What is the training for the HR people that decide salaries/comp? I know there's glassdoor & co, but before that, HR people must have used some sort of methodology to figure out what the top & bottom range of a position is, no?
"Neutral" third parties conduct salary/compensation surveys. Many companies (HR departments) participate in these surveys and disclose their information to the third party.
The third party then packages it up into an anonymized 'market compensation analysis' that companies can subscribe to/purchase.
So companies can get a good read on how much is enough for their market, without ever directly revealing their hand to anyone except for the neutral middleman that's washing all the data.
A lot of HRIS (HR Information System) companies and payroll companies also package up and sell aggregate analytics on compensation figures. And the companies using those vendors are pretty complicit about their employee data being used in this manner, because at the end of the day they're the same ones buying the end result and peeking at the aggregate data to ensure they don't offer any more or less than they have to for a role.
If you want an answer to “what’s the median base salary for SWEs in the US?”, they can sell you the answer. If you want to know “what’s the P75 SWE salary?” or “what’s the P90 SWE salary?” or “what’s the P75 SWE total comp in the cohort of these 12 top employers of SWEs?” those are all also available, but numerically quite different from each other.
You could call any of them a “market reference point” with a straight face.
Levels has a data feed service they sell that’s an initial dump of N months of history and then a daily dump of additional ones. That was modestly priced (I think that was low 5 figures per year, but in any case it was quite inexpensive relative to running a software operation).
As an individual or small employer, I think levels is an amazingly valuable service (for free). When you get larger, buying these products makes more sense. (We’re remote-first and even still spend more on soda and coffee in a year than on salary data; I’m sure the salary data does more to attract, motivate, and retain than the soft drinks do.)