Personally I wouldn’t start with a huge list of negative keywords. Generally I like to build negative KW based off of the search terms that show up after some time of the ads running. Make decisions based off of data not assumptions
Your strategy depends on whether or not the “12-15 important keywords” are related or not. Are they similar intent? Are they short-tail keywords, long-tail keywords, or a mix of both? If you’re grouping too many unlike keywords into one ad group just because they’re all “important”, you will likely not see the results you’re hoping for.
Phrase match keywords will still match to search terms that would’ve matched to exact match keywords. Once you start getting traffic, review your search terms report and take note of how search terms are matching to your keywords. If the search terms are frequently matching via exact match, then maybe you should add some additional keywords on exact match, as well.
Group your keywords by similar intent, even if that means those 12-15 keywords are now divided into 3 different ad groups of 4-5 keywords each.
It’s challenging to know which keywords are working or not. That’s why we let AI find the right keywords. With platforms such as [Beam.city](https://www.beam.city/dna) DNA you can run your ad campaigns on Google Search or 12 other top ad channels using the best performing keywords. See how it works here: [www.beam.city/dna](http://www.beam.city/dna)