The Onsøy clay deposit is representative of soft structured natural clays. These are materials with brittle mechanical responses that are important to represent correctly in design. For this purpose, numerical models are employed, but calibration of the appropriate constitutive models is difficult. Calibration based on laboratory tests alone is problematic because the natural structure is damaged during sampling to an extent that is difficult to identify. In this work, we explored a calibration strategy for structured clay models based on the joint interpretation of laboratory and field tests through simulation. For this purpose, we made use of a large strain simulation of two mechanical in-situ test results where probes were pushed into the ground: The cone penetration test (CPTu) and the Marchetti dilatometer (DMT). The results obtained for the Onsøy clay showed that the DMT is advantageous in this respect, because it has lower measurement uncertainty and because the initial membrane contact pressure measured in the test is less sensitive to confounding factors.