Week 35 Day 4: How Repeatable Processes Scale Trust Across the Organization

In a team of five, trust is personal. Everyone knows everyone. Handoffs happen through direct conversation. When something goes wrong, you talk to the person. In an organization of fifty, trust must be systemic. You hand off work to people you have never met, through processes you expect to work. If the handoff process is reliable, trust scales. If it is not, the organization fragments into tribal groups that trust each other but not anyone outside the tribe.

Here is how process reliability scales trust. Consider the deployment process. In a small team, someone deploys and the team watches in real-time. If something breaks, everyone knows immediately and the person who deployed fixes it. Trust is personal: 'I trust Alex to deploy carefully.' As the team grows to 20, not everyone can watch every deployment. The team creates a deployment process: automated tests, staging validation, production canary, rollback procedure. Now trust is systemic: 'I trust the deployment process to catch problems.' When a new engineer joins, they do not need to be trusted personally before they can deploy. They need to follow the process. The process carries the trust. Scale this to an organization of 200. Teams that have never met each other depend on shared infrastructure, shared APIs, shared data. The team that owns the payment API has never met most of the teams that use it. The consuming teams trust the payment API not because they trust the people -- they have never met them -- but because the API has documented SLAs, automated monitoring, and a reliable incident response process. The process enables team A to trust team B's output without knowing team B's members. This is how organizations scale beyond Dunbar's number (approximately 150 people, the cognitive limit of personal relationships). Below 150, trust can be personal. Above 150, trust must be processed-based. The organizations that fail to make this transition experience the 'scaling crisis' -- the point where the organization is too large for personal trust but has not built the systemic trust to replace it. The symptom is: everything slows down, approvals multiply, decisions queue up, and the organization becomes bureaucratic not because someone designed it that way but because the absence of trusted processes forces everyone to seek personal verification.

Dunbar's number (Dunbar, 1992) -- the theoretical cognitive limit on the number of stable social relationships an individual can maintain, estimated at approximately 150 -- provides the empirical basis for the personal-to-systemic trust transition. Research by Zhou, Sornette, Hill, and Dunbar (2005) found that organizational communication patterns shift qualitatively around the 150-person threshold: below 150, communication is predominantly peer-to-peer (personal trust); above 150, communication becomes increasingly mediated through formal channels and processes (systemic trust). The 'scaling crisis' is documented by Greiner (1998) in his model of organizational growth phases, specifically the 'crisis of control' that occurs when organizations grow beyond the ability of founders and early leaders to personally oversee operations. Greiner's research found that organizations that successfully navigated this crisis did so by implementing formal coordination mechanisms (processes, standards, procedures that carry trust), while organizations that failed to implement such mechanisms either remained small, fragmented into autonomous units, or collapsed. Research by Gulati (2007) on 'managing network resources' demonstrates that inter-team trust in large organizations is primarily driven by 'institutional trust mechanisms' -- standardized processes, SLAs, and governance structures -- rather than by interpersonal relationships between team representatives. His research found that teams with reliable institutional trust mechanisms coordinated 40% faster and with 30% fewer conflicts than teams that relied on interpersonal relationships for cross-team coordination.