I don't think we necessarily have a full appreciation of how difficult it has been for some agencies to manage signface life in a way that avoids premature sign replacements. There are three basic approaches out there: (1) replace all signs in a given corridor at a set time interval regardless of actual retroreflectivity; (2) log all signs by install date and add them to a sign log or sign inventory and replace all signs in a certain date cohort based on sampling of a few signs within that cohort to assess the retroreflective performance of the cohort as a whole; or (3) use remote sensing on a periodic basis to find the signs, measure their retroreflectivity, and mark them up for replacement as needed.
(3) is cutting-edge and the bugs are still being worked out. Finding the signs is done as part of photologging and machine vision is used to identify the signs. The computer typically has access to a catalog of standard signs but is not necessarily reliable at identifying one-off designs. (2) is conceptually straightforward, but experience suggests that many agencies struggle to maintain up-to-date sign logs. Several software solutions are available for sign inventorying but without automation, the task has to compete for scarce staff resource not just for adding new signs to the log but also for sampling retroreflectivity. (1) has the advantage of establishing a clear baseline and thereby simplifying planning for sign replacement without the need to dedicate resources to actually going out and checking sign retroreflectivity periodically.
In the fifteen or so years I have been collecting sign panel detail sheets (accumulating about 70,000 from 40 US states), I have seen a few instances of the same sign being replaced three or more times in a ten-year period without any changes in message. My personal favorite example:
SH 290/SH 349 Iraan/SheffieldHowever, what is going on here is also comparable to Griliches' account of the diffusion of hybrid corn. It takes time for agencies to study the "smart" options for scheduling sign replacement, which includes familiarizing themselves with evolving technology, and develop business cases showing that the initial costs (in equipment, software, and staff resources) involved in going "smart" will eventually deliver savings compared to sticking with an existing "dumb" approach. This all follows a logistic curve. At some point in the future many agencies will transit rapidly from "dumb" to "smart" and then there will be just a few holdouts left replacing the same signs pointlessly at too-short intervals.