My development process seems to go:

1. Write code
2. Believe that said code is amazing
3. Write more code
4. Revisit earlier code and start to doubt it's amazing-ness
5. Get frustrated that it could be done better
6. Redo 1 and repeat cycle

Seems a massive waste of time but I tend to like taking a different approach as soon as I find I'm getting stuck with the previous one.

I then get encouraged to take the quick/easy approach which seems like a backwards step and not worthwhile because I know it won't be as fast/efficient.

It depends what you're working on. In a capitalist society your output matters to your employer.

Spending a lot of time on writing fast/efficient code prematurely via. guesswork (multiple levels or refactoring) is probably seen as a waste of time. A great learning opportunity for you, but possibly not valuable to the company.

If the product does have performance/scaling issues, sometimes it's more cost effective to throw more resources (servers) at it than to spend time optimising. If that's not true in your co's case, well then your employers probably have it wrong.

Do a proto project in paper first with final objectives.
You can Change your approach at a point you can precise in the project and always keep In mind your final goal.
Had the same problem, solved it in a more lengthy way, that's the easy approach