Relative Prize

Relative prizes reward the best performance among competitors rather than requiring achievement of a specific threshold. This approach is valuable when you want to encourage innovation but cannot precisely define what success looks like in advance, or when you want to maximize performance beyond any particular target. Competitors are measured against each other, with the best entry winning regardless of absolute performance.

The Netflix Prize demonstrated the power of relative competitions. By offering $1 million for the best improvement to their recommendation algorithm, Netflix attracted thousands of teams and ultimately achieved a 10% improvement—a target they couldn’t have specified in advance but recognized as valuable once achieved. The competition also generated significant knowledge spillovers as teams published their approaches.

Relative prizes excel at pushing the frontier of what’s possible and discovering innovative approaches that funders couldn’t have anticipated. They work well when performance can be objectively measured and compared, when you want to encourage maximum effort rather than just meeting a minimum bar, and when the competitive dynamic itself might spur greater innovation than individual grants would achieve.