Yes, mass-spectrometry-based metabolomics can be performed on as little as 20ul in certain cases, and 80ul is certainly enough. Below are a few studies from other authors and platforms, and there are many more out there.
Regarding your Theranos comment: You are absolutely right. This is something we will have to deal with, and where are actively working on our messaging.
The blood test accuracy problem that Theranos ran into wasn't so much about the specimen volume, but rather that they drew the blood from capillaries too close to the skin and thus it wasn't representative of blood circulating in larger veins. For some tests where they were just looking for any presence of certain substances that was good enough, but it couldn't work reliably for any test that needs consistent, quantitative results.
Does one do a benchmark by running the same tests, at different volumes with the same blood across multiple testing machines/methods, and then compare the results for accuracy/variance?
Yes, exactly. We are running internal validation tests to (a) compare the quality of the measurements with larger amounts of blood collected the "regular" way, and (b) validate that the storage and transportation at room temperature does not have detrimental effects on the measurements.