The Two Prices of a Machine Breakdown
A machine breakdown has two prices: the repair, which the maintenance budget sees, and the production it interrupted, which nobody invoices. Factories that only track the first price systematically underinvest in exactly the machines that are quietly costing them the most.
Two numbers per machine
The vocabulary is old and still underused. MTBF, mean time between failures, says how often a machine or machine class betrays you. MTTR, mean time to repair, says how long each betrayal lasts. Together they turn maintenance from anecdotes (that overlock is always down) into a ranked list: this machine class fails every 46 running hours and takes 90 minutes to restore, and it sits on your bottleneck operation.
Both numbers exist only if breakdowns are logged as events with timestamps: reported, attended, restored. That is a floor discipline measured in seconds per incident, and it is the entire foundation.
Delivery impact is the missing column
The step most factories skip is joining breakdowns to consequences. A stoppage on a bottleneck station during a tight order is a different event from the same stoppage on a slack line, and the difference is measured in delivery risk: which orders, how many lost minutes, recoverable or not. When the breakdown log carries that join, maintenance priorities rank themselves by money and promises rather than by who complained loudest, and the 07:30 meeting can show open breakdowns with their delivery impact instead of a bare count.
The backup pool is capacity insurance
Sewing floors have a cheap hedge unavailable to most industries: spare machines. A deliberately maintained backup pool per machine type, with swap as the first response and repair as the second, converts many 90-minute stoppages into 10-minute swaps. The pool's size is a calculable decision, failure rates times repair times against the cost of idle spares, not a habit inherited from whoever bought machines last.
Preventive maintenance meets the plan
The last integration: PM windows and the production plan should negotiate, not collide. Servicing a line's machines during its changeover day costs nothing; pulling them mid-run costs the plan. That negotiation is only possible when maintenance schedules and production schedules live where they can see each other, which is one more argument for the connected model over the separate-systems status quo: uptime is not a maintenance metric, it is capacity, and capacity belongs to the whole factory.