[2022] Speed running.

"Speed running" is (typically)from a fresh start to any well defined end goal.
Using the examples from speedrun.com you have
Starter%; Complete the starter quests as fast as possible.
Grace%; Locate and kill Grace (The radioactive boar)
For a longer one, Tier5%; From a new save complete any T5 job.

As long as you have a well defined end goal that ideally produces an interesting run, you're speed runnin.

I would suggest the following as an example of a 'new' one.
SteelBlock%; On Navezgane from a new save acquire the materials to upgrade a single block to Steel as fast as possible.

It's certainly a different kettle of fish for a sandbox game. You could think up some arbitrary goal for yourself, but I think the better fit would be looking for those areas where the game provides structure. And quests seem like the go-to there, because they have an unambiguous start point and end point.

It's precisely because of speed running that you always have to clear the zombies in higher tier quests. By this I mean that the effort versus reward didn't work when you could smash and grab the courier satchel in under a minute, and get full credit from the trader for completing a Tier 3/4/5 quest. So one can argue quests are designed with speed running in mind: designed in a way to be more interesting and less trivial, even if you're speed running them.

Because the game has a high degree of randomness, I don't think requiring one to start from 7 AM on Day 1 is the best metric. Even on Navezgane, most any end goal for a speedrun (like your proposal to speedrun getting a bicycle from the trader) would depend too much on where you happen to start. So I think whatever you do, you'd want to strip away those random elements, for instance by starting the stopwatch when the player is at a predetermined set of coordinates. Again, the quest marker provides a good way to calibrate that. And quests are mostly self-contained - other than which biome they're in, it doesn't really matter where they are in a world.

Perhaps you could speedrun how fast you can kill every zombie in a Blood Moon, for a predetermined gamestage, horde size, zombie speed, etc. You'd want to pick a gamestage high enough to be challenging, but low enough that killing them all before morning is feasible. That would be neat, because you'd hopefully see the creativity of the speedrunner come through, with different base designs all angling for the same goal of optimizing killing speed.

Like, a lot of speedrunning is just trying to shave off fractions of a second here and there, copying what others do but doing it a little better. Perhaps where speedrunning in 7DtD could be different is that there'd be more opportunities for unique solutions. A fastest 'meta' would no doubt emerge in the end, but perhaps the game has enough complexity that it would take longer to get there than, say, speed running the campaign for a shooter.