We type in the tracking number in the report. It would be easy for an automated program to extract those numbers, and to calculate the delivery time by tracking number (you can do it online for batches of 20 packages using the USPS online tools - you do have to "manually subtract days and times, however). No need for any human intervention. Collect the data automatically. Use a simple data program to calculate the percentage delivered early, on-time, and late. As far as damaged packages, if they want to identify them in photographs, they could use AI to do it. There are automated programs to identify burned potato chips on conveyer belts and then blow them off the line with a puff of air - I have seen them in potato chip factories. Have you seen the IBM TV ad that shows identifying wild animals using AI? These days, there is no need for "someone" to scan millions of pics. Is AI perfect? No. Is it good enough? Probably, and getting better.
As far as identifying early, on-time, and late packages, that is truly trivial. I have personally done it for thousands of shipments using Excel and for millions of customer call center calls using Minitab.
By the way, they already use a customer survey. The clerks are supposed to circle the survey request on the bottom of each receipt.
@sandyf wrote:
I doubt they have someone looking through the millions of pics of packages that go by or even have some computerized program to alert them if a package or its contents are or more likely look damaged. Yes they could do that, but do they? They could also have a camera on the exterior and interior of every post office so they can see if the flag is flying, or the interior is messy and they could have customers fill in a short survey as they leave to say whether or not they thought service was good and or fast. It would save us rate payers a lot of money.
Shopping Southeast Pennsylvania, Delaware above the canal, and South Jersey since 2008