there is no rule of thumb for the swabbing area, if you want you can swab the entire surface of your equipment. The adequacy of the area chosen is more an analytical issue (swab an area to get sufficient analyte to be quantified, but no so big as to saturate the swab with analyte), and 100 cm² has been proved as a good compromise. As a matter of fact, there can be a lot of sampling points smaller than 100 cm² simply because the swabbed piece is too small.
For micro a swabbed surface of 25 cm² is more usual, but only because this is the surface of a RODAC plate, and it is easier to handle the data if all the swabbed areas are the same.
Regarding the cream, I do not agree completely with you, in that it can be expected to leave “no” residue. If it is an emulsion, it has a non water soluble component, which may be uniformly spread over the surface after cleaning, and thus being not readily visible and even be oxidised or decomposed, and if this is the case, rinsewater analysis with TOC makes no sense, because if the cleaning water did not remove that component, why should the rinsewater? My advice is to perform the limit calculation with the toxicity of the most toxic component or degradation product, compare it with the 10 ppm limit, and take tho lower of both (most probably the 10 ppm limit). Then calculate the amount of cream that could remain if the entire formulation was made of the most toxic component (worst case), and spike a coupon/surface with the corresponding proportion of each component of the cream. Spike the entire surface, not only one 100 cm² spot, because a spot is far more visible than a uniform film covering the surface. If all of them are readily visible you have a good reason to convince the customer.