What a SPOF is and why they stay invisible
A single point of failure is the only element a critical flow passes through: payments, shipping, data, decisions. While it works, it is invisible — the system looks reliable. Its failure turns it into the bottleneck of the whole company, and it turns out there is no backup, no substitute, and the instructions live in one person's head.
SPOFs stay invisible for two reasons: they have merged with routine («it just grew that way»), and nobody is tasked with finding them. SPOF hunting is the highest-return work in all of continuity.
Where to look: six typical places
- IT: a single server, channel or data center; a database without a replica; a homegrown system one developer understands; a single admin with full rights.
- People: the irreplaceable chief accountant; the only holder of a key client relationship; the technologist with everything «in the head».
- Suppliers: a sole supplier of critical input, one logistics route, one bank for all payments.
- Sites: one production facility, one warehouse, an office without which the call center stops.
- Data and documents: a key file on a local drive; licenses and accesses registered to one employee.
- Clients and channels: one client with half the revenue; one marketplace as the only sales channel.
A one-week search method
Take 5-7 critical processes (from your critical process register) and walk each chain with one question: «what happens if this element disappears for a week — is there a substitute, and what does the downtime cost?» Everything answered «no substitute» goes on the SPOF list with two attributes: cost of a day down (how to price it) and cost to fix. Sorting by the ratio gives you a ready work plan — the top will be holes that close almost for free.
Fixing: documentation, diversification, redundancy
Three strategies in order of cost. Documentation — nearly free: instructions, knowledge sharing, access handover. Diversification — a second supplier, second bank, second sales channel: cheaper than it looks if started before the crisis. Redundancy — a second site, a database replica, a standby channel: the only strategy that needs budget, and exactly where not to overdo it — reserve according to the cost of downtime, not according to fear. A large share of SPOFs is fixed by organizational measures within existing salaries — resilience growing without new spend.
A note on people
Human SPOFs are the most common and the most delicate: backing a person up can look like distrust. Working moves: paired work on critical operations, mandatory vacations (the best irreplaceability test), documentation as part of the role rather than a one-off chore. A deeper dive — in the key person risk article.
FAQ
How is a SPOF different from a normal risk? It has a binary profile: not «may degrade» but «stops completely». That earns SPOFs their own flag in the risk register.
How many SPOFs are acceptable? Zero is unreachable and unnecessary: the goal is to know all of yours and decide each consciously — fix, reserve, or accept with open eyes.
How to keep the list current? Review on every change: new product, relocation, supplier switch. In our practice it is part of the annual BIA cycle and visible on the dashboard among weak spots.