What a photo at the meter actually proves.

Three things it proves

• Somebody was there. The visit happened, which is the first thing contested.
• What the index showed. Not what was typed — what was on the dial.
• The state of the installation. A flooded box, a broken glass, a seal cut: the context that explains the figure.

And what it does not prove on its own

A photograph carries less than people assume. On its own it does not establish when it was taken, since a file date can be anything; nor where, unless the frame shows something that identifies the place; nor which meter, if the serial number is out of shot. A clear photo of an index that could belong to any of four hundred meters settles nothing.

Four rules that make it hold

• Index and serial in the same frame. One photo, two facts, no argument about which meter.
• Readable without zooming. If the agent has to pinch to check it, so will whoever opens the dispute.
• Taken before touching anything. After wiping the glass or lifting the lid, it documents your intervention, not the state you found.
• Attached to the job, not to the phone. This is the rule that matters. Evidence is a photo tied to a date, a round, an agent and a meter — everything else is a picture.

What else is in the frame

A meter photo routinely catches a door number, a parked car, sometimes a person at a window. That makes it personal data in most jurisdictions. Two habits are enough: frame tight on the meter, and keep the photos as long as the billing dispute period requires, not forever. A library of street photographs accumulated by accident is a liability nobody chose.

The number to watch

The share of contested readings closed without a second visit. It is the only measure of whether the photos are doing their job. If the agent has to go back, the photo was taken but it was not evidence.