Bystander effect in open teams

Child note of bystander-effect: teams with open boundaries — open-source projects, volunteer communities — where members join and leave continuously. Here diffusion has three layers, and the countermeasures differ from both large and small closed teams.

Layer 1: who acts right now. The classic effect among whoever happens to be present: who fixes this issue, who answers this question. Same as the parent note.

Layer 2: who knows what, across time. Fluid membership erodes transactive memory — the shared map of who knows what. Totally reconstituted teams perform worse than stable ones (Moreland, Argote & Krishnan, 1996); partial change keeps the expertise structure but lowers how efficiently it is used (Lewis et al., 2007). High information sharing from departing members removes the penalty of losing even critical members; anticipated change is itself harmful, because people under-invest in specialization when they expect to leave or be left. A review of dynamic teams finds change harmless when it is predictable and anticipated, when transactive memory and shared social identity exist, and when the team runs on swift trust and implicit coordination.

Layer 3: the latent group at ecosystem scale. An open-source project has millions of beneficiaries and a handful of contributors — Mancur Olson's latent group (The Logic of Collective Action, 1965): a public good anyone can use whether they contributed or not, so rational users free-ride. The maintainer becomes the single point of failure: a small cadre, often fewer than ten people, carries widely used libraries, and the Log4j crisis showed what happens when that cadre is exhausted. Free-riding is invisible and normalized — the code is visible, the unpaid maintenance hours are not — and corporate users face structural barriers to giving back (donation rules, contract overhead, improvements that also help competitors).

Countermeasures, labeled. Verified: preserve transactive memory (departure handovers with high information sharing; train newcomers together with old-timers); make change predictable (announced departures, maintained onboarding paths); build shared identity. From Olson: selective incentives (benefits only contributors receive) and federations of smaller groups — Olson's own prescription for mobilizing a latent group. OSS practice (not lab findings): bus-factor management (co-maintainers, no single maintainer on critical paths); lower contribution barriers (contributing guides, good first issues, mentorship); triage automation (bots that label and route — a responder that does not diffuse); funding streams (public funding on the Sovereign Tech Fund model, Tidelift-style collective support, employer-allocated hours — time over money); norm-setting (maintainers are not support teams); and making free-riding visible (publish who uses versus who contributes).

Evidence status. Direct lab evidence for open-team countermeasures is thin. The transactive-memory findings are experimental; Olson is theory; the OSS items are field practice and case evidence. Do not cite the practice items as experimental findings.

Parent: bystander-effect · large-organizations · small-teams