There's probably a business case to be written here about total addressable market. The world is a big place, but after all the people willing to pay $50-$100 for some novelty info buy their dossiers, where does new revenue come from to keep the lights on?
New people turning 18. It's no different than the FL studio model, which has worked admirably well for much longer than this has been a company. The population isn't static.
one would hope that they could find a way to be sustainable with a payment of $100 per record or whatever it was. (i don't have a good sense of what their costs per commercial sequence are.)
but the difference between food and 23andme, beyond the obvious necessity v curiosity, is that 23andme is a 1 time purchase. once someone has seen their results, they've extracted all the value and have no reason to buy anything else.