A single point of failure is any one vendor, person, system, route, or approval whose loss stops the mission. They hide in dependencies that have never failed, which is why nobody is watching them. You find them by tracing one critical output backward to everything it needs, then asking of each link: if this disappears tomorrow, what happens, and who notices first? The exercise takes a day. The crisis it prevents can take a year.
Supply Failures Are Mission Failures
In the Army a unit that cannot be fed, fueled, or resupplied does not continue the mission at reduced capacity. It stops. Our Principal directed readiness, budget, and logistics for more than 3,000 personnel across four combat deployments, and the lesson that survives from that work is simple: the plan is only as strong as the weakest thing it depends on, and the weakest thing is usually one nobody has looked at.
Civilian organizations run the same way. A school district with one bus contractor, a city with one payroll administrator who knows the system, a company with one supplier for a part that goes in every product. None of these is a problem until the morning it is the only problem.
Where They Hide
Single points of failure cluster in five places. Check all five.
One Vendor
A sole source for a critical input, often chosen years ago for price and never revisited. The contract has a termination clause. The operation does not.
One Person
The employee who is the only one who can close the books, restore the server, or calm the largest customer. Usually your best person, which is why nobody wants to admit it.
One System
A platform with no tested fallback. The backup exists on paper. Nobody has restored from it, and the last person who tried left in 2023.
One Route
A single road, port, carrier, or data link. The military calls this a line of communication, and it is the first thing an adversary targets because it is the easiest.
One Approval
A signature that cannot be delegated. When that person is on a plane, the organization waits. In a crisis, the wait is the crisis.
Why Nobody Is Watching
Dependencies that have never failed produce no data. There is no incident report, no ticket, no line in the risk register. The organization's attention goes to the things that break often, which are by definition the things it already has workarounds for. The dependency that will actually stop the mission is quiet right up until it is not.
The second reason is ownership. A single point of failure usually sits between two departments. Procurement owns the contract, operations owns the output, and nobody owns the question of what happens when the contract fails. Risk that belongs to everyone is managed by no one.
Find It in One Day
You do not need a software platform or a six-month assessment. You need a whiteboard, the people who do the work, and one day.
- Pick the three outputs that define the mission. Payroll going out. Students arriving. Product shipping. Not thirty. Three.
- Trace each one backward. What does this output need, and what does each of those need? Vendors, people, systems, routes, approvals. Stop when you reach things outside your control.
- Ask the question of every link. If this disappears tomorrow, what stops, how fast, and who notices first? Write the answer down. Vague answers are findings.
- Mark every link with no tested alternative. A backup that exists only in a document is not an alternative. That is your list.
- Rank by time-to-stop. The dependency that halts the mission in an hour outranks the one that halts it in a month. Fix the top three. Assign an owner to each.
Then test one alternative. Actually switch to the backup vendor for one order. Actually restore from the backup. Actually have the second person close the books. The test is the audit.
What a Fix Looks Like
Not every single point of failure needs a redundant system. The fix is sized to the time-to-stop and the cost of the outage. The options, in rising order of cost:
- Document it. Written procedure, kept current, tested by someone other than the author.
- Cross-train it. A second person who has done the task, not watched it.
- Pre-qualify it. A second vendor who has already passed onboarding and filled one order.
- Pre-delegate it. A standing authority for the approval when the principal is unreachable, decided before it is needed.
- Duplicate it. The expensive option, reserved for the links that stop the mission in hours.
Run the Audit With Us
The On-Site Battle Rhythm Audit is one day at your location. Ask for the dependency map as the focus and we will trace your critical outputs with your team, test one alternative before we leave, and hand you a ranked fix list with owners.
Services are provided on a best-efforts basis.
Need a second opinion first?
Executive Consultation is $275 per hour. Bring your three critical outputs and your vendor list.
Related reading
How Do You Run an After-Action Review?: the habit that keeps the fix list honest.
Questions
What is a single point of failure?
Any one vendor, person, system, route, or approval whose loss stops a critical output. The defining feature is the absence of a tested alternative, not the likelihood of failure.
How do you find single points of failure in an organization?
Pick the three outputs that define the mission, trace each backward to everything it depends on, and ask of each link what happens if it disappears tomorrow. Every link with no tested alternative goes on the list, ranked by how fast its loss stops the mission.
Is a backup vendor enough to remove a single point of failure?
Only if the backup has been onboarded and has filled at least one real order. A vendor that exists only on a list has not been tested and may fail the same way the primary does.
How often should the dependency map be reviewed?
Annually at minimum, and after any change in vendors, key staff, or core systems. Retest one alternative at each review.