Years ago, through a strange twist of fate, I found myself part of a small development crew which created the product now known as Honeywell Pro-Watch. One of the things we marketed (and I believe they still do) was the broad ability to support about a half-dozen other manufacturer's panels, not just Honeywell's Mercury-powered panels. We would go into sites supporting non-Honeywell boards for existing construction, with the premise that new construction would use Honeywell/Mercury panels, and older construction would be gradually replaced. While that generally proved true for greenfield work, the existing installations remained. To this day I am aware of hybrid sites where nearly all the system is on Honeywell/Mercury, but the orginal non-Honeywell panels, existing when PW was installed more than a decade prior, remain in place.
There are some truths in this industry:
First, access control is, at its best, invisible. At its worst, it is extremely disruptive. Workers don't necessarily know if a camera is down, a camera retains some deterent value even if it's down, and certainly workers don't need to be rebadged if you change out cameras, but everyone knows if a door isn't working.
Second, many door controllers are slaved to a host system or appliance, at which a "top-box" integration is usually implemented, extending usable life of older boards. I know of installations of 1980's-era panels, which nonetheless participate in modern IT integrations, due to logic implemented at the head-end.
Full disclosure: I now work at Mercury and of course our partners find value in our compelling features like SNMP, Zeroconf, TLS, biometrics, etc. That said, I know of end-user sites with 100 Mercury panels, and maybe 5 Thorn/Grinnel boards from the 1980's, but will not replace those 5 until they fail.