Encouraged by the emerging disaggregation trend in optical transport networks, operators are willing to offer flexible and future-proof optical spectrum as a service (OSaaS) instead of classical capacity services. To define this modern service in commercial offerings, a commonly accepted, reliable, and easy-to-use service characterization method has to be developed. This is explicitly challenging in terrestrial brown-field, black-box scenarios, where open line system (OLS) data are unavailable or outdated to employ calculation or simulation-based methods. The objective of this work is to identify a characterization method that reliably captures wavelength-dependent performance variations, possible filtering penalties, and the operation regime of the optical spectrum services. First, all OSaaS configuration options in the OLS are systemized into four primary configurations, and their operational aspects are discussed. Then, two of the most used OSaaS configurations are tested in the lab and in a live network environment, providing links up to 5738 km in length. Generalized signal-to-noise ratio profiles are captured with a channel probing method, using a single-probe sweep and multiple simultaneous probes. Then, the accuracy and suitability of the amplified spontaneous emission loaded optical signal-to-noise ratio profile is tested in a lab environment. Finally, we discuss the required time and hardware resources to characterize the service and bring out the pros and cons for each investigated profile. In conclusion, we propose a single-probe sweep as a reliable and easy-to-use method to characterize the OSaaS in terrestrial brown-field networks.