This is a difficult question to answer with a high degree of accuracy, but could someone estimate to within an order of magnitude for me the chances of a zip code for a given address changing over the course of a year?

For example, would you say the chances of this are:

a) ~10%

b) ~1%

c) ~.1%

d) ~.01%

Just trying to get a rough sense of the constancy of these things for some code I'm writing. If the answer were ~10% I would proceed very differently than if it were ~.01%.

  • ZIP or ZIP+4? If just ZIP, I'd be shocked if it were higher than 0.01%.
    – nmpeterson
    Commented Aug 22, 2014 at 19:04

2 Answers 2


I maintain the Zip Code feature class for the USPS. Of our 34,000~ Zip Codes, I alter the boundaries of maybe 5~ a month.

Some Zip Codes contain thousands of addresses, some contain less than one hundred.

Hard to estimate an average yearly change per address, I would say .01%

  • 10
    Wow. I think you may actually have been the best person in the world to answer this question. That is really helpful. Thanks man.
    – Peter Berg
    Commented Aug 22, 2014 at 19:15
  • @Maksim Is this Zip Code feature class you speak of publicly available? Or is it an internal USPS only product? The only polygonal ZIP code dataset I was aware of was ZCTAs from the Census Bureau.
    – dmahr
    Commented Aug 22, 2014 at 19:39
  • 1
    @Maksim Understood. Can you tell us anything else about the dataset? For example, do the individual features for ZIP codes ever overlap? And how are point addresses (e.g. P.O. boxes) represented? Mostly I'm just curious about how well ZCTAs are a proxy for the official thing. Thanks for weighing in!
    – dmahr
    Commented Aug 22, 2014 at 19:49
  • 3
    There is no overlap between Zip Code polygons. We map P.O. Boxes as point locations, located within Zip Code boundaries. The ZCTAs are fairly accurate, we update our boundaries based on how the routes that deliver to those locations change.
    – Maksim
    Commented Aug 22, 2014 at 20:02
  • 2
    Our Every Door Direct Mail application displays our Zip Code boundary layer, and the routes associated to each.
    – Maksim
    Commented Aug 22, 2014 at 20:29

Years ago I was advised to use census topology for thematic mapping rather than zip codes whenever possible. The reasoning was that zip codes are designed by the USPS for mail delivery purposes, not for our mapping purposes. Zip codes can be arbitrarily redrawn anytime. Post office boxes can have their own zip code which for mapping purposes is effectively a point inside another zip code. You could conceivably have things arranged so that commercial buildings have one zip code and residences have another, all in the same topological area.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.