We rely on what is reported by the airline and work under the assumption that what is published is based on their definition of on/off block (i.e. aircraft movement or door close/open). Whenever possible we rely on direct airline operational feeds but will use other trusted indirect sources such as a Global Distribution System (GDS) source or airports to supplement such reported data.
Unfortunately, we typically are not informed of what the specific criteria each airline utilizes. In some cases we find some airlines may use two different methods. What might be helpful is to look at a taxi-time analysis to identify variances in taxi in and out times.