See, I lack replayability in the levels of my game, is too linear and doesn’t last for long. I don’t really see this as a big problem, but I can understand why some people think it is. I have a few ideas on how to solve it, but my game is complicated to explain. The easiest comparison that comes to mind is Captain Toad Treasure Tracker: the levels are linear, short and small.

Simply making the level bigger is not feasible, since the camera generally needs to target the whole level at once, and adding more content is not an option, since I don’t have the resources and I know for sure that my game has ENOUGH content.

So how should I present it in order to gain replayability?

Again, I have certain limitations: the level can’t be too big, which obviously means that it can’t have several divergent paths either; doesn’t have to be linear, but given its small size it’s basically bound to be; I can’t (or should) increase the amount of content, and finally, procedural generation is not an option since the levels are handcrafted and really procedural stuff doesn’t improve quality.

Basically, if you had to develop a game similar to Captain Toad, how would you make it more replayable?

One solution (the only one that I have right now) is to focus more on bosses, since those aren’t limited by the size of the stage (but that doesn’t impact the levels themselves).

Also, I admit that I posted this question somewhere else, but am also putting it here to reach more people. And something that I forgot to mentiojn before is that I can't rely on many different character-based playstyles. I mean, I could, but that essentially forces the player to revisit 95 % of the game to see a 5 % difference, which I don't like. I also did consider reusing space, like Captain Toad. If you analize the goomba galleon level, you'll see that the player needs to backtrack a lot and the space is recontextualized after lowering half of the ship. The problem is that it can become predictible if done in every level and is difficult to implement given its nature (and this only makes the level longer, not more replayable).

I often hear replayability builds on things like playing a game with different RPG classes, or the motivation to see all story branches and endings.
But this does not apply to me personally. Not at all. Almost all games i've played more often than once are linear games, and one playthrough did not differ from he other.
The only reason to play them again was simple: The games are so good, it is fun to play them again, having the same experience again.

You could add secrets to your game, or hidden Stars, etc. This will increase my motivation to replay it, but it will not increase the probability, which only depends on how good it is.
You can make the game huge with a lot of potential exploration. But if the initial experience is not great, i will not replay it just to get more of it.

Personally i do not care about replayability at all when making a game. Doing so does not help to make it better. If you also think that's not too important, you're right imo.
In general i would say: The more options the player has, the more replayable. But if giving more options messes up your design or causes confusion, it's not worth it.

@JoeJ I know, right? That is my perspective as well. However, am making a puzzle platformer, genre that isn't that popular right now especially compared to rogue-likes, precisely because of the replayability. I read somewhere else the case of a puzzle platformer that seemed good, had good reviews and in general was received well, but regarding sales... well… yeah.

Reading that was partly what brought me here. In that same post the author said that part of the lack of popularity of said genre is the lack of replayability (if it has). So basically, is a marketing strategy. If you see someone online playing a game that catches your intrerest, but you are dumb and end up watching the entire walkthrough of the game (that results in not wanting to spend 60 bucks on it since you've already seen everything; sadly this happened to me once) then the game needs to have replay value to encourage you to buy it anyway. So basically, am worried over this and the fact that a lot of people think different than us and might not want to buy the game if it “doesn't last a 100 hours” and review it poorly because of this.

Thanks @Vilem Otte for your long reply, seriously. First of all, my game is a 2.5D puzzle platformer, so the sandbox aspect kind of doesn't apply here, same wirh simulators. I also forgot to mention that I already decided to implement a multiplayer precisely for the replayability, which is working well, but right now I wanted to focus on the single player replay value. I can consider my game story driven, but not really with many branching paths. Your suggestion of dealing with situations in multiple ways is a better solution. If puzzles are physics-based (and they are), they tend to have multiple solutions (just think of BOTW or TOTK). However, because my game has a central mechanic that is used for every puzzle (unlike the 4 items in Zelda), I generally can't create many solutions, since other solutions involve getting rid of that central mechanic. Take for example one of my puzzles: it was about moving and resizing the terrain (the central gimmick) to move 2 honey drops to the same spot to create a bigger one. It's physics-based, but it has only one solution due to the nature of the game. And because I can't make the levels too big, this problem becomes harder to solve.

feref28 said:
I read somewhere else the case of a puzzle platformer that seemed good, had good reviews and in general was received well, but regarding sales... well… yeah.

Sales a re a measure of success, but they do not necessarily tell how good the game is. If it's good but there are no sales, likely the problem can't be fixed with an even better game.
I assume your problem is elsewhere - marketing or other business mumbo jumbo, where i can't help with. ; )

Your conclusions / worries seem founded on bad reasoning, because:
Puzzle games may be niche, but some people want them.
And those people know that puzzle games are not about multiple ways, endings, story branches, or offering individual play styles.
They expect that there is only one way to solve the puzzle, and the fun is to figure it out.
After that the game is completed and done. But after some time, when we have forgotten the solutions but still remember the game well, we may replay it.

So to me, the assumption of adding replay value ‘by force’ to 'fulfill expectations' does not make too much sense.
On the other hand, maybe this attempt gives some new ideas we did not see before, because it's so unusual.

Now knowing it's a puzzle game, here's the first thing coming to my mind on adding replay value: Ship it with a level editor, and build up a community to generate content. (not saying that's easy!)
I don't know Captain Toad, but i play a free Super Mario clone for Windows: SMBX. I play it daily for many years already, because it has a Splitscreen mode so i can play with my wife.
They have a webpage and forum, and people release their episodes they have made with the included editor. The quality of the community content is very good on average. Some excellent puzzle episodes as well.
We have played some of them 5 five times or more, i guess.

Vilem Otte said:
not just “collect 500 stars”, which is mainly boring and annoying.

Obviously you do not know the satisfying feel of collecting a star. Not mention getting all of them. ALL STARS!!! \:D/
But maybe you will figure it out one day. There is always hope.
I still have not made it past the first dungeon appeared in Fallout 2. So many buttons on screen.
But it's still on my todo list to proceed, so i would finally learn what people love about RPGs… : )