The obvious hack for 13th Age + Eberron is to make the 13 Dragonmarked houses into the icons. Yeah, sure, you'll want some feats for Dragonmarks, blah blah blah, but that's the easy part. I'm more curious what happens when you refocus Icons in this fashion.
The most self-evident change may be the least important. Yes, replacing the Icons as individuals with organizations removes the possibility of a personal relationship, but as I've noted before, Icons are also their organization, and that organization is the part players will usually interact with. With the switch to houses, that part remains the same. Now, you'll probably need to populate the houses in a way that interests your players (since names and faces are still critical) but that's a good practice anyway.
What intrigues me is that in doing this you are explicitly *not* encompassing the world with the Icons. The Dragonmarked houses are just one axis of the setting, and using them (rather than, say, the various kings and such) makes a statement about what kind of game this will be. That is, it is a game that is going to center around the intrigues, conflicts and alliances of the great houses. The other setting elements still exist, but they will be encountered through this lens.
This fascinates me. It takes the broad, kitchen sink nature of the average setting, and pare it down to a thematic core. Want to use the setting again for a different type of game? Use a different set of icons!
Now, there's something similar that happens when you look solely at the subset of icons that players choose, and one might argue that you could offer any number of icons in a setting, then focus on using only the ones that players choose. This definitely makes for a strongly player-directed game, but I don't like it quite as much as the great houses approach because it produces too clean a dataset. There's nothing thematically tying the players interests together, and there are no rough edges of things that are important to the game, but not personal to the players.
Why does that matter? It provides a necessary contrast. When a setting revolves too strongly around players, it can start to ring false. One good safeguard against that is to make sure that the setting has elements that are important, but not personal to the PCs. Not too many, of course, but enough to make the world feel alive.
Anyway, I keep thinking of other ways to apply the Icons model, and for some reason, Eberron popped into my mind today, and I figured I'd capture it.
1 - And while we're at it, make them cool. Dragonmarks were always much more interesting as described than as mechanically implemented.