The use of multibeam echosounders (MBES) for mapping benthic habitat has gained widespread acceptability. Multibeam backscatter imagery provides an objective tool for scientists and managers to chronicle the extent and condition of the benthic resource. However, there are no standardized methods describing how best to process backscatter data to derive meaningful segmentations, although several acquisition parameters have been identified as having the capacity to affect the classification result. This research attempts to determine how the orientation at which a feature is insonified can affect classification outcome using commercially available software (QTC-Multiview), and to evaluate this significance related to vessel speed as a proxy for data density. A complex 2-km2 area of Stanton Banks, UK, was selected as the test site for the study. The area was insonified using a Kongsberg Simrad EM1002 MBES at perpendicularly opposing orientations, at two different vessel speeds within the same 24-h period. The classifications displayed 53% (k = 0.396) similarity at 4 m s−1 and 49% (k = 0.342) at 2 m s−1 from opposing orientations. Common orientations at different speeds were 68% (k = 0.583) similar (east–west) and 53% (k = 0.384; north–south). Most of the variation was in topographically complex areas, which coincided with shallow depths (<60 m). Meteorological and oceanographic conditions at the time the data were collected were evaluated as having had the potential to influence the outcome of the classifications. Interpretation of the results suggests that the orientation at which insonification occurs has a greater ability to influence the classification result than vessel speed using an image-based technique.