Everyone seems to be making post mortems, and one of them said "do it for future you", which convinced me. So. Post-mortem.
The idea for Finding Sam started out much more point-and-clicky, but we (@hangarter and I) spent a lot of time on the story (worth it), got in a fight and lost with Fungus (the Unity plugin that we chose to manage story interactions, because @hangarter had already used it, on smaller projects, and it had done the job well), and got both better and worse art than expected. So now it's very graphic novelly.
the Story
We talked a lot about the story, and once we had the outline and had asked "why?" (what lead to this thing happening?) enough times, we role-played it together, recording it, then reworked and rewrote it. It was fun, and helped get started. Overall, it was nice seeing something grow that none of us would have created on their own.
In the end it was 11 pages and 4500 words long (more like 4000 if you remove the speaker's name).

Fungus
Since our story is very non-linear, and it was important to be able to interrupt a dialog at any time, I lost probably half a day trying different ways to do that.
Also, Fungus saves all the story information in the scene itself, which makes me paranoid I'll lose it, so I decided to save every character dialog in a prefab. Bad Idea. The Worst Idea ever. Merge hell started then, and even when we broke the prefabs it kept going (to the point that at the end, we decided that only one person was to work on the scene/story at the same time; you know, the end, aka when you're struggling to integrate everything together, aka when you need all hands to work on the scene/story at the same time).
It's also lousy at adding dialog, it takes forever and way too many mouse clicks (you basically have to add it sentence per sentence, or the dialog won't interrupt like we needed).
I felt like it fought us at every step of the way and decided never to use it again (but @hangarter didn't seem to feel that way, and our needs were very specific).

the Art
I originally didn't want to make art, but programming instead for this jam (I ended up doing mostly story). Since I had failed to motivate any of my artist friends, I said I'd do the bare minimum for the art. Two hours on monday morning, I said when we did the planning, would be enough. I hadn't counted on needing to do the scene blockout in 3D (because who knows what a big living room where you could host a party looks like? Apparently not me), and a variation of the background for the flashbacks (which I'm very proud of), and character portraits. I ended up rushing a lot in the end, but once integrated into the game it looked surprisingly good.

Time management
The first evening, we made a planning: we had spent the day brainstorming, tracing the outline of the story and the choices that would make the game, and tested the basic functionalities we'd need with fungus. The evening would be for writing dialogs, until the end of sunday afternoon where we would stop and start coding (there was relatively little coding to be done). Art I could do in half a day, then putting everything together at the end of monday.
It was mostly followed, even if we spent the last night scrambling madly with exports and forgetting to set a game-breaking boolean.

Conclusion
I regret not allotting even a few hours to music and sound: I think it would have made the game much better for comparatively little effort.
And next time I'll try to form a team with more separate roles: not only do you lose a bit of time going from art to code to story back to code, and trying to understand what the other has done, but mostly I just missed the feeling I get when someone else does a part on their own and I look at what I've done and suddenly it's magically bigger thanks to their contribution (the art I did gets interactive or the programming I did gets pretty, seemingly on its own), which is a great feeling.
On the other hand, we did manage to finish the game, which is more than I got last ld (being in a team helps me organize and prioritize).
And we got great feedback very quickly which let us fix the (sometimes major) bugs that (since we hadn't actually tested it) had slipped into the game.
I kinda want to continue it (the story is about a quarter of what it should have been), but "I'll keep working on this jam game" is a well-known mirage. We'll see how it goes!