There are many different ways to start learning a second (or third) language, but there are not really ‘right’ or ‘wrong’ ways. As someone who is struggling to self-learn a language, I have searched valiantly for the best way to learn Japanese – and everywhere says something different. So I can only offer the way think is most straightforward, and which is (kind of) working for me.
First: learn the alphabets. Japaense uses three, all at the same time, and it is necessary to learn all of them. Start with Hiragana (most commonly used, and will help with pronunciation) then Katakana (same pronunciations, different character set; for words taken from foreign languages, generally English).
Second: Build vocab. If possible, learn the pronunciation, the word, and the kanji at the same time. Building a vocab will help when you want to try reading.
Third: Grammar and sentence structure. This should be easier after learning some words, but should possibly be mixed with vocab learning. This is blurry for me.
Another point is that as a self-taught learner, I skipped around a lot trying to find ‘the best’ site or book to learn from. Often this meant having to drag through the hiragana and katakana tutorials after already learning them. An important feature for self-learners – or even those being taught – is to add a feature in which one can skip the things they already know.