1) Put obsessive attention into making the first five minutes of use of your product awesome. It is the five minutes people are most likely to see. It is also a gate through which all subsequent use of your product must pass. (Seriously, folks, if you value conversion optimization, I want you to play five minutes of WoW and five minutes of any other MMORPG, and take notes on what happens. WoW will drag you around by the nose. WoW will point you in the direction of the next thing to do. WoW will show you success in those five minutes, even if you're terrible at what you're doing. We should all aspire to making products that nail those first five minutes as much as WoW does.)
2) Do not devote disproportionate developer resources to content/systems which the majority of the user base will not interact with. (It took them years to learn this, but they eventually go around to it.)
3) Make it simple. Then, make it simpler. WoW is one of the most complicated pieces of engineering the world has ever known, but six year olds can pick up and play it.
4) People will put up with any amount of drudgery if you give them small, frequent, random rewards for doing so. This is in fact so powerful I'm kind of scared of doing it.
5) If people are bringing their boyfriends into your product so that they can spend more time together, that is probably a good sign you have reached mass market success.
6) Monthly. Billing. Minor discounts for cashflow in advance.
7) High quality visual design can be made rot-proof by using iconic representations, bright colors, and timeless aesthetics. If you chase the leading edge of graphical sophistication, on the other hand, three years from now plan on doing a total graphical rip-out or you'll look dated.
As a game developer, I'd argue that thinking along the lines of #2 is very dangerous. Great games have a ton of subtle touches and content that only a small percentage of players will appreciate. The trick is that each player will notice a completely different set of features/content, and it makes them really happy when it feels like the developers went out of their way to polish some little bit of the game.
Think of Super Mario Brothers, one of the best selling games of all time. There are a stunning amount of hidden areas, tricks etc (even using an emulator with save+load, and a guide, it's not easy to explore the entire game world), but this is very closely related to how stunning the game world felt (especially for the time). An average player may only come across 10-20 secrets in the game, out of hundreds, but each time they'll have that warm fuzzy feeling where they are thinking, wow, they went to all the trouble to put this here?
Another good example is No One Lives Forever. They did a TON of funny lines with the guards talking to each other and notes that were left around (eg a note to employees saying no fornicating in the evil death ray storage facility). The chance of picking up on each of those was low, say around 10%, but since there were hundreds the game world, each player will have 20-30 moments where they were exploring some back room and they found something cool, which gave the game world immense amounts of character. If they instead made 20-30 moments and forced the players through them linearly, it would have felt cheap.
I'd also like to mention that, unlike many forms of software, it IS actually a requirement that game developers have fun making the games they play, because otherwise making the game itself fun is pretty much impossible. I don't think it's a coincidence that the developers of Diablo 2 took the time to put in the cow level, which most people never see, and spent a huge amount of resources tweaking leveling up for characters all the way up to level 90+, at the same time that they made such a fantastic game for the average player. If they made the game specifically for the average player, I think a lot of that magic that goes into development would have been lost, and paradoxically the game wouldn't have been as fun for anyone.
Of course you have to be somewhat rational about where you spend your development resources, but if you try to make sure that the majority of the players see the majority of the content, it's so easy to make a sterile game. And if you are often completely irrational about where you spend your time in the name of being artistic, it's surprisingly easy to make a game that is actually fun and gets attention. The difference in sales between the former and the latter, without a marketing plan B, will be orders of magnitude.
The issue with WoW in the original WoW and The Burning Crusade wasn't that only a small fraction of the people saw the content, it was that only a small fraction of the people could see the content. If I wanted to even set foot in the final dungeon, I'd need 24 other players with me, and all 25 of us would have had to have beaten all of the previous dungeons many times (and they're limited to being done once a week) and done a pile of quests to even get into those dungeons. So I'd literally have to have been max-level for about 4-6 months and playing 20+ hours/week during that time before I could attempt the final dungeon, and then the bosses were so hard that only a few hundred teams could beat them.
The solution Blizzard implemented in Wrath of the Lich King (the newest expansion) is to scale the big dungeons (raids) so that 10-man teams and 25-man teams can do them, and also have easy modes and hard modes, so that Ensidia (the best WoW guild) can be challenged by the Hard Mode and the more casual players can struggle against (but ultimately beat) the normal mode. They also kept adding new ways to get gear, so that people trying to break into raiding on the top level wouldn't have to spend months in the earlier raids, and they were pretty successful at that.
You're absolutely right that a more fully developed world with content that not every player would see is very important, but that content's only valuable if it's accessible to all the players. Anyone can go to the Cow Level in Diablo II, which is what made it great. But not everyone could kill Illidan or Kiljaedan in WoW, which was why it was a waste of money.
WoW was spending literally tens of millions of dollars on content seen by a fraction of a percent of its player base. That was unjustifiably bad. That it continued for quite some time was a problem of the dev team culture and also, probably, a symptom that they did not have good metrics. (I have been there, done that, and got the T-shirt. Not to the tune of tens of millions but I've certainly frittered away man months on things seen by less than .1% of customers and, even worse, 0% of trial users.)
It does sound like they took it too far (I've not played WoW after losing a significant percentage of my time as a teenager to Diablo 2).
But I was talking about the general rule. I'd say that the opposite mistake is much more common amongst game developers, and it's a big reason why so many crappy games are released.
Certainly 0.1% of customers is too low. But I wouldn't call 1% too low - many of the best games ever made include lots of 1% content. And I think that's no coincidence - the mindset that results in developers putting 1% content into a game also results in a fun core game. 1% content only happens when developers care a lot about the game.
I should clarify that by 1% content I mean a different 1% of the audience will see each piece of content.
Yet interestingly enough only 30% of people make it past level 10 and keep on playing the game according to Blizzard CEO Mike Morhaime. (http://www.escapistmagazine.com/news/view/98245-Only-30-Perc...). Also related to this is the issue that WoW seems to have peaked in its growth.
So even Blizzard has a lot to learn in this regard.
1) Put obsessive attention into making the first five minutes of use of your product awesome. It is the five minutes people are most likely to see. It is also a gate through which all subsequent use of your product must pass. (Seriously, folks, if you value conversion optimization, I want you to play five minutes of WoW and five minutes of any other MMORPG, and take notes on what happens. WoW will drag you around by the nose. WoW will point you in the direction of the next thing to do. WoW will show you success in those five minutes, even if you're terrible at what you're doing. We should all aspire to making products that nail those first five minutes as much as WoW does.)
2) Do not devote disproportionate developer resources to content/systems which the majority of the user base will not interact with. (It took them years to learn this, but they eventually go around to it.)
3) Make it simple. Then, make it simpler. WoW is one of the most complicated pieces of engineering the world has ever known, but six year olds can pick up and play it.
4) People will put up with any amount of drudgery if you give them small, frequent, random rewards for doing so. This is in fact so powerful I'm kind of scared of doing it.
5) If people are bringing their boyfriends into your product so that they can spend more time together, that is probably a good sign you have reached mass market success.
6) Monthly. Billing. Minor discounts for cashflow in advance.
7) High quality visual design can be made rot-proof by using iconic representations, bright colors, and timeless aesthetics. If you chase the leading edge of graphical sophistication, on the other hand, three years from now plan on doing a total graphical rip-out or you'll look dated.