Single point of failure
In short. A single point of failure is any one person, supplier, system or customer whose loss would stop or seriously damage a business, because nothing else can step in. Identifying these dependencies in advance lets a business add backups before a problem occurs.
The concept comes from engineering, where one component failing can bring down the whole system. In a small business the weak points are usually human or commercial: one person who holds the passwords, one supplier for a key part, one customer providing most of the revenue.
A quick test is to ask what happens if each person, tool or supplier is unavailable for a week. If the answer is that work stops, or that nobody else knows how to do it, you have found one.
A small agency whose only developer holds the hosting login is exposed. A café with one coffee machine supplier is exposed. A firm with a single customer providing half its income is exposed to that customer's decisions.
Mitigations are usually modest: documenting procedures, sharing access through a password manager, training a second person, qualifying an alternative supplier, and spreading customers. The aim is not to remove every risk but to make sure that no single event can halt the business, and to do it before it is needed.
Questions people ask
Is a key employee a single point of failure?
Yes, if the business cannot continue without them and nobody else has the knowledge or access. Documentation and cross-training reduce the risk.
Can a customer be a single point of failure?
Yes. If one customer accounts for a large share of income, losing them can threaten the business, which is why concentration is tracked.