To make any real money in free to play, that is to say more than a "shit salary" as you put it, you need critical mass of players engaging with your game and having a high retention rate with your game.

I am by no means an expert on F2P, as I work for a AAA developer, but there seem to be two successful models.

1) Pay to win in a non-competitive game. Pay to win allows people desiring to spend the money to do so to progress faster. Faster progression in a competitive game gives no incentive to people who don't want to spend to win so in general you will lose these players and you won't have a critical mass to support your paying customers, which results in your game being abandoned. In non-competitive games this doesn't necessarily disincentivize non-paying customers, as they can still progress naturally at a more deliberate rate. This type of game tends to be supported by people called "Whales" (it's a gambling term for big spenders). Where a small percentage of the population supports the game financially.

2) Pay for aesthetic/customization. Another model doesn't stall progress at all, but rather users can pay to make their characters look unique. Special skins, cool effects, items to decorate your house, etc. This type of game doesn't turn off non-paying users as much, and can turn non-paying users into paying users more easily. Some ideas for increased monitization could be to have limited time/quantity items for sale, which might convince people to spend money if something cool is a limited edition skin. You could also have a resell option and take a cut off of the sale of those items.

In the end, neither of these models will work if you don't have a game that will both attract and retain a critical mass of players for an extended period of time. This comes down generally to having a great game that is continuously fun to play, and a LOT of marketing spend. At the end of the day, research other successful F2P games out there and choose the model that appeals most to you.

I've worked only in mobile f2p since joining the game industry. Here are a few tips:

1) Limit players daily gameplay through artificial means (stamina, energy, etc)

a) If a player wants to play more a day, they pay more

2) Have monthly, weekly, daily, and even hourly events

a) You want players to develop appointment style gameplay where the game becomes a habbit and they check the game multiple times a day

3) Use exponential curves for difficulty

4) When it comes to items/monsters/classes, never let the player buy what they want, let them buy a chance at getting what they want (i.e. gotcha in puzzle and dragons)

5) At minimal, have some social mechanic

6) Make sure everything is data driven

a) Track the metrics for everything

7) You will lose 50-90% of players within the first 2 minutes, so make sure your NUF (new user flow) is amazing

8) Your art (style and theme) and game icon will be a large component of drawing in players, much more than interesting gameplay

9) If they game doesn't work, drop it and start a new one. The games as a service phenomenon means don't try and find the silver bullet, build a new gun

10) Correct economy sinks are CRUCIAL, if your game takes off you don't want to have numbers growing to an unmanagable size

Hope this helps

http://www.gamasutra.com/view/news/200157/Freetoplay_tips_from_Triple_Towns_studio_head.php

Here's a bit of advice from my 4 years developing F2P games:

1) Don't scam your players. Do the introspection; how do you feel when the game is trying to rip you off?

Understand that your game may be funded by only 10% of the players, and 90% of them will play solely for free (its quite possible, even perhaps, desirable!).

If the players that are playing for free have a great time, they'll contribute by spreading the word, and getting you valuable players from the 10% you need to stay afloat.

2) Provide meaningful, but not game-breaking advantages to paying players:

Though customization items only represent roughly 2-4% of the entire purchases in F2P games, they can be considered valuable in that they don't affect the game balance. If the game is MMO, it gives the paying players bragging rights to look "involved" in the game.

If these customization items are also available to free players, albeit through grinding, it could help establish a social order within your game. Some games have managed to finance themselves solely on this model.

3) "Side-Grades"

As per the above, you can't include balance-breaking items for paying players, but you can give them versatility.

The Planetside 2 model is especially great here as each gun is more or less the same (they are fairly balanced) but players will feel more comfortable playing with guns with faster/slower rates of fire, better/worse accuracy and damage, etc. Let them choose their playstyle, but attach a price to it. Make them understand they can kill everyone with the basic gear, but that choosing their favorite gun will allow them to match THEIR playstyle.

If you game allows specialist-type guns by default (say, sniper, shotgun and smg), make sure you put a versatile gun for sale (auto-rifle).

4) Launch and watch

It's possible that your model will be flawed day one (technically, all game's model is flawed upon release). See what people ask form, consider it, and apply as needed.

If players ask for a feature in sufficient numbers and there's a way to make it balanced, do it and sell it, it will sell. Though its often true that players don't know what they want until they see it, once they've seen enough of your game, they can have very insightful feedbacks to provide. YOUR game becomes THEIR game upon launch.

This is critical.

5) Audience

I'm part of the school of people that believe you're not trying to create fanatics, but rather, want to convert fanatic players to your game. The best way to do so is make it appealing to them.

Define the playerbase you're trying to get to, see what they like.

If you're making a City-building game with elements of combat, make an Evony account. Log into that game. Join a guild. Play for 6 weeks.

Notice who spends money, and what they do it for. Look at the chat what their reasoning is for spending this amount of money.

Learn and watch.

Once you feel confident you understand these players, build your game with these people in mind. Remember what matters to them, and choose accordingly.

You can't build a successful F2P game without having an intricate (psychological) understanding of your "whales" (expression that defines roughly a user group that is a minority in your game, yet constitutes the majority of your sales).

When you consider item #4 above, remember that your whales carry more weight. If a decision would not be unanimous, check whether the whales will be happy about it, and choose accordingly. While it's important to keep the free players happy, your whales are more important.

For anything more specific, you'd need to show us a concept of what you're trying to achieve.

One thing to consider is, making your game cheaply, hence you won't need to spend a lot to keep the game running.

For example, a game i play used to only use images, not moving visuals, and added the option for players to download an imagepack and directly load images from there instead of from the server, which was faster for the players and decreased server-usage.
(this is an example, server-cost are getting lower and lower, and thus the game i talked about is allowing high-res images to be taken from the server during play nowadays)

Indeed, if there's one thing you want to optimize is server usage. This is the single reccurring expense you want to cut on, so using as little server power as you possibly can is key to making a profitable F2P game.