Outline techniques for raising the quality of operational processes and lowering the chance of human mistakes.
The idea
Process quality and human error are tackled together. Standardise and simplify processes, reduce hand-offs and re-keying, and move toward straight-through processing. Each manual step adds a chance of error. Let p be the chance of error at one step and n the number of independent steps. The chance of at least one error across all the steps then follows from the formula below. Chance = 1 − (1 − p)ⁿ. Build in error-proofing such as mandatory fields, drop-down lists and format checks that make wrong input hard. Use maker-checker or four-eyes review, checklists and clear documentation. Manage workload, staffing and fatigue, since errors rise at peaks and with turnover. Train staff, and encourage a culture where mistakes are reported and root causes are analysed rather than individuals blamed. Track error and rework rates. A checker lowers errors only to the extent the review is genuinely independent. The exam rewards linking each technique to the cause of the error.
Formulas
Probability that an error survives a multi-step process and a final review
P(error in output) = [1 − (1 − p)ⁿ] × (1 − c)
- p
- the error probability at each independent step
- n
- the number of manual steps
- c
- the share of errors the final review catches
Steps and review must be independent. Using the catch rate as the share that remains, or a catch rate measured in isolation when checker and preparer lean on each other, are the traps.
A practice question
A trade entry screen uses drop-down lists, mandatory fields and format validation so that invalid values cannot be entered. Which technique for reducing human error does this illustrate?
- Post-event reconciliationWhy not: Reconciliation finds errors after they are recorded; the screen prevents invalid input in the first place.
- Incentive alignmentWhy not: Incentives affect motivation; they do not make an invalid entry impossible.
- Periodic internal auditWhy not: Audit is an after-the-fact check on the process, not a feature of the entry screen.
- Error-proofing at entry - correct
Constraining inputs so that wrong values cannot be entered is error-proofing, a way of designing the process so that mistakes are hard to make.
A harder one
Why does reducing the number of manual hand-offs in a process typically lower operational risk?
- Fewer points exist where information is re-keyed or lost, lowering cumulative error probability - correct
- It removes the need for any control over the process, because errors can no longer occur at allWhy not: Fewer hand-offs reduce error opportunities but controls are still needed.
- It increases the independence between the maker and the checkerWhy not: Hand-offs are about transfers of work; fewer of them do not make reviews more independent.
- It shifts losses from operational risk to market riskWhy not: Process design does not change the type of risk borne.
Every hand-off is a chance for error. With n steps at error probability p the chance of at least one error is 1 - (1 - p)^n, which falls when n falls.
The hardest tier
A process has five sequential, independent manual steps, each with a 2% chance of error. A final independent review catches 80% of any errors made. The probability that an error is present in the output after the review is closest to:
- 9.6%Why not: Gives the probability of at least one error before the review and ignores the review.
- 1.9% - correct
- 7.7%Why not: Treats the 80% catch rate as the share of errors that remain instead of the share removed (9.61% x 0.8).
- 90.4%Why not: Reports the probability of no error at all, 0.98^5, rather than the probability of an uncaught error.
P(at least one error) = 1 - 0.98^5 = 9.61%. The review leaves 20% undetected: 9.61% x 0.2 = 1.92%, closest to 1.9%.
In the app: 6 practice questions on this outcome, spaced review that brings it back before you forget it.
Start free →More in Risk Mitigation
- Show the various ways firms deal with operational risk exposures and weigh them against each other.
- Contrast the kinds of internal control and illustrate each kind with examples.
- Describe automating controls, designing internal controls and testing them, noting the risks and difficulties in each and how to make them work better.
- Explain why new products, fresh business initiatives and mergers or acquisitions can create operational risk, and how firms can curb that exposure.
- Identify the approaches a firm ought to adopt to reduce the damage when an operational risk event occurs, and describe them.
- Describe how operational risks are transferred and how reputational risk is handled, and judge how well these work across differing circumstances.