Building Human Centered Algorithms in [email protected] Pt. 130 Aug 2021
It’s been a few months since I graduated from UC Berkeley.
Looking back, one of my proudest achievements was building the nonprofit student org [email protected] with my friends during my last year of school.
I think one of the most impactful programs we built was our resource matching system.
We first started with a simple MVP where users would fill out a Google Form and we would email back each respondent with a personalized list of resources based on what we could find in our custom database. Throughout the first few months of testing, we conducted focus groups and looked at our own analytics to see what users valued the most.
Here were some key takeaways we had:
- The main way people find resources is through word of mouth from their friends.
- Hearing how a resource is used by a friend contextualizes how the resource can be used by people
- Hearing from a friend lends credibility and makes it more likely for them to actually use the resource.
- Sometimes people don’t know where to start or what to search for when trying to figure out what they want to do in college.
- A key thing for students during college is discovery: finding their passions and interests. How can we aid in this discovery phase for students?
- For example, I might be interested in building tech with a focus on social impact. I might be interested in a variety of clubs, research, or professional frats. Which sort of organization would be the best fit for me? We found that this was dependent on a variety of factors including your background, experiences, and motivations.
From there, the business teams worked on developing a robust tagging methodology. The tech team built a few algorithms and tools that helped students discover resources more efficiently and helped speed up our internal resource matching system.
In my next blog article, I’ll talk about how we built these tools and systems.