Walk onto any shop floor and ask the metrology technician: "Why is this calliper on a 12-month calibration interval?" The answer, almost universally, is: "Because that is what the manufacturer recommends." Ask the same technician: "Has anyone ever verified that 12 months is the right interval?" The answer, almost universally, is: "No."
Calibration intervals are one of the most unexamined assumptions in quality management. We set them once — usually by copying the manufacturer's recommendation — and we never question them again. The instrument goes out for calibration on schedule. The calibration certificate comes back. We file it. The auditor checks it. Everything is in order.
But is the interval correct? Nobody knows. And the cost of not knowing goes in both directions.
The cost of too frequent calibration
If your calibration interval is too short, you are spending money on calibration you do not need. At a plant with 500 instruments on a 6-month interval, extending the interval to 12 months — if the data supports it — saves 500 calibrations per year. At an average cost of €60-100 per calibration, that is €30,000-50,000 per year in direct cost, plus the indirect cost of instrument downtime and administrative overhead.
I have seen plants where instruments were on 3-month intervals inherited from a previous quality manager who was "being conservative." The intervals had no data basis. They were set by fear, not by analysis. The calibration budget was three times what it needed to be.
The cost of too infrequent calibration
If your calibration interval is too long, you are producing parts with instruments that may have drifted out of tolerance. You do not know when the drift occurred, so you do not know how many parts are affected. You discover the drift at the next calibration — which may be months after the parts shipped.
I investigated a case where a coordinate measuring machine drifted over a 12-month calibration interval. The drift was 0.04 mm — small, but significant for the tight-tolerance feature it was measuring. Every part inspected by that CMM over a four-month period was potentially misclassified. The containment cost: €180,000. The root cause: the calibration interval was set for a stable environment, and the CMM had been moved to a location with temperature fluctuations that accelerated the drift.
Calibration intervals based on assumption rather than data are not conservative. They are arbitrary. And arbitrary intervals fail in both directions.
What the standard actually requires
ISO 9001 and ISO/IEC 17025 both require calibration intervals to be determined based on risk and supported by data. The manufacturer's recommendation is a starting point, not a permanent answer. The standard expects you to review calibration history, analyse drift patterns, and adjust intervals based on evidence.
This is almost never done. Most companies set the interval once and never revisit it. The calibration records accumulate year after year, and nobody looks at the data they contain. Which is a shame, because that data tells you exactly what your intervals should be.
The method I use
I implemented a calibration interval analysis programme at a plant with 1,200 instruments. The method is straightforward and does not require sophisticated statistical software:
1. Collect calibration history. For each instrument, gather the last five calibration results. Each result shows whether the instrument was within tolerance, and if not, by how much it had drifted.
2. Identify drift patterns. Instruments that consistently pass calibration with wide margin can have their intervals extended. Instruments that show drift approaching the tolerance limit need shorter intervals. Instruments that fail calibration need immediate investigation and interval reduction.
3. Adjust intervals based on data. Extend intervals for instruments with stable performance. Shorten intervals for instruments with drift. The adjustment should be documented with rationale — the auditor will want to see why you changed the interval.
4. Monitor continuously. Every calibration cycle produces new data. Review intervals annually based on the accumulated evidence.
The results
After implementing this programme, we extended intervals for 340 instruments (mostly gauges and hand tools with stable performance), shortened intervals for 47 instruments (mostly precision instruments in variable environments), and left 813 instruments unchanged. Net effect: calibration costs decreased by 22 percent, and calibration failures decreased by 40 percent — because we caught drifting instruments earlier.
The auditor reviewed the programme and noted it as a best practice. The standard requires evidence-based intervals. We had the evidence. Most companies do not.
Your calibration intervals are guesses. Some are good guesses. Some are bad guesses. But they are guesses until you prove them with data. Start collecting the evidence. The savings — and the risk reduction — are sitting in your calibration records, waiting to be used.
