What experts say
"(...) a key message of this book is that a survey isn't my first choice of method when you really don't know much about the people you want to hear from. You'll be much better off starting with some interviews or finding a way to observe people doing whatever you are interested in, and then maybe consider doing a survey later when you know more about what you need from it."
— Caroline Jarrett, in the book Surveys That Work
My two cents
Interviews lead to unexpected learning
I like to start bigger research projects with interviews because interviews are not 100% scripted conversations. This means that people will often go in directions you didn't expect.
And if you are okay to improvise a bit, you can follow people further in those conversations and suddenly arrive at places you would never had the question that made people say that.
These days, when I go in an interview, I limit myself to one key question, as I know the rest will follow up naturally. This shows how much interviews are to me more like an exploration than a search for an answer to an exact question.
Surveys are great when you know are under time pressure
But yes, I know, interviews take a shit load of time. Especially the recruiting (even if that part can be outsourced pretty easily today to recruiting companies like TestingTime).
In projects where time is a big pressure, having a tiny bit of data is better than relying on just your assumptions. So, if the timing doesn't allow for interviews, surveys can still be a good starting point.