SAI Industrial ConsultantBusiness Consultant · Vadodara Start a Conversation
  1. Home
  2. Insights
Production & Planning Systems

Machine Downtime Recording: How to Record It So It Actually Falls

5 min readBy

How do you record machine downtime?

Write one line for every stop longer than five minutes: date, machine, stop time, restart time, minutes lost, one reason code, the action taken and who attended. Write it when the machine stops and when it restarts, not from memory at the end of the shift.

Ask an owner how many hours his machines lost last month and the answer is usually a guess. Ask why, and the answer is "breakdowns" or "no operator". Sometimes that is true. Often it is not.

You cannot reduce a loss that you cannot see. Downtime recording makes it visible.

Start with the simple sum

Take the time a machine is planned to work in a shift. Remove the breaks (lunch, tea, shift change). What remains is the available time. Every minute of it falls into one of three groups.

  • Run time: the machine was making parts.
  • Set-up time: the machine was being changed over, including loading and unloading.
  • Downtime: the machine should have been running but was not.

Run time + set-up time + downtime must add up to the available time. If they do not, some time is missing from your records.

Keep the three groups apart. If you call everything downtime, you lose the true downtime figure. If you leave set-up out, you cannot see it at all, and you cannot reduce it later.

The log: one stop, one line, one code

The log is a single sheet kept at the machine. The operator fills it. The supervisor checks that every line has a code. A good line looks like this (example only):

MachineFromToMinutesCodeActionAttended by
Lathe 110:0510:3025Tooling: insert chippedInsert changedOperator

The code is the heart of the log. "Machine problem" cannot be fixed. "Tooling, insert chipped, no spare at the machine" can be fixed with a box of spare inserts.

Why codes from the first day

Many plants begin with free writing and plan to add codes later. That does not work. Free writing gives words like "problem", "waiting" and "power". The history you have written cannot be sorted later.

Make a short list of codes, about 20 to 25, grouped under a few heads.

HeadTypical owner
Set-up and toolingProduction
Breakdown and powerMaintenance
Material not availableStores and purchase
Waiting for first-piece checkQuality
Drawing or papers not at the machinePlanning
No operatorProduction and HR
No order for the machineSales and planning

The head tells you who owns the fix. This removes the blame game. The question is no longer "who is at fault?" but "who owns this head?"

Weekly Pareto: find the few big losses

Once a week, add up the minutes by code and arrange them from the biggest to the smallest. This is called a Pareto. In most plants, three or four codes make more than half of the lost time.

Example only: in one week a machine lost 400 minutes in 12 stops. Tooling and set-up made nearly half of it. Breakdowns made about a third.

Do not try to fix everything. Pick the top two or three codes. For each one, write an action, an owner and a date. Review it at the next weekly meeting.

You cannot manage the hours you do not record. Write each stop when it happens, and the plant will tell you where to look.

From downtime to actionFrom downtime to action: 1. Available time (Run time, set-up and downtime); 2. Log each stop (One line, one code); 3. Weekly Pareto (Reasons in falling order); 4. Action, owner, date (For the biggest losses)1Available timeRun time, set-up anddowntime2Log each stopOne line, one code3Weekly ParetoReasons in falling order4Action, owner, dateFor the biggest lossesFrom downtime to actionFrom downtime to action: 1. Available time (Run time, set-up and downtime); 2. Log each stop (One line, one code); 3. Weekly Pareto (Reasons in falling order); 4. Action, owner, date (For the biggest losses)1Available timeRun time, set-up and downtime2Log each stopOne line, one code3Weekly ParetoReasons in falling order4Action, owner, dateFor the biggest losses
From downtime to action

Common mistakes

  • Writing from memory at the end of the shift. Small stops vanish and the wrong loss gets the blame.
  • Writing only the "big" stops. Five-minute waits for a crane, gauge or drawing add up to more than the breakdowns, and they are the easiest to remove.
  • Mixing set-up and downtime. You lose both figures.
  • Blaming the operator. Look at the system first. Was a spare insert available? Was the material at the machine?
  • No review. A log nobody reads is just extra paper.

How to start in your plant

  1. Ask one question: "How many machines do we have, and how many ran yesterday?" The answer is often a surprise.
  2. Pick the machine that limits your output and start the log there.
  3. Give it the code list on day one.
  4. Let the operator write, and the supervisor check every line.
  5. Include machines available, machines run and hours lost in the structured daily meeting.
  6. After four weeks, make the first Pareto and choose two actions.

Show the lost hours in money too: hours lost × the hourly cost of the machine. Owners act faster when they see rupees, not minutes.

Key points

  • Record every stop over five minutes, at the time it happens.
  • Keep run time, set-up and downtime apart, and make sure they add up to the available time.
  • Use reason codes from day one, grouped by owner.
  • Review the biggest losses weekly and give each action an owner and a date.
  • Start with the machine that limits your output.

Questions owners ask

What counts as downtime?

Time when a machine was planned to run but did not. Lunch and shift change are breaks, not downtime. Planned set-up is shown separately from unplanned stops, so the real downtime figure is not hidden.

How small a stop should be written down?

Write every stop longer than 5 minutes. Small waits for a crane, a gauge or a drawing look harmless, but together they often cost more than the breakdowns.

Why use reason codes instead of writing the reason in words?

Free words such as "machine problem" or "waiting" cannot be counted or fixed. A code can be counted, and its group also shows who owns the fix.

Written by Deepak B. Gondkar

Principal Consultant & Founder, SAI Industrial Consultant

Deepak B. Gondkar founded SAI Industrial Consultant in 2015. He brings 34+ years of industry experience and 11+ years of active consulting, and works on-site with MSME manufacturers in Vadodara and across India.

Want your downtime recorded in a way that leads to real fixes? Start a conversation with us.

Start a Conversation
Back to Insights