Look at it like this: Steps of player retention.
1st step: Finding the game. A lot of people just won't even know your game exists at all. This is why Steam and advertising on social media is important. Yes, word of mouth can help but only after someone has already played the game and recommended it to their friends. You need to get them to play it first and if they don't know it's out there, there's not going to be people playing it.
2nd step: Judging the cover. Hey, images and blurbs are important. Those are the honey trap that get those who find your game interested in playing it and if you don't show something that will capture their interest you're not going to get anyone to play. No players means no word being spread means no players means no audience. So those who get past this point and actually play your game are going to be the group who net you more players.
3rd step: The first foray. The first 10-20 minutes of your game are vital! You might have a long-ass game with slow building set-up but you need to get the players to stick around past those first 10 minutes before they can even bother with the rest of your game. If they don't stick around they're going to either ignore the existence of your game completely or leave a review that probably isn't as shiny as you'd like. So nailing the presentation and showing something that will make the player want to stick with the game in those first few minutes of play are extremely important.
I just want to take a step aside here and point out that it's not just the gameplay itself that is vital so far. Presentation of what is shown should be high-quality tier. If the first thing a player sees is a glaring spelling mistake, some weird grammar, clashing colours, overly empty/busy map or mapping errors their thoughts on your game are going to sink. They will start the game already predisposed to shut it down if something they don't like comes along instead of giving it more of a chance because they see issues piling up. This is why testing and polishing are extremely important for a game dev. Spit and shine that first half hour til you can see your reflection. If they're still playing after that half hour they're more likely to forgive a few minor issues here and there but starting off a game with shoddy presentation? Yeah...
4th step: Retention. So they've gotten past the first 20 minutes and found something that has hooked them in. Great. Now you need to keep them in.
If a player gets to the 4th step, then they're more likely to be in it for the long haul unless you really mess up somewhere.
One thing I recommend is looking at when people start dropping your game and seeing what's going on at that point. Are some of the instructions confusing? Did you change something or introduce a new gameplay aspect? Did enemy difficulty ramp up or did something in the story change to make no sense?
But yeah, it can be hard to keep players to the fourth step if you ignore the first three - because the first three are vital to getting people to the fourth. And every step you will lost a LOT of people who might have otherwise played, due to whatever reasons. So put your best foot forward and do what you can to present your game well in all of those first three steps so that players at least get to the fourth step.
@Dreadshadow I wouldn't say a jRPG has to have some action to start off with - talk talk and scene scene can work just fine as long as it's interesting and not just text scrolling on a screen. Sometimes dumping the player in the middle of a battle can be just as much of a turn off as throwing text at them - it depends on the way it's presented.
For example, Trials of Cold Steel is a great game but the first 10 minutes or so was kinda bad (thankfully the company has a name behind them from their other games, but if it had been the first of the series I'd played I probably wouldn't have continued playing). It starts you off at a base full of enemies with a party of different characters who have a quick chat, go through a short dungeon fighting enemies and then watching some nukes being shot off from failing to stop the enemy. It didn't work because we didn't know the characters and didn't care about their plight, what they were doing or why. Only after playing the game and meeting the characters and their story do we actually care, but that start was really bad.
Another example is one of my games which has a 10-15 minute series of scenes giving you the background of the characters. They're fully automated and you watch them but everyone who's given feedback on the game have said that they really liked them and got interested in playing the game more because of that introduction. You don't actually get to play until about the 15 minute mark, but you've gotten to know the story of the characters, the hook of the story that makes you want to play more and find out the what's, why's, when's and how's of the rest of the game. Said game got cancelled years ago but I still have people PMing me at times asking when the game is going to come out because they loved what they saw and what they did get to play.
So a dialogue, scene-heavy start isn't necessarily a bad thing as long as it's done well. Hell, one of my favourite games, Suikoden V, takes a very long time to set up the main story. Again, it's a series that has a history and NAME behind it so if you've played the other games before you know that it's worth the weight (and boy-howdy, is it ever), but it takes literally 15 hours to get into the real start of the game. And it's worth it and it works.
It's how you do it, not what you do.