Scoping is one of the toughest tasks in game/app design. How much is too much? How much do you need to succeed? If you're designing well, all your features will flow in and out of a core loop, so understanding what's in scope is critical.
One technique I use is to find a comp--a game or app that has a similar set of features that you consider a good rough map for what you're trying to build. For example, if you were building a multiplayer word game, a comp might be Words with Friends.
Sit down with the comp and jot down each screen. Then, more important, sit down with a walkthrough/guide/doc and make a summary of the experience, from the user's POV. What exactly can she do in the app/game? What is the moment-to-moment?
Sometimes you only need a subset of the app/game. For example, maybe your goal is "a game on mobile that captures the experience of WoW raids." You'd only need to do a screen list and walkthrough for the WoW raid experience (and everywhere it relies on other game loops).
Once you have your walkthrough, make a copy and edit in your game's nouns and verbs. Keep the same format. Make sure everything you change is just a direct swap--no new features (yet). How much of what you want to build is captured? What's missing?
List out the missing elements, then evaluate each one against the walkthrough. If the walkthrough sounds complete as is... why do you need that feature? If you consider it essential, what are you willing to cut to make room for it?
Are things REALLY an even swap? Be honest with yourself. For example, if your goal is "Stardew Valley but in 3D" don't just tack "in 3D" on there like it comes without cost. For some people/teams, 3D is harder and more time-consuming.
For everything you add to the walkthrough, make a sincere effort to cut elsewhere. If you're adding the ability to create your own character, can you make fewer levels? Or fewer types of enemies? Or fewer weapons (and thus fewer animations)?
Try to make the walkthrough as narrow and small as it can be and still have it accomplish your core goals along with meeting your user's needs. Add, whittle, debate, whittle some more.
Once you think it's at the most narrow state it can be to meet your goals, then prioritize it, line by line (or feature by feature). Every item must have a distinct priority order, even if two features are intertwined.
Once you have features/lines prioritized, take one last look at the bottom 20% of your list. When you run out of time or money (and you will), how does the game/app look without those features? How can you plan now for those to be released after the main launch?
This priority list becomes your product backlog. Figure out any tech/art requirements for each line and slot them in only above that line. For example, if you don't need character wearables until customization at item 10, don't start that tech until you finish item 9.
One caveat: ALWAYS design and build with future work in mind. Build items 1-9 in a way that supports item 10. But don't get sidetracked by the bits that seem fun or interesting--you prioritized your features, so stick to those priorities.
Things will change, and as they do, revisit the walkthrough and revise it. Stick to your guns: try to cut a feature every time you add one, and prioritize the new items. Don't get distracted by the lure of "but it's just..." or "but that new game is so cool..."
Maintaining your own focus over the months and years of a game's development is really tough. Maintaining a team's focus and momentum over those years is even harder. Having a clear player-facing walkthrough everyone loves can be the key to moving forward.
