We have developed a model that describes the transverse load degradation in Nb3Sn CICCs, based on strand and cable properties, and that is capable of predicting how such degradation can be prevented. The Nb3Sn cable in conduit conductors (CICCs) for the International Thermonuclear Experimental Reactor (ITER) show a significant degradation in their performance with increasing electromagnetic load. Not only do the differences in the thermal contraction of the composite materials affect the critical current and temperature margin, but mostly electromagnetic forces cause significant transverse strand contact and bending strain in the Nb3Sn layers. Here, we present the model for transverse electro-magnetic load optimization (TEMLOP) and report the first results of computations for the ITER type of conductors, based on the measured properties of the internal tin strand used for the toroidal field model coil (TFMC). As input, the model uses data describing the behaviour of single strands under periodic bending and contact loads, measured with the TARSIS set-up, enabling a discrimination in performance reduction per specific load and strand type. The most important conclusion of the model computations is that the problem of the severe degradation of large CICCs can be drastically and straightforwardly improved by increasing the pitch length of subsequent cabling stages. It is the first time that an increase of the pitches has been proposed and no experimental data are available yet to confirm this beneficial outcome of the TEMLOP model. Larger pitch lengths will result in a more homogeneous distribution of the stresses and strains in the cable by significantly moderating the local peak stresses associated with the intermediate-length twist pitches. The twist pitch scheme of the present conductor layout turns out to be unfortunately close to a worst-case scenario. The model also makes clear that strand bending is the dominant mechanism causing degradation. The transverse load on strand crossings and line contacts, abbreviated as contact load, can locally reach 90 MPa but this occurs in the low field area of the conductor and does not play a significant role in the observed critical current degradation. The model gives an accurate description for the mechanical response of the strands to a transverse load, from layer to layer in the cable, in agreement with mechanical experiments performed on cables. It is possible to improve the ITER conductor design or the operation margin, mainly by a change in the cabling scheme. We also find that a lower cable void fraction and larger strand stiffness add to a further improvement of the conductor performance.