Every site that’s run a clipboard CMH process for any length of time and then switches to telemetry has the same first-week conversation. The numbers don’t agree.

We’ve seen the gap as small as 4% and as large as 22%. Twelve to fifteen percent is typical. The interesting part isn’t the gap itself — it’s which side is wrong, and the answer is usually “both of them, in different ways.”

The clipboard is rounded

Operators round. Whether they mean to or not. A shift that ran from 7:14am to 3:48pm gets logged as “7am to 4pm — 9 hours.” Multiply that by 250 working days and a fleet of 30 machines, and you’ve over-counted by hundreds of hours per year.

Operators also forget. The end-of-shift logbook entry happens after a long day, sometimes after the operator has already left. If the maintenance person fills it in the next morning, they’re estimating.

And operators are people. If the supervisor is measuring utilization and yelling at low numbers, the logbook hours go up. If the supervisor is measuring utilization and rewarding high numbers, the logbook hours also go up. The clipboard is a social document.

The telemetry is also wrong, just differently

Telemetry CMH is precise. It is not always accurate.

The most common failure modes we see:

  • Wrong definition. The telemetry tracks ignition-on; the operator was tracking engine-running-under-load. They are not the same number; we wrote about that here.
  • Sensor dropouts. A gateway that loses connectivity for an hour didn’t see the machine running for that hour. If the dropout isn’t backfilled correctly, that hour disappears.
  • Cold-boot bookkeeping. A new telemetry install starts at zero. The machine is at 4,200. Until somebody enters the offset, the dashboard is wrong by 4,200 hours and looks confidently precise about it.
  • Engine swaps. Replacement engine, same chassis. The engine has zero hours; the chassis has 7,000. Both numbers are correct; reporting one when the other was needed is wrong.

A telemetry number with the wrong definition or a missed offset will be off by hundreds of hours and present itself with five-digit precision. That’s worse than a clipboard number that’s off by 12% and known to be approximate.

Who wins

Once both sides are configured correctly, telemetry wins. Continuous, unbiased, doesn’t get tired, doesn’t round. But “configured correctly” is the load-bearing phrase in that sentence.

We’ve stopped running automated reconciliations between the two for new customers. Instead, we run a 30-day shadow mode: telemetry CMH is captured but not authoritative. The operator continues clipboard logging. At day 30, we sit down with the ops lead, walk through the gap on every machine, and decide on a per-asset basis what the correct CMH offset and definition are. Then we cut over.

The 30-day overlap is annoying. It’s also the only thing that makes the cutover painless instead of a month-long argument about which number to trust.

What it costs to get it wrong

A fleet that under-counts CMH by 10% over a year:

  • Misses about 10% of scheduled service intervals, or hits them late.
  • Has warranty conversations get adversarial when the manufacturer’s number disagrees.
  • Loses the ability to plan capex based on remaining engine life, because “remaining” is measured against a number it doesn’t trust.

A fleet that over-counts by 10% does the opposite: services equipment too often, ages out machines that have life left, and burns cash on unneeded preventive work.

Both directions are expensive. The middle isn’t a happy compromise; the goal is the right number, not the average of two wrong ones.

What we recommend

  1. Pick the CMH definition that matches your service intervals. Document it.
  2. Capture an explicit offset for every asset at install time. Don’t trust the dashboard until that offset has been signed off by someone who looked at the on-board hour meter.
  3. Run telemetry and clipboard in parallel for 30 days. Reconcile per-asset, not in aggregate.
  4. After cutover, the telemetry number is authoritative. The clipboard is for incident notes, not hours.
  5. Audit quarterly. Walk the yard, read three random machines’ on-board meters, compare to the dashboard. If the gap is more than 1% per quarter, something is misconfigured — find it.

The goal isn’t to win an argument with the operator. The goal is one number, agreed on, used by everyone. Telemetry can be that number, but only after you’ve done the boring reconciliation work.