The key factor is the delivery mechanism. The RNA has so somehow make it inside the cell intact, then migrate to the nucleus, be transcribed into viral proteins and then the viral proteins have to make it outside the cell to be detected by the immune system. Most of the time this process fails, but in principle (quite a few animal models and a few human studies), it can work just enough for seroconversion to occur.
Rogue RNA simply doesn't survive outside the cell and has little chance of being internalised, so traditionally these sorts of vaccines simply didn't work.
The key innovations are "nanotechnology" based (hey, that was my major at UNI!

). So there are various technologies that involve novel lipid/polymer/dendrimer coatings that allow the RNA to survive in tact and to enter the cell via endocytosis.
Some details on the technology:
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC5906799/
The second key is how and where it is delivered into the body. Evidence in humans shows that delivery using a needle and into the muscle results in poor immunity - most likely because the RNA never manages to undergo the process I described above. Efficacy is increased by targeting the lymph and high velocity fluid 'jet' injection. The latter likely improves efficacy since it spreads out the payload over a larger area.
There are DNA vaccines that work in much the same way, DNA is more stable, but it still has the other disadvantages described above, and there is additional risk of the viral genes being incorporated into the genome.
The vaccine that
@Jonathan Edwards was talking about is not strictly a DNA vaccine, but a recombinant viral-vector vaccine. The typical example is a an Adenovirus (which is a double stranded DNA virus) which has been engineered with recombinant technology to include proteins from other pathogens. Do note however, that these additional proteins won't actually be incorporated in the virion (complete virus particle) as some sort of hybrid virus. This vaccine also has the risk of genome incorporation.
As far as I know, this type of vaccine has only had widespread use (albeit in an experimental sense) to control Ebola virus. I personally doubt this type of vaccine will be approved for universal vaccination due to the additional risks compared to conventional vaccines.
https://www.ncbi.nlm.nih.gov/pmc/articles/PMC4994731/
The other key type is engineering bacteria to produce the viral proteins using recombinant technology. The downside to this is they might not fold in quite the same way as in a human cell and hence the resulting antibodies that are developed upon exposure might not work as well.
Compared to traditional vaccines, all of these have a key advantage, namely the basic technology can be re-purposed to pretty much any pathogen (so long as we know the genetic sequence), in a matter of weeks.