Service mode observing simultaneously provides convenience, observing efficiency, cost-savings, and scheduling flexibility. To effectively optimize these advantages, the observer must exactly specify an observation with no real time interaction with the observatory staff. In this respect, ground-based service-mode observing and HST observing are similar. There are numerous details which, if unspecified, are either ambiguous or are left to chance, sometimes with undesirable results. Minimization of ambiguous/unspecified details is critical to the success of both HST and ground-based service observing. Smart observing proposal development tools which ave built in flexibility are therefore essential for both the proposer and the observatory staff. Calibration of the science observations is also an important facet of service observing. A centralized calibration process, while resource-intensive to install and maintain, is advantageous in several ways: it allows a more efficient overall use of the telescope, guarantees a standard quality of the observations, and makes archival observations more easily usable, greatly increasing the potential scientific return from the observations. In order to maximize the scientific results from an observatory in a service mode operations model, the observatory needs to be committed to performing a standard data quality evaluation on all science observations to assist users in their data evaluation and to provide data quality information to the observatory archive. The data quality control process at STScI adds value to the HST data and associated data products through examination and improvement of data processing, calibration, and archiving functions. This functionality is provided by a scientist who is familiar with the science goals of the proposal and assists its development throughout, from observation specification to the analysis of the processed data. Finally, archiving is essential to good service observing, because a good archive helps improve observing efficiency by not allowing unnecessary duplication of observations.