Sometime in mid-2022, I noticed a pattern. My friends and I were spending a stupid amount of time tracking sneaker drops — refreshing SNKRS, watching Twitter accounts, missing releases by minutes because we found out about them too late. Every release felt like a coordination problem. The information was out there, but it wasn't where we were.

Most people would have shrugged and accepted L's. I happened to be a CS minor in Entrepreneurship at the time, which gave me a useful side effect: I was actively looking for problems to build something around. So one Tuesday evening I opened Discord, set up a server called Jersey Supply, and started populating it with channels for the major brands and a real-time alerts feed.

The first hundred

The community grew slowly at first. Friends, then friends of friends, then strangers from sneaker subreddits. By the end of summer 2022 we had maybe forty active members. By November we crossed a hundred. By spring 2023 we hit 350+ members, with roughly thirty active on any given drop day.

What I learned in that first six months — and what no class ever taught me — is that growth was almost entirely a retention problem disguised as an acquisition problem. We didn't need to find more people. We needed to make the people we already had want to come back.

Growth was almost entirely a retention problem disguised as an acquisition problem.

The data, kept on a spreadsheet

I built a Google Sheets tracker with about ten formulas across two tabs: one for inventory and restocks, one for member activity and engagement. Nothing fancy. But it was the first time I'd looked at numbers that I owned and asked them questions.

The questions were the kind you only learn to ask when you have skin in the game. Which channels have the highest message rate per member? (Yeezy and Jordan, by a mile.) What's our drop-off after a missed alert? (Brutal — if someone missed three drops in a row, the odds of them coming back fell off a cliff.) Which day of the week did our most engaged members join? (Saturdays, weirdly — it correlated with a content cadence we'd set up that turned out to matter more than I realized.)

350+
discord
members at peak
10+
sheet formulas
tracking inventory
11mo
total
operating window

What I'd do differently

I wound the community down in the spring of 2023 to focus on graduation and my first job. Which was the right call, but I left things on the table. If I were starting today with what I know now, three things change.

i
Tracking events, not just members.
I tracked who was in the server. I should have been tracking what they did — which alerts they clicked, which drops they bought through, which content kept them around. The shift from member-centric to event-centric measurement is the same shift product analytics teams make at scale.
ii
Cohort retention from week one.
A cohort is a group of users who joined in the same window, plotted by how many of them were still active N weeks later. If I'd built that view from day one, I would have spotted the third-missed-alert cliff six months earlier than I did.
iii
Pricing, even informally.
Premium tiers, partnerships with smaller resellers, even a tip jar — anything that turned engagement into signal about willingness to pay. I treated it as a hobby project. It deserved at least an experiment.

Why I'm still talking about it

Jersey Supply was small. It made no money. By the metrics that matter to most people, it was nothing. But it taught me something that's still pulling me forward four years later: building something for users you actually understand is wildly more interesting than building something for users you've imagined.

That's the thread that runs from Jersey Supply to my work with Gametime Vault now, and to the kind of product analytics co-op I'm targeting after Northeastern. Sneakers, sports cards, sports betting, live commerce — the businesses I'm drawn to are the ones I'd be a customer of anyway. That's not a coincidence. It's the whole point.