Reed's Law
David P. Reed's 1999 claim that the value of a group-forming network grows as 2^n because each subset of users can form its own subgroup.
Reed's Law, formulated by David P. Reed in 1999, argues that the utility of large networks, particularly social networks that allow arbitrary subgroups to form, scales as 2^n, where n is the number of participants. The reasoning is combinatorial: a population of n members can form 2^n possible subsets, so any platform that lets users self-organise into groups derives value from the explosion of possible groupings, not just pairwise links. Reed proposed the law to push back against the assumption that Metcalfe's Law's n-squared bound captures all of a network's value. He distinguished three regimes: broadcast networks (value scales as n), pairwise networks (Metcalfe's Law, n-squared), and group-forming networks (Reed, 2^n). Critics including Andrew Odlyzko argue 2^n vastly overstates real value because most theoretically possible groups never form and most that do form are tiny, and that n log n better fits observed network economics. Even as an upper bound, Reed's Law is cited to explain why platforms with rich grouping primitives, such as Slack workspaces, Discord servers, and subreddit-style communities, can sustain engagement well past the point where pairwise interactions saturate. It is one of three competing scaling models invoked in discussions of Network Effects in Knowledge Platforms.