How would you scope a first playable build for an eight-planet action RPG?

I’m prototyping Far Meridian, an original third-person action RPG. The current build is a small offline browser demo. I’m considering Unreal for the first native build and would value advice from developers who’ve shipped a 3D action game or run a multiplayer playtest.

The first 30 minutes would start in Beacon City, a modern hub. The player hears about a nearby settlement, prepares in town, explores a wilderness route, learns a gatekeeper’s attack patterns, and earns access to the settlement. The larger game would expand that loop across kingdoms, continents, and eight planets. Each planet introduces a local power system. The local level resets on arrival, while weapons and permanent stats carry over.

My proposed first milestone is one city, one route, one settlement, and one boss on PC. After that I’d test a small dedicated-server session. Multiple realms of around 100 players, mobile and console versions, and the remaining planets are long-term goals.

  1. What should this first build prove about combat, exploration, and progression before I add a second region? What would you cut from the first milestone?
  2. Which decisions about replication, save data, or world streaming would you make early, and what could safely wait until small multiplayer sessions are fun?
  3. What would you measure before adding more players per realm or beginning mobile testing? If you went through a similar transition, which test exposed the first bottleneck?
  4. How would you test whether a planet-level reset feels worthwhile and whether later planets are distinct enough to sustain the game through the finale?

A concrete example of what you built first, what surprised you in testing, or what you would do differently would be especially helpful.