Trick question - let's start by trying to define "True".
Consider that you mix several pigments into a white paint base precisely following the recipe. This recipe will produce a 100% opaque paint when applied. You then spray the paint onto two test panels. On one panel, the final "appearance" is quite glossy: the paint went on wet and dried to a smooth, almost mirror-like surface. On the other panel, the finish has a dry, dusty look.
Same paint, same recipe, same substrate, different application technique.
What will the human observer see if he visually compares these two panels in a controlled lighting environment? The answer generally is that the smooth, glossy sample will "appear" to be much darker, richer, more robust, than the dusty, dry sample.
The difference in "appearance" here is NOT color: we know these panels were painted with the same can of paint - - there could be no difference in recipe. The difference is GLOSS; an appearance attribute.
Back to the question. "What measurement mode should I use to measure "True" color?" The industry accepted answer is Specular Component Included (SCI). If you configure a sphere instrument to collect all reflected light, specular as well as diffuse, then it makes little or no difference if the sample is almost mirror like or quite matte in appearance. The recipe and only the recipe controls the final "Color". Thus for match prediction, it is almost universal that companies use SCI when formulating color recipies.
While this makes sense in the laboratory, it can sometimes trip you up in Quality Assurance. Let's consider our glossy vs. dusty example in the context of a finished product. A manufacturer is making vertical tool storage cabinets. The left-hand door is painted by employee "A" who produces the almost wet looking result. The right-hand door is painted by employee "B" whose technique yields the dry look. Will this product be appealing to the buyer? Probably not! It will appear to have one gleaming, dark door, and a second weathered looking door. Looks used, damaged, repaired.
If you were to use your sphere in SCI mode, you would argue with the customer that the "Color" of the two doors was identical. But that's not how it "appears". No matter how you view the two doors they look different.
Enter Specular Component Excluded measurements (SCE). As the specular component of appearance IS gloss, and as the gloss of the two panels is very, very different, a measurement that can ingore or exclude gloss actually is quite sensitive to its change. Specifically, the recipe controls total light absorption and reflection ("Color"); it's X% no matter the wet or dry look. That said, the glossier the surface, the less of X% returning to the SCE instrument's optical receiver. If the surface were a perfect mirror (100% gloss), then the surface would "appear" to be a perfect black - no matter the "True" color.
Bottom line. For recipe creation in the lab, almost always SCI - "True" color. For Quality Control in the production process, a mix but leaning towards "Appearance" measurement - SCE.