We play games because they amaze and delight us -- good ones, at any rate. We still often hear that the right way to launch a F2P game is "create the minimum viable product," launch the game, and if it gets traction, invest in improving it -- with "improving it" meaning A/B testing features. To put it more brutally, this approach can be described as "throwing shit against the wall and A/B testing it until it doesn't stink."
This doesn't work, and never has; without a sound core game, no amount of tweaking will save you. And what constitutes a "minimum viable product" is subject to the same ramp up in quality over time as every other aspect of game development; the pure HTML social network RPGs that were successful in the early days of Facebook gaming would never get traction today, because we expect much more from browser games now.
Moreover, it's now true, in F2P games, as it always has been in the conventional market, that "you only get to launch once." Data shows that a game's first cohort, the first people who flock to it, monetize more effectively than later ones; and once you have lost a player, they very rarely return. You want to launch not with a "minimum viable product;" you want to launch with something amazing, because that's your best hope of snowballing a big audience and achieving long-term success. And of course, you can do geoalphas and short-term tests with small audiences to massage your FUE and your metrics before going full-bore and doubling down on your marketing spend; but you absolutely should not be launching your game broadly until you have something polished and cool.
Your players are not guinea pigs to be A/B tested; they are living human beings, and if you are in the entertainment business, it is your obligation to provide them with quality entertainment, from the very start.
We need to stop thinking in start-up mentality terms, and start putting in the extra 20 percent effort that produces polished games.
Spam Not; Message Endogenously
Time was, you heard a lot of talk in the F2P space about "K-factor." It's a term borrowed from epidemiology, actually; K-factor is the number of healthy people infected by one sick person, and in fighting an epidemic, you try to get K-factor as small as possible, by doing things like getting sick people to stay home, wear face masks, take medication, and so on. In games, we use it to refer to the number of new players recruited by each player acquired, and we want to increase K-factor. Once upon a time, Facebook games saw very appreciable K-factors, which meant that a new game could grow by leaps and bounds -- and also meant it was clearly in the interest of game developers to get players to send as many messages, and make as many newsfeed and wall posts as possible, because their visibility attracted more players.
So every game encouraged you to post virals at every turn, and in some cases pretty much coerced you to do so -- hence the "staffing" mechanic common to so many builders: to open a new building, you need six friends to agree to staff it, invite them today.
I have quit games when staffing requirements became too onerous (14 friends? you've got to be kidding); and I am sure I am not alone. They do foster messaging; they are also bad for long-term retention. And what are we planning for again?
Here's an example of dishonest practice: Most games ask you to invite friends early in the first user experience. Fair enough, as far as that goes: But many also do not allow you to close the pop-up that asks you to invite. There's only a single "invite" button. The user is forced to click it -- or quit. This pops a friends-list dialog, and you can close out of that, and return to your game without sending invites; but you have just told your players "We treat you like sheep and try to trick you into sending virals." This is the opposite of respect.
A/B test your invite dialog with a close button and not, and I guarantee you will see two things: First, yes, more invites will be sent if there's no close button. And second -- you will see more drop-off at this point in your FUE, as some players say "screw you, I'm not doing that." In other words, you are trading off keeping this user against the hope of acquiring some other user; which is better in the long term is hard to say, but it is not hard to say, "You are disrespecting this user." Don't do that.
Messaging is even more problematic in mobile F2P; yes, a player can easily send messages to someone in their phonebook, or if they signed on via Facebook Connect, to FB friends. But a phone is not a social network, and this kind of messaging is even less natural than it is on a social network; few of your players will do so, unless you force them to -- and forcing them to is unethical.
The reality is that K-factor is now small, mainly because Facebook has nerfed the virals (and messages were never a main source of players in mobile). You will no longer see newsfeed or wall posts from games, except for those you already play. Typically, no more than 3 percent of your users will be acquired by virals; is it really worth annoying your players for such a small number?
Virals still are useful in terms of player retention; a player who gets a game message from a friend is more likely to return to the game, which you want, of course. But this is best fostered by positive messaging, like "free gifts," rather than negative barriers to play.
If you want to foster messaging, don't force it; instead, make it endogenous to your game, make it fall naturally and comfortably out of your design. As an example, in the (no longer extant) game, Knighthood, you played as a medieval nobleman. If you invited a friend and they joined the game, they began as your vassal -- and you received a small portion of the resources they generated each day. This perfectly suited the core fantasy of the game, and provided a strong incentive to send invites; it also served a useful purpose in terms of gameplay, because you wanted to keep your vassals happy lest they leave you for another lord, and therefore had a strong incentive to help them out. Good social engineering as well as good design, in other words.
Stop asking "What would Zynga do?" We know what Zynga would do, and it hasn't worked out so well for them. Instead, start asking "How can I get my players to want to send messages? How can I make doing so a positive experience for them? How can I build messaging into the core of the game, make it natural, make it emerge spontaneously from play?"
League of Legends is currently one of the most successful F2P games on the market. As with most F2P games, it has a number of different monetization points, but among its most lucrative is the purchase of champions. It's a DOTA-style game, and champions are playable characters, each with its own advantages and disadvantages; playing each one is a different experience, they combine with others in interesting combined-arms ways, and there's always a desire to unlock more.
While the currency used to purchase them can be ground for, it would take an enormously long time to unlock all (and to unlock the "runes" used to further customize champions -- which are only available for soft currency, and the main sink for such). Consequently, there's a strong incentive to pay, at least at times, and many players do.
This is a perfectly honest, and reasonable system; I know what I am getting, and I can see the value. Moreover, if I'm content with playing just one or a handful of champions, I can reasonably get all the runes I want for them without paying; it's only if I become very attached to the game that I'm likely to pay, and if I am enjoying it that much, I presumably feel I'm getting value for the money. (I'm not claiming LoL's monetization practices are wholly ethical, by the way; sale of rune sheets strikes me as pretty skeevy; but selling champions is pretty reasonable.)
As another example, Plague, Inc. provides seven playable disease types (for free, in the Android version, or for a buck with iOS), unlockable through play. The impatient can pay to unlock them early; but in addition, there are three additional content packs you can buy, two unlocking diseases that work very differently from the original seven, and the third providing fourteen additional scenarios. Again, an honest transaction, and I have paid for all three packs, and gladly, since I love the game and actively like giving a little money to its developers.
Content, of course, requires development time and money; it is not as lucratively monetizable as consumables or gates. It probably can't be your only point of monetization; but it is something your players will feel far more willing to pay for than a one-time bypass for a timer. I guarantee you will see a higher percentage of paying players (though not necessarily higher revenues per paying player) if you provide content for pay.