Practising Failure

Practising Failure

The information technology industry has developed a vocabulary for recovery that delivery companies have been slow to adopt outside their IT departments. A recovery time objective is the maximum time a system may be down before the consequences become unacceptable; a recovery point objective is the maximum amount of data that may be lost, measured as the time since the last good copy. Setting these objectives for each critical system forces a conversation about priorities that otherwise happens only during a crisis: which systems must be restored first, how much manual work can substitute for each, how long the business can survive without them.

The objectives are worthless unless tested. The more sophisticated technology companies deliberately break their own systems in controlled ways to see what happens, a practice sometimes called chaos engineering, and run exercises in which teams must restore services from backups under realistic conditions. Delivery companies can do the same with their physical operations: run a hub for a night on manual labels and paper manifests, reroute a region's volume to a neighbouring hub as if the first had burned down, operate a delivery station with its handheld devices switched off. Such exercises are expensive and disruptive, and they are almost always cheaper than discovering the same weaknesses during a real event. They also reveal something that no plan can: whether the people who would have to improvise know how to.

from Delivering the Future: The Companies, Technologies and Decisions Reshaping Delivery (2026)

Next
Next

Links