Perhaps it might be more useful for me to answer the specifics...though this might require a lot of context. Again, I don't run 5e, but in theory the same ideas can be loosely applied to 5e.
A question for 5e DMs out there (and, I suppose, players of 5e DMs). When a PC fails at something, what happens in your game?
For example: the rogue fails to pick the lock.
Picking locks is an application of the Thief move,
Tricks of the Trade. Full success means it succeeds no problem. Partial success means the player must choose between two of the following three options (the two being picked by the DM): "danger, suspicion, or cost."
I don't have a Thief in my game, but our Bard has taken this move as one of his multiclass moves, so it counts. (Character had a long-established connection with a Robin Hood-esque thieves' guild, so this move choice made perfect sense. Ironically, the character has ended up being very similar to a classic 1e Bard, minor Druid abilities and all, purely by the player following their interests.)
Each situation will make different options more plausible or interesting. So, for example, he recently picked the lock of a door on the "guest house" of a woman's private estate (far from the city) while trying to find out what happened to her and why her estate was so trashed. There was no social cost to picking that lock, so "suspicion" didn't make much sense to me. That meant "danger" or "cost" were the two main options. Danger is nice and simple, getting ambushed while picking or alerting unseen enemies to your presence are great forms of danger. Cost is trickier--the character has a lot of money and hasn't picked a lot of locks, so many standard costs would be a pittance to him. Instead, I'd probably go for something magical. Maybe at one point he has to use one of the magic daggers from his shadow-magic sword as a tension source, and it turns out the lock itself is magic and the two things react badly to one another. The player loves that sword, so negatively affecting it is a good motivating cost, opening up new story paths, and repairing it would be tricky as many of the people he could turn to to repair it are antagonists, or otherwise morally suspect.
Of course, the player actually DID succeed, so nothing of this came up at the time. But this is the kind of thinking I would have used if they
did miss the roll.
The bard fails to seduce the queen (or the barkeep, or whomever).
This one is less clear-cut, as there is no dedicated "seduce" move. If seduction comes up often enough, I'm sure I could write one, find one online, or mix those two approaches to generate something. But as it stands, it's the generic DW resolution. Full success (10+ on 2d6+mod) gives you most or all of what you wanted. Partial success (7-9) gives you either success-with-cost, usually for indivisible effects like "I attack the orc"; or it forces an uncomfortable choice, e.g. between taking a risk or playing it safe; or you only get small progress or fewer benefits, some of which you hate to give up. Straight up miss/failure (6-) means I as DM can make a "hard move" (actually doing something hurtful to the characters or which directly impedes them, rather than simply being the
threat of impeding them.)
So, if the Bard tried to seduce the Sultana and failed? Well, in this case, the biggest costs would be social, political, and resource-based. The party has enjoyed a good rep with the Sultana thus far. She generally likes them, and they have done good things for her personally and for the city-state she rules. It would be stupid of her to simply write the whole group off after a single botched seduction.
However, if he did this in a public place or in a conspicuous way, she would
have to respond publicly to avoid seeming weak or cowed. This would likely result in a loss of face for the Bard, probably being stripped of his diplomatic envoy status and being publicly admonished for his uncouth behavior; the party would find it much harder to make use of official resources, and the Sultana herself would cease to have the warm affection she has for the party, becoming more cool and businesslike with them.
That doesn't mean the relationship can't be repaired. It certainly could be. But it would be a challenge, especially due to losing the diplomatic envoy status.
Bartender is easier. He or she kicks the party out and says nasty things about them to other barkeeps/cooks/etc. in the area. Unlikely to affect their reputation or more "high class" joints, but the party would find the seedy underbelly much less friendly than it had been before. Again, this effect is not
irrevocable, but it won't change unless the party actively works to fix it (which is far from guaranteed.)
The party fails to find a path through the mountains.
This would be an
Undertake a Perilous Journey roll. A miss on this one delays the party significantly, exposing them to environmental dangers (in my game, this would usually be thirst more than hunger because it's a desert), costing precious time, and generally putting the party's objectives at risk (e.g. they would be passing the mountains to get to the other side, presumably to enlist the aid of the peoples there...a delay might let their enemies get in first and poison the well.)
The party uses up a lot of resources fighting cultists, and decides to take a long rest instead of rushing to stop the evil summoning.
This is what Dungeon World calls a
golden opportunity, and is an invitation to use a hard move against the players. I tend to be a softball GM but if my players did this I
would exploit it. Perhaps the summoning is nearly complete when they arrive--they now have mere
moments to stop it rather than hours. Perhaps the summoning is already "started" but not finished: their dalliance has allowed the Dark One to get a foothold in their world, but they can still prevent Her from fully entering and unleashing her dark powers unrestricted, and then
after that they'll need to find out how to banish Her completely from their world. Perhaps there is a sacrificial victim, and by the time they arrive, the victim is already dead--but freshly dead, so they can try to revive them or keep them in a
gentle repose type state until they can revive them properly. Etc.
What happens next? Are there stakes? Do you tell the party the stakes? What happens when a player wants to try again?
Some actions permit trying again, but most do not, because the world (or at least the party's knowledge thereof) should
change as a result of their actions. I don't usually think in terms of "stakes," but rather in terms of goals and significance. If there are stakes, generally I will tell them everything their characters could reasonably know (and I err on the side of telling too much, rather than too little.)
If possible, I'd especially like examples based on actual play. I'm not looking for general advice, so much as examples of how people have actually dealt with this in play.
Well, most of the above examples are "what if" scenarios rather than actual events from my game. But hopefully they are nonetheless illustrative.