Well, I am far from expert, but we all strive to excel this question.
All in all, I think the crucial point is to effectively define research frontier of what has been studies so far, when it comes to the topic you want to research. Than you should decide onto which particular aspect, standpoint and further direction you would like to take. Of course, it is way easier to write it here then to properly do it - well known story. But, I always try to spend as much time as needed on initial browsing - which will consequently define further steps [e.g. methodology].
Anyway, Google Scholar is rather abundant with resources on this topic -> https://scholar.google.com/scholar?hl=sr&as_sdt=0%2C5&q=how+to+write+a+good+research+paper&btnG=
Good research paper requires good research work and new findings (novelty). It's a very dedicated work. First you read good publications in your domain. Try to understand your work in-depth and design the way of presentation of your research.
Please see the links; these may be useful to understand the basics of good writing:
A couple of observations: firstly, you appear to be making the assumption that you have already carried out some research,or perhaps you are trying to explore the writing process apart from the study being written about; secondly you have used the word impact without further explanation. Impact could be interpreted in very different ways. Journals and articles within them have established objective impact metrics but these do not fully explain the breadth of the construct.
Your question is essentially about writing quality. I would recommend that you first consider your writing ambition in context. A good model for this is Day's IPACE model.
Personally, I have found the most effective and productive approach to journal article writing is to have a single co-author with an agreed first author. The supervisor-student relationship is sometimes inappropriate for this because of the pre established power imbalance. It is better conceived as a mentor-mentee relationship where the mentor sometimes takes the second author role in order to give the mentee experience in leadership.
Writing planning can significantly improve clarity and reduce the time to draft an article. Critical thinking should also be applied to the amount of background reading and the pitch of the selected journal and article style.
Impact is often unplanned. It can be the result of fortuitous circumstances, but often follows from good preparation, good relationships, a positive motivation and having the right material at the right time which resonates with the academic community.
Why not study articles you perceive in your discipline as having high impact and their authors? What were the critical factors leading to their high impact? Perhaps you should even try asking the authors themselves. Maybe the writing element is really just a minor aspect to the whole process.
I want to understand one thing: have you already received an outstanding scientific result and only have to give it to the scientific community qualitatively? If this is the case, then a good example may be Grigory Perelman. He published three articles not in a prestigious scientific journal, but on the Internet, and yet these articles made him famous throughout the world. The quality of writing articles does not make any difference if the scientific result is very strong. Perelman wrote some parts of the proof very briefly without detailed explanations. Thus, the main goal is to get a strong scientific result, which will have the impact you are asking about. The degree of this impact depends on the urgency of the problem at the moment. Today's current problem may cease to be relevant tomorrow, it must also be understood. Sometimes there are cases that the received scientific result can be in demanded many decades or centuries later, but for now it will be a long time to lie in the general list and will not even be noticeable. For example, in music, some composers are well known only for their so-called short-song hits, and not for their large symphonic compositions, which can be very much. It is very difficult to predict the fate of a scientific contribution, so it's best to work honestly and hard to get the best result, and then the scientific community will tell you about its impact on science. My short answer is the following:
do not think about it, everything will be resolved by itself.
There were several good answers. So it may be difficult to write something new. I simply share my experience. The first trade-off is: a good research paper for what purpose? If the goal is to get scores for promotion from publishing in a prestigious journal, then you need first to find an interesting question, add a bit and do the literature survey. Otherwise they will simply reject it, especially in economics. If you do all this, they also can reject it, unless you have a famous coauthor. So people who are successful here, very often contribute marginally to science, but select a popular question.
If you want to write an article that is important for people, that will receive a lot of reads on RG even as a working paper, the best strategy (as I think) is to take a socially important question that practically was not investigated before. You may fail to publish it in a good journal (because editors dislike to be responsible for publishing something completely new), but (if the article is good) it will raise public interest. But then it might be better to become a journalist: they have much more reads comparing to scientists. However, some analytics of social processes (and not just reporting news) is more a scientific work.
In math it is a bit different: you need to have a strong new result, like theorem, algorithm, etc. But many publishable contributions there are also just small generalizations of the previous work.
In historical fields, gaining a thorough knowledge of the primary sources so that you are able to draw new conclusions from them that effectively explain other historical data is, in my judgment, the most important factor.
When your search includes a new idea with a compelling presentation, and includes the basic elements of the research, in this case inevitably the research will have a positive impact on specialists and readers in general.