.
Next Steps
Although my UX capstone project is finished and I am graduated from Purdue University, I still have plans to continue with this research. I think I have captured multiple issues within this system, but I feel that there is so much more to learn and improve upon.
​
For next steps, I plan to take this prototype and put it in front of various stakeholders and users to see what they think about it. I wasn't able to do this to the extent that I wanted during the semester, so I believe a follow-up to this project would benefit from a series of usability or desirability tests with some conversations with a variety of users and stakeholders.