Salganik, Dodds & Watts (2006), Science 311:854–856: 14,341-participant artificial music market; "increasing the strength of social influence increased both inequality and unpredictability of success"; "success was also only partly determined by quality"

Claim: Salganik, Dodds & Watts ran a 14,341-participant artificial music market in which participants could see (or not see) what others had downloaded. Two findings:

• "Increasing the strength of social influence increased both inequality and unpredictability of success." When people could see others' choices, the winners ran further away from the pack — and which song would win became less predictable from quality alone.
• "Success was also only partly determined by quality." Quality matters; social-influence dynamics dominate the rest.

Implication for a new site: when reviews are absent, the "rich get richer" social-influence dynamic that helps incumbents is unavailable — making first-impression design and clarity carry disproportionate weight (Stanford Web Credibility — Fogg et al. (2002/2003): ~2,684 participants, 100 sites; "design look" present in 46.1% of comments (most-cited), ahead of Information Design/Structure (28.5%) and Information Focus (25.1%)).

Source: Salganik, M. J., Dodds, P. S., & Watts, D. J. (2006). "Experimental Study of Inequality and Unpredictability in an Artificial Cultural Market." Science 311(5762):854–856.

Confidence: Verified. Large-N controlled experiment in a peer-reviewed venue.

Caveat: The study models cumulative-advantage dynamics — not a per-site conversion lift. Use it for the structural point ("social proof produces winner-take-most dynamics, and a new site is excluded from those dynamics by definition"), not for a quantitative prediction.