Runbook

Definition

A runbook is a step-by-step procedure for handling one specific, recurring technical situation – restoring a server from backup, rotating a service account password, clearing a print-spooler fault – written so that a technician who has never seen the environment can execute it correctly. Unlike an SOP, which describes how the business does something in general, a runbook is scoped to one task, one environment, and one outcome, with exact commands, credential locations, order of operations, and the points where the tech must stop and escalate.

Why it matters to an MSP

Runbooks are where an MSP's margin lives. The same ten ticket types typically make up 60–70% of volume, and handled from memory, time per ticket varies threefold between your best tech and your newest hire. A runbook attached to the ticket type collapses that spread: the new hire resolves it in the senior tech's time, first-call resolution goes up, and senior hours go back to project work. It is also how you hire: your first technician is productive on day one only if the common work is written down, even as one-page versions.

The other half is recovery. A restore that depends on the one person who knows the domain controller comes before the application server is a single point of failure with a phone. The per-client recovery runbook – restore order, credential locations, vendor support numbers, the local-versus-cloud decision tree – is what turns an RTO from a number in a contract into something you can deliver at 2 a.m. Keep them in your documentation platform, linked from the ticket type, and review them when the environment changes; an outdated runbook is worse than none because the tech trusts it. Structure and ownership are covered in documentation standards.

Related terms: SOP, RTO, Escalation