Call the fleets closest to a rebuild, not the ones closest on the map.

Your call list starts from the installed-base register. It has the serial numbers, the last owner you recorded, the warranty and service history, and the parts orders that came through your own channel. For the newest machines there may be telematics too.

What it does not have is the rest of the fleet. Older machines that changed hands twice. Units sitting on a site whose owner buys parts wherever the price is best. A contractor who is working his machines much harder than the last time anyone rang him. None of that shows in the register, so the register cannot rank it.

So the list gets built the way it always has. Reps are spread across the territory by region and account size, and they call on the customers who pick up the phone. It works well enough until a fleet owner sends a machine to an independent shop for a rebuild you never heard about, or a site shows up in your parts data only after the work has been done. When the quarterly planning meeting asks why one account got the attention and another did not, the honest answer is that it was on the route.

This page describes a score meant to add a second opinion to that list. From high-resolution imagery, each machine we can see in your territory gets a rough age band, a read on how hard its site works, and a direction of travel over the quarters. You get the ranking and the three components behind it, so the list can be argued with in the meeting.

The offer is at an early stage and is shaped with the first territories that sign up. What follows is what we are building towards and what early access customers would see.

1. S1

   Send the territory

   Your service area as a boundary, or an account list with site locations.

2. S2

   Imagery pass

   A high-resolution satellite or aerial pass over the territory each quarter.

3. S3

   Machines scored

   Each machine we can detect is scored on an estimated age band, a duty proxy read from site intensity, and the change since the previous quarter.

4. S4

   Ranked list delivered

   The list comes back with the score components, ahead of your planning meeting.

Use the list before you build the quarter's call plan. It gives you a reason for who gets called first that you can put in front of the regional manager, and it sits alongside the parts forecast rather than replacing it. Where a score disagrees with what the account manager knows, the account manager wins.

These are the ways the score can mislead you.

L1This is not a fault signal

The score says nothing about whether a component is failing. It ranks fleets that look closer to a rebuild or a parts spend, using age and duty only. It is a ranking, not a forecast.

L2Duty proxy is inferred

It comes from how busy a site looks in imagery. It is not operating hours from the machine, and a quiet-looking site can still run its machines hard.

L3Age is a band, not a date

Age is estimated from visible cues and condition. It is not a build date or a maintenance record, and it cannot replace the serial number.

L4Units, not models

Machines are scored as detected units. We do not confirm make, model or serial from the image. Matching a unit to your register is done on site or account.

L5Resolution sets the floor

At 0.3 m a pixel covers about a foot of ground, so small machines are only a few pixels across. Very small machines may not be separated from their surroundings.

L6Quarterly, and a pass can slip

Cloud or a missed tasking can leave a site's score one quarter old. The list says which sites were not refreshed.