My guess is that where there are 100+ users, every factor was
"near-perfect". I'd also guess that one of the near-perfect factors was the
application's requirements, with a high ratio of read/view to update
activity. BTW, because of the reputation of the people reporting the 100+
user DBs (Kaplan, Groh, and Forte, to cite just three), I believe those
reports.

In the days of Access 2, I'd start to get edgy when a client talked about 20
or more users, and downright uncomfortable when they talked about 50. But
30 - 70 users with adequate performance seems routine in 32-bit Access, and
even more is not all that uncommon.

One reason to be edgy when the estimate for number of users seems on the
high side is that, in over 40 years in the computer business, I never
encountered an initial user estimate that did not at least double during the
life of the application. On one Access project (fortunately a client-server
arrangment for reasons of reliability and recoverability), when I came on
board the project in 1995, the client's project administrator said, "35
users, never more than 50". When I last worked on that project in 2000,
there were just under 200 users, and more being added every week. So much
for firm estimates! (And, just for the record, the same held true in my
previous incarnation as a mainframer... figure at least twice as many users
as estimated, and maybe many, many more than that.)

 Larry Linson
 Microsoft Access MVP