Could one live of consulting for people with simple statistics questions or should one have a PhD in statistics for a consulting job to be financially stable? Some recent flow of basic questions on the website made me think that there is a sufficient market for this but I wonder if there is explicit proof from the field.

@MartijnWeterings, in university towns there are certainly people who do paid tutoring for undergraduate, or lower level graduate, statistics classes (stats 101). With a big enough school, there can be adequate demand for that. But it wouldn't be a great living.

@MartijnWeterings I know for a fact that (especially in the US tech sector) the increasing prominence of machine learning is driving a demand for people who are competent programmers or competent "subject matter experts" to gain skills in statistics and machine learning to improve their employment prospects/shift careers.

Often people will pursue an advanced degree to do this, but there are some employers who only care about what a person is skilled to do, rather than whatever paper they have

@MartijnWeterings One way to live off consulting is to have some clients who need your services to help them make substantial amounts of money. These clients will not be students and they won't be government workers, nor will they be people who are involved in the cost of doing business. Look to big pharma (biostats) and corporate execs (management consulting, data mining, business analytics).

Another way is to secure a large long-term contract with one client (think government consulting).

The worst possible business model for a consultant is to try to serve large numbers of clients who cannot pay much.

@MartijnWeterings I can personally attest that there are data science consulting firms which pursue interesting work with large organizations -- sort of like management consulting, but with less suit-and-tie-and-powerpoint bullshit.

There's usually a sharp divide in the maturity of their data processes -- either they've gotten as far as putting data into a database, or your job is to build a better mousetrap

I'm not sure which mod will eventually review this flag, but it seems that I have inadvertently flagged the same answer twice as "not an answer", once recently and once a two years back. Sorry! That's an oversight on my part.