This goes back to your conversion rate. It's easy to forget, after working on your labor of love for a year or so, that the average user will probably miss all of the things you think are obvious. For instance, if the only method of spending cash in your app is through a "store" button in your menu (and not implanted somewhere in the main play cycle) then the truth is, many players will never see it because they're going straight to "Play."
There is a reason that games like Triple Town integrate virtual item purchases right into their tutorial -- you want everyone to know it's there, and to know how (and why) they would want to use it.
It also goes without saying that the companies that succeed at F2P almost universally use analytics (such as Google Analytics or Flurry) to determine exactly how many of their users are landing in the store -- and they're tuning this flow continually. Something like the placement and color of your buttons in the menu can have a surprisingly large effect on user behavior.
I understand that these things feel so minor when you're focused on the gameplay. And don't interpret what I'm saying to mean that you shouldn't focus on gameplay (see my plea here.) But if you're making a F2P game, you need to care about how well you're guiding users into purchases, if you hope to be financially successful.
Again, this goes back to lifetime value and conversion rate. Most paying users aren't going to give you money the first time they play your app, or even the second time. And remember that of those users who do invest -- the most valuable are those who come back and make repeat payments.
If your app has a fixed amount of content -- for example, a puzzle game with 30 levels -- then you're essentially cutting off your most valuable users, and your LTV will suffer. There are, of course, exceptions. But take a look at any F2P top-grossing chart, and you're going to find mostly games with evergreen mechanics, and the top IAPs will mostly be virtual currency to fuel this play over a long period of time.
This one should go without saying, and probably should have come first in my list, since it's the most important. In the "bad old days" of retail packaged goods, you could get away with a mediocre product to a certain degree, as long as the box was pretty. And in the paid app marketplace, this is still true (but to a much lesser extent, since users tend to be more savvy and the user reviews are right there to read).
But mediocre free-to-play games do not succeed financially. Oh sure, there are exceptions -- and since quality is so subjective, it's a hard thing to prove. But you absolutely need to create a game that has a set of users who love it. If few people love it, few will talk about it, and even fewer will stick around to pay. One of the cool things about free-to-play is that you can make a living with a niche app -- but it still needs that burning core, and that burning core needs to stay engaged for a long time. This is your first priority, and is also the hardest thing to get right!
It's important to realize that there are a lot more opportunities to fail with a F2P app than there are with a paid app. With a paid app, you've earned the revenue at the moment of download, even if the user never launches it. With F2P, Not only do you need to grab users with your app's concept and presentation in the app store, but you need to delight them before they spend a cent. And then you need to give them a great reason to spend money, and to guide them through this process, and then get them to repeat it. If anything in this process breaks down, it's going to kill your revenue potential.
It's not enough to test your gameplay -- you need to test the whole loop. Companies that succeed at F2P test extensively before launch, and then treat the release as the beginning of a product's life, using analytics and business experience to improve their core metrics until the product has either proven viability, or is abandoned.
With Monkey Drum, we tried here, but honestly could have done better. We learned fairly on that most users weren't finding the store, and were able to improve this significantly through some UI changes. Unfortunately, this didn't seem to be the biggest factor in Monkey Drum's failure to monetize, which leads to the next point...
When we chose to make Monkey Drum free-to-play, we had this idea that the very young users would probably enjoy the free-play drum mode and interacting with the characters, and a much wider audience would get into the sequencer and music creation tools. We imagined a community of users materializing around the song creation tools, and kids and adults alike wanting to purchase additional instruments to create more expressive creations.
Only after we shipped, however, did we realize that our biggest audience was very young children -- and that their favorite activity was playing the drums, which didn't have much incentive to pay.
As I mentioned earlier, we went out of our way to not be "evil." We clearly mentioned the in-app purchases in the app description and when the app first launched. The hard truth is, educational music apps have a much smaller audience than games. And educational music apps with a specific aesthetic that appeals to young children have an even more limited audience -- and one that we suspect doesn't drive many IAPs in any situation.
Despite being an almost universally loved app, and going out of our way to guide users into our store, the downloads and LTV portions of the revenue equation just didn't make sense with this audience. We believe now that it was the wrong business model altogether.