It is important to remember that not all analysts covering a company will necessarily be in the IBES database. Each broker should agree to provide their analyst estimates to IBES, but the association is voluntary, not mandatory. Thus, there may be analysts covering a company that does not have their estimates in IBES.
The number of estimates "NUMEST" in IBES Summary History is a reasonable approximation for the number of analysts covering a company. Note, however, that this number does not include estimates that were excluded by IBES because they were deemed to be outliers (excluded estimates are contained in the "exclude" file on WRDS Unix server or can be downloaded from IBES Detail History - Excluded Estimates) nor does NUMEST account for the number of analysts whose estimates were stopped according to IBES records (the stopped estimates are contained in the "stop" file on WRDS Unix server or can be downloaded from IBES Detail History - Stopped Estimate).
For a better approximation of analyst coverage, you can construct a measure of the "analyst following" from the Detail History file (see more on replication Summary History from Detail History, the sample program that you can use can be found at ibessum sample program). The advantage of using Detail History file to measure analyst coverage is that Detail History contains the "stopped" and "excluded" estimates, which according to our research, were not taken into account by IBES when constructing NUMEST variable in Summary History.