I think that your abstract should feature a statement about in which domain(s) your approach works better. When listing related work, emphasise on their weakness compared to your work (was that mentioned by someone before? cite their work!). Then, in your results/discussion, make clear that you indeed found an approach superior to existing work. If possible, show that through simulations/experimental results/surverys/data.
Also, when introducing your method in detail, focus on pointing out differences to prior work. Keep in mind that since papers are densely populated with complex information, it is easy for readers to skip a sentence or two. (it does not help that a lot of reviewers seem to be very busy).
My current advisor holds the opinion that writing a paper is like selling a used car: while you should not keep quiet about downsides, you should still be very convincing.
Clear and structured writing is also important: to keep the readers entertained, the paper should work like an essay where you argue that your approach is indeed novel.