Prokaryotes are thousands of times smaller in mass/volume, so most cells (and maybe even most DNA) can be bacteria while these are nonetheless minor players.
This is definitely part of it, the estimates claiming 90% bacteria by cell count end up suggesting perhaps a half gallon of actual bacterial volume. As far as DNA, I think the parent stat can't hold for base pairs. Not as a volume issue, but because mammalian genomes are ~3 orders of magnitude larger than bacterial ones. Even at a 10:1 cell count, our base pairs would be 99% human.
Most importantly, though, the 10:1 stat is just outdated. Newer estimates put the number around 1.3:1 instead.
A side note to passing readers: thousands of times smaller in mass or volume means of course only tens of times smaller in diameter, which agrees with the pictures shown in the article and elsewhere.
I thought we ARE 95% germs, DNA wise.