I found a curiosity in this area that the papers with wrong results have many citations as every next paper with right results would like to show their own distinct vision. But I do not suggest such a solution.
As well as I'm concerned, if your paper well be written or well be explained in the simplest language, it will get more citations and it's not matter it is review article or research paper. It's essential to select a hot topic for the paper as well (in Computer Science, Deep Learning, Machine Learning, Data Mining, Distributed Systems, Big Data, as well as IOT are the cases in point)