I don’t think you will find a hard and fast rule, but knowing something about the construction of kanji helps a lot.
Check wikipedia here: Chinese character classification - Wikipedia
You are interested in compounds. These are mainly either 1. compound ideographs (all parts are taken for their meaning) or 2. phono-semantic compound characters (one part meaning, one part just shows the reading).
For the first kind there will be exceptions to any rule you might find, even though they were often simplified to match the style of the phonetic compounds. But there are still things like 譱 out there.
For phonetic compounds there will be only two main parts, a “sound” part and a “meaning” part. The meaning part is probably what you want to consider as “more priority”.
In your example 誘 has two parts, the kanji 秀 that shows the reading (“read me as しゅう”) and the determinative part 言 (the しゅう related to “words, speech” → invite). Granted, the actual reading of 誘 changed to ゆう (damn those Japanese), but that’s the general idea. The “pine” in there is not at the top level.
A nicer example is:
(Sound ゆ relating to words, sickness, mouth, feelings, carts, …)
You can check my userscript that integrated such information in WK.