Why we built a leaderboard before we had a community

Most products bolt rankings on after they have critical mass of users. We did the opposite — and the leaderboard ended up doing the work of the community we hadn't built yet.

The standard advice

The standard playbook is: build the core product first, get usage to a level where rankings have signal, then surface them. The reasoning is that a sparse leaderboard looks embarrassing — top producers with #1 next to a five-digit save count make the product feel established; top producers with #1 next to "12 saves" make it feel like a graveyard.

We made the opposite call. Built the leaderboard at week six — when the catalog was 3,000 beats deep but the user count was well under a thousand. The number-one producer that first day had something like 18 saves total.

What it actually did

Three things happened we didn't expect:

1. The leaderboard became the social object

Without a community page or feed or any of the standard community-product surfaces, the leaderboard was the most directly social thing on the site. Producers could see their rank. Artists could see who else was being saved. Numbers were legible.

Producers we hadn't talked to started DMing us screenshots of their first rank. They tweeted it. The screenshot itself was the social object — sparse leaderboard or not, having a visible position in something matters more than the absolute size of the something.

2. The numbers grew from being watched

Early leaderboards bend the curve they're measuring. The producers near the top start posting more, sharing more, improving more. We saw it within a week. Producers in the top 20 averaged ~35% more uploads in the following week than producers ranked 40-60. Causation is hard to prove on small samples, but the directional signal was strong.

Visibility creates a feedback loop. The leaderboard wasn't just measuring — it was steering.

3. The empty-state was honest

The thing we feared — sparse numbers looking embarrassing — didn't happen because we framed the page correctly. We didn't hide that the catalog was small. We just made what was there visible. An honest small leaderboard reads better than a fake big one. Visitors saw the actual catalog size, saw the actual save counts, and could decide for themselves whether to engage. The ones who did stayed.

An honest small leaderboard reads better than a fake big one.

The mechanism

Why does this work? Because rankings convert information into identity. A producer's beat count is information. Their rank is identity. Identity is shareable. Information isn't.

Once a producer has a rank, they have something to share. Once they share, the recipient sees the leaderboard for the first time. Each share is a free unit of distribution that wouldn't have existed without the ranking.

Numbers create acquisition. Not because the numbers are big. Because they exist at all.

How to do this without it being cringe

Some ground rules from the past month:

• Pick a metric tied to outcome. Followers don't measure anything real. Saves do. More on this here.
• Show the full ranking, not curated picks. A "featured producers" widget feels like marketing. A leaderboard feels like truth.
• Update continuously, not weekly. Frozen rankings feel like a magazine. Live rankings feel like a sport.
• Make rank shareable. Every producer profile should generate a unique OG card with the rank baked in. We took ~3 days to add this and it doubled producer-side share rate.
• Show the delta. "↑ +3 this week" is more interesting than "rank #5". Movement is news. Position is fact.

The takeaway

If you're building a marketplace, content platform, or any product where contributors care about visibility — don't wait for critical mass to ship rankings. The rankings create critical mass. Frame them honestly, tie them to a real metric, surface the deltas, and let producers do the distribution work for you.

We were six weeks in with a sparse leaderboard. A month later the same producers were sharing their ranks on Twitter with twenty thousand impressions. The leaderboard caused that directly.