The knowledge practices that work at 10 clients start showing cracks at 25 and break completely at 50. This is not because the tools were wrong or the team was negligent. It is because knowledge operations have different failure modes at different scales, and the practices that solve 10-client problems do not address 50-client problems.
Here is what shifts at each threshold and what to do about it.
Under 25 clients: the memory phase
Below 25 clients, most MSPs operate primarily on team memory. Senior techs know each client environment well. Runbooks are written but are more reminders than true standalone procedures. Credential lookups are fast because everyone knows roughly where things are. New hires learn by shadowing.
This works until turnover. When a senior tech leaves who held institutional knowledge for 12 clients, that knowledge does not transfer. The documentation that existed was too sparse to replace the person. The next hire needs to rebuild understanding from scratch, which costs time on every ticket during the learning period.
25 to 50 clients: the search problem
Between 25 and 50 clients, memory no longer scales and documentation density becomes the bottleneck. There are now enough clients that no single tech can hold all of them in working memory. Every ticket requires a lookup. The question is whether the lookup takes 30 seconds or 5 minutes.
Teams at this stage typically have some documentation but it is inconsistently organized. Some clients have detailed runbooks; others have a few notes. Credential organization may still be centralized in a shared vault with no client-level isolation. The search problem: with 50 clients and uneven documentation, a tech looking for a specific runbook has to know where to look and what to call it. If either is uncertain, the search fails.
The fix at this stage is structural consistency, not more content. Standardize the per-client template: every client has the same sections in the same order. A tech who has never worked a client before can still navigate to the right place because the structure is identical to every other client they have worked.
Above 50 clients: the governance problem
Above 50 clients, the challenge shifts from "can we find the right information" to "is the information current and authoritative." At this scale, multiple techs are updating documentation simultaneously. Content drifts. A runbook that was accurate six months ago may have been partially updated by three different techs in ways that are now internally inconsistent.
The documentation quality problem at 50+ clients cannot be solved by individual effort. It requires process: a defined update trigger (when does a tech update a runbook, not just when can they), a review mechanism for high-risk documentation, and visibility into which records have not been updated recently.
The access control inflection point
At under 20 clients, most MSPs give all techs access to all credentials. It is simpler and the team is small enough that everyone is trusted. At 50+ clients with a team of 10-15 techs, the same approach creates a compliance and offboarding problem. An offboarding at 50 clients means rotating every credential that the departing tech could access, which is potentially hundreds of credentials if access was unrestricted.
Role-based credential access is not bureaucracy at this scale. It is a practical constraint on the blast radius of offboarding or a compromised account.
The benchmark that matters
At any scale, the benchmark for knowledge operations is the same: mean time to information during an active incident. If a tech opening a ticket can find the relevant runbook and credentials in under 90 seconds, your knowledge operations are working. If they are calling a colleague, searching in a second tool, or looking through old tickets, you have a problem worth measuring and fixing.