Construction & fleet field method

Don't Average Two Disagreeing Hours Sources

When a cab meter and a telematics hours feed disagree, name both readings, do not average them into one utilisation number, and keep one locked hours source for the file.

Imagine a plant or fleet lead opening a utilisation file when two hours readings for the same machine and window do not match. The cab meter shows one total. The telematics or gateway feed shows another. Someone suggests splitting the difference — “take the average and move on.” That average invents a third series neither source produced, and it hides which reading the office actually trusts. This article is a practical office and field method for that moment: keep both readings visible, refuse the average, and use a locked hours source for comparisons. It complements how to lock the hours source when a cab meter and a telematics feed disagree; it does not replace the separate rules that GPS movement is not engine hours and that a missing or delayed hours reading is not zero.

Illustration of construction machinery used to discuss disagreeing hours sources, not a customer result

Illustration for this article; it does not show a customer deployment, measured result or certified installation.

An average is not an hours source

A hours source is a named reading the office can still explain: which instrument, which interface, which export, and which person can show how the number was obtained. An arithmetic mean of two disagreeing readings is none of those things. It is a spreadsheet convenience that looks tidy and loses the conflict.

When the cab meter and the telematics hours field disagree, you have two evidence lines, not one soft truth in the middle. The same discipline used for engine hours vs powered-on time applies here: keep labels separate. Averaging two engine-hours candidates collapses the conflict into a fake single label.

What a disagreement can and cannot show

Before anyone blends the numbers, write down what each reading is, and what the gap does not prove.

SituationWhat it can supportWhat it does not mean
Cab meter ≠ telematics hoursTwo named readings for the same window disagree; the conflict is visible and worth resolving.That the midpoint is the “true” engine hours for utilisation.
Small gap (tenths of an hour)A possible rounding, lag, or sampling difference that still needs a named source for the file.That a small gap is safe to average away without locking a source.
Large gapA stronger signal that sources, clocks, installs, or cut-over windows are not aligned — pause utilisation comparisons.That picking the higher or lower reading “for fairness” is the same as locking a source.
One source blank or delayedFollow the missing-reading rule: keep the gap visible; do not invent a fill.That you may invent the blank from the other source’s delta, or from a track — see also GPS movement ≠ hours.

A disagreement is a reason to re-check install, export windows, and which instrument the office already locked. It is not a reason to manufacture a third series.

Do not average two disagreeing hours readings

Refuse the average as an office habit. If both readings exist, keep them side by side with event time and receipt time where relevant. If the file needs a single hours column for that window, that column must come from the locked source — not from (A+B)/2.

Also refuse close cousins of averaging:

  • Taking the higher reading “to be conservative” without naming that choice as a temporary rule and without locking it.
  • Taking the lower reading “to avoid over-claiming” and then comparing that machine to peers that use a different rule.
  • Blending cab meter, telematics hours, and powered-on time into one utilisation KPI.
  • Smoothing a jump after a gap by spreading the delta across days — that is a different error covered in hours jump after a gap is not one day’s work.

Keep both readings visible, then lock one source

The practical sequence is short:

  1. Name reading A and reading B (instrument, export, person who can show the screen or photo).
  2. Record the gap without filling or averaging it.
  3. Decide whether the existing locked source still holds, or whether the office must re-lock after a documented change — not a silent blend. See a change of hours source is not one continuous record.
  4. Until that decision is written, pause machine-to-machine or week-to-week utilisation rankings that depend on that window.

If you have not yet sampled one machine end-to-end, do that before scaling a blended rule across the fleet — see sample engine hours on one machine before fleet rollout.

What you can and cannot conclude

You can concludeYou cannot conclude
Two named hours readings disagree for a defined window.That the average is the utilisation truth for billing, hire, or KPI boards.
The office must use one locked source for the file, or mark the period unresolved.That “close enough” gaps can be averaged without leaving a written rule.
Hardware and platform choices can supply multiple hours-related fields; they still need label discipline.That more feeds automatically justify blending them.

Pause rules before utilisation comparisons

Do not rank machines, crews, or weeks on a blended hours column while a disagreement is open. Do not treat an averaged patch as continuous with earlier weeks that used a single locked source. When the conflict is resolved, write which source won and from which date — then resume comparisons under that rule.

For excavator rental and construction telematics contexts, keep the same evidence posture you would use on a sample brief: named source, named window, no invented fills. Product and sample pages such as telematics for excavator rental, GPS tracker sample validation, and GPS tracking for construction machinery are for configuration and evidence design — not a substitute hours reading.

Frequently asked questions

If the cab meter and the telematics hours differ by a few tenths, can we use the average?

No. An average invents a third series that neither source produced. Name both readings, keep the gap visible, and use the locked hours source for the utilisation file.

Is averaging the same as locking one hours source?

No. Locking means the office names which source the file will use and can still explain. Averaging mixes two sources into one number and hides which reading was trusted.

What if we need a temporary number for a weekly report while the disagreement is open?

Keep both raw readings and mark the period as unresolved. Do not publish a blended hours total as if it were a single source reading. Resolve or re-lock the source before comparing utilisation across machines or weeks.

Does this replace the rule about GPS movement and engine hours?

No. Movement is still not an hours reading. This article covers a different failure: two real hours readings that disagree, and the office averages them instead of locking one source.

Who wrote this

Ziyi Wang — LEXUNYU, Shenzhen. Independent sourcing and integration support for fleet and industrial telematics hardware. LEXUNYU does not manufacture devices and does not operate a SaaS fleet platform. This article is a practical field method, not a field test report.

Manufacturer documents and field results vary by model, configuration, site and market. Share the intended use case and evidence requirements for a written discussion — do not treat this page as a certification or a measured result.