This description is written as if Eldrazi were already banned into acceptable levels of playability. Before the Eldrazi craze occurred when OGW was spoiled, I was expending my brewing energies working on Grixis Faeries. I'm going back to that for the moment.
Playable faeries has three options in Modern: UBR, UBW, or UB (pre-Twinban, UR TwinFae was super fun). I don't think that Faeries currently has enough options to be playable. An unbanning of Ancestral Vision in April or a printing of Innocent Blood in Shadows Over Innistrad block would do it (and in happychristmasland where I get both, we're REALLY in town). But right now, you can play a slow grindy control deck. I don't like that, and doing so loses out on Faeries' biggest strength--being an aggro control shell. Forsaking the flexibility and power of a true aggro control deck (I'm of the PV school of Faeries thought) seems bad, because when an aggro-control deck is good, it is very difficult to pilot but the best deck in the format (see PV's articles on faeries for this argument. While written for a different format, the arguments still apply). So, until we get a few additional effects, UB is out.
That leaves Grixis or Esper. Each provides a strong way to interact with the early game, which Faeries desperately needs in greater quantities than Thoughtseize and Inquisition of Kozilek. Esper gets you Mana Tithe, which is easier to play around, but a little more flexible than Grixis' Lightning Bolt since it can also protect your opening hand Bitterblossoms with a little more upside later game (ie turn 4-6 as a Spellstutter follow up). Of course, while Bolt doesn't protect your hand, it's an early game removal spell and can WIN the game later on, rather than just having a mild upside. Slightly less relevant early game, much better late game. Esper also, of course, gets access to Path to Exile; but path is really a late game removal spell far more than Bolt is, because giving your opponent extra mana early game means you likely lose.
So, with that in mind, looking at secondary cards made the current decision an easy choice for grixis: Terminate and Kolaghan's Command along with Bolt provide a better maindeck, worse sideboard than Esper options.
In Esper, I'd care about: Mana Tithe, Path to Exile, Azorius Charm, and Zealous Persecution, along with the typical sideboard options of Stony Silence and Rest in Peace.
I have an esper list moderately assembled elsewhere. This list emphasizes the aggro of aggro control by trading some Fae bodies for Bolts and trading Cryptic for Kolaghan's maindeck. Otherwise, the deck plays like a Fae deck does. Bitterblossom will win you the game on its own, and is your favorite opening hand draw with support. Once you can land and protect a Bitterblossom, you can keep tempo permanently by switching all the rest of your plays to defense.
Treat your spellstutters as cheap disposable cards until/unless you have one late game with a Bitterblossom presence. As an alternative which is probably more important in the Modern context. Late game Spellstutters are hard counters for every playable counterspell in Modern except for Cryptic Command when played to protect a larger Fae threat such as a Vendilion Clique. If you draw some at that point in the game (turn 5 or later) at that point the card is a Counterspell and should be treated accordingly.
Scion of Oona is a counter-target-Abrupt Decay and/or a win condition. In prior periods when Faeries were worth writing about, the arguments for/against the card were very much either 0-1 or 3-4. In Modern, 2 has been very good to me. (See LSV and PV on faeries, respectively).
VClique can target yourself, which is often better in situations where having your opponent draw an extra card can end the game (combo decks, for example). On the flip side, you can also Vclique the opponent just to look at a (combo deck's) hand and not give a draw, as that clause has a "may" trigger.
MClique is an extremely versatile, powerful piece, and remains so in Modern. Flash, Flying, Timewalk in the form of it's Champion trigger, and a 4/4 body all in one! Immune to Bolt and Abrupt Decay, MClique can turn off your Bitterblossom once you have an established board of 3-4 Faeries if you run the risk of being bolted to death. You can tap a Mutavault and Champion that, you can respond to removal and cause a fizzle by taking the target, and you can layer the triggers such that if you have no Faerie in play and 8 Mana, you can end up with a single Mistbind Clique after casting two.
Final thoughts about Esper and Grixis Fae. Esper: This deck has a very solid longer game, and by natural inclination prefers to stick an early Bitterblossom and proceed to win with Fae tokens alone. Easy access to lifelink in cards like Sorin, Solemn Visitor or Azorius Charm, as well as a poor man's Force Spike and a fabulous sideboard make this relatively easy. Ghostly Prison is a real hidden gem here, letting you preserve more removal spells against decks like Merfolk and other flood the board aggro variants. Worse game against Affinity and Elves than Grixis, but a better game against every other deck (including Eldrazi, since that deck seeks to tap out every turn it can, and Eye mana can't be used to pay for Mana Tithe). Roughly even against Tron variants because access to Stony Silence really slows the Tron deck down. Once again, you have a solid answer to turn 3 Karn for super value in Mana Tithe.
Grixis: this deck pushes the aggro on the aggro control spectrum almost by default. Access to Lightning Bolt and Snapcaster Mage leaves you with an 8 damage swing in your favor on the end step of your opponent's turn 3 or 4, sealing the game sooner. Interestingly, improper sequencing can actually make the turn 4 BoltSnapBolt a liability because the cards push you to end the game quicker, and also provide fewer absolute answers. As a result, you stand the possibility of losing more life to Bitterblossom than the card is actually worth (balance this against lifelink access in Esper deck design).
Stronger in metas with more early creatures where Bolt is at maximum value, compared to a meta with a higher number of midrange and/or combo decks that rely less on creatures or have a lower deck percentage of creatures overall.
This deck also has a markedly worse later game thanks to Bitterblossom's life loss, if, as in metas described above, you stand a solid chance to lose your creatures by endgame frequently and can't BoltSnapBolt or KCommand. In that case, the current construction of creatures may be incorrect: a pair of Kalitas, Traitor of Ghet may be required for access to a solid late game plan that makes good use of Bolt and provides lifelink. I'd be cutting a Vendilion and a Mistbind in their place.
