The source universe starts from public PPP loan records — nearly ten million American businesses — but we deliberately don't include all of them. A majority of PPP borrowers were individuals: freelancers, contractors, gig workers. Real people, but not operating businesses a buyer could acquire.
A company makes the cut when any of three things is true: it has real staff (three or more employees), its estimated owner earnings clear $50k a year, or it's a professional practice — medical, dental, legal, accounting, insurance, financial advisory — which we include all the way down to solo, because books of business are genuinely bought and sold. The result is a dataset of businesses someone could actually buy, not a phone book. See Understanding the data for how the estimates behind that screen are built.