"15 Ways to Solve the Chicken or the Egg Problem" — CJ Gustafson (Mostly Metrics)

Why this is in the vault

Marketplace cold-start playbook with the "Flintstoning" pattern named explicitly. Useful as a reference when small-bet conversations turn into "we need both supply and demand" — particularly relevant for any future RDCO bet that has a two-sided shape (Squarely social/leaderboard expansion would have this if it ever ships, and a Sanity-Check-as-marketplace-of-experts variant has been floated). Also good Sanity Check fodder — the "Flintstoning" naming is the exact kind of concept-that-needs-a-handle move SC could lift.

⚠️ Sponsorship

Sponsored by LockSimple (appears in related-content slot). Lower-density placement than the Brex issues. Doesn't bend the editorial — the playbook is from James Currier / NFX research that CJ is summarizing. Note: full 15 tactics are paywalled; we have the visible ~4 plus the framing.

Issue contents

How-to essay with the four core frames visible on the free tier; the remaining 11 specific tactics are gated behind paid subscription. I'm filing the framework, not the exhaustive list.

Core thesis

Marketplaces fail at cold-start because they need both sides to show up before either side has a reason to. The fix is sequential, not simultaneous: pick the harder side, start in a deliberately narrow niche ("white-hot center"), and accept that the early version will look manual and unscalable. Scalability is the prize, not the prerequisite.

The four visible frames

1. Harder Side First. Identify which side has more friction (usually supply for high-trust marketplaces, usually demand for commodity ones) and concentrate cold-start spend there. The other side will follow if the harder side is real.
2. White Hot Center. Don't try to be a "global marketplace" on day one. Pick a niche so narrow it's almost embarrassing, dominate it, then expand tangentially. The narrowness is the wedge.
3. Flintstoning. Manual processes hidden behind a thin automated facade. The user thinks the marketplace is matching them; the founder is doing it by hand at 2am. This is correct early. The mistake is automating before you understand what you're automating.
4. Reference: Currier / NFX network-effects manual. CJ defers to the canonical NFX taxonomy for the deeper cuts.

Voice tactics

• Parachute metaphor: "like jumping out of a plane and building the parachute on the way down" — overused industry-wide but CJ owns it with the matter-of-fact delivery.
• Caption-as-aside: "Founder, chasing chickens" as image caption. Visual-pun-as-pacing-device. SC could use more of this.

Mapping against Ray Data Co

• "Flintstoning" is the harness-engineering thesis applied to marketplaces. Both ideas say the same thing: the early version is a human doing what looks like automation, and the path to actual automation runs through doing it by hand long enough to know what to automate. I should cross-link these explicitly. Naming the concept ("Flintstoning") is itself the move SC could lift — concept-with-a-handle compounds.
• For Squarely's hypothetical leaderboard expansion. If Squarely ever adds a friends-leaderboard or a community puzzle-trade feature, that's a marketplace. The harder side would be active puzzle-creators (supply). White-hot center would be a single niche (likely the founder's own X audience as the seed). Flintstoning would be the founder personally pinging the first 50 puzzle-creators. Useful as a not-yet-needed reference.
• For Sanity Check's marketplace variant. If we ever monetize SC as a marketplace of practitioner essays (founder + invited operators), the harder side is operators with both expertise AND time to write. White-hot center: data/AI operators specifically. Flintstoning: founder personally edits the first 20 guest pieces. Filing as a reference for the conversation if it comes up.
• Sanity Check candidate: "Flintstoning Your Newsletter." The founder has been Flintstoning SC since launch — manually picking topics, manually editing, manually engaging. The piece would be about why that's correct, not a stage to grow out of. Logging as candidate.

Related

• [[2026-05-11-mostlymetrics-owning-the-control-point]]
• [[../05-concepts/harness-engineering]]
• [[../01-projects/squarely/positioning]]