The as-laid embedment is a key parameter in the design of non-buried rigid pipelines. Large discrepancies are observed, for example in the number and length of predicted free spans, or in the global response to thermo-mechanical expansion, when this parameter is misevaluated. The employment of wide ranges to cover the uncertainties and ensure safety may hugely increase the project costs by mitigating what could not be an actual concern. Conversely, there is no reliable method to accurately predict this embedment. The large strains imposed to the soil dishearten the use of finite element analyses, while physical model tests are as well subjected to the scatter in soil parameters along pipeline route (increased by the sampling under harsh conditions); and mostly to the dynamic motions that occur during the lay process which are difficult to quantitatively foresee. This last effect causes pipelines to embed deeper, as the seabed soil is both softened and physically displaced by the pipeline motion. Engineering practice has been using empirical models, which are always subjected to calibration and applicability range issues. This paper presents the results of feedback analyses of six rigid flowlines (ranging from 174mm to 353mm diameter) recently installed on clayey seabed offshore Brazilian coast, in water depths between 1087m and 1628m. The embedment values, measured by as-laid surveys, were used to reassess model parameters calibration. Results indicate encouraging agreement with published works. Statistical treatment of the data led to smooth distributions, which might be useful in future projects employing probabilistic design approach.