The mistake people make now is to add these percentages up - probably because the idiom is to see what it “all adds up to”. This is wrong - now you need to multiply them together!
Let’s assume our game demo is 5MB and the very latest JRE is required to run it.
We get 0.001% * 100% * 20% * 25% * 75% * 20% * 20% * 20% * 75% * 75% * 99% * 10% = 0.00000001670625%.
With those 600,000,000 customers on the internet, that’s 10 sales of your game. Rock on! If this were Alien Flux, I’d have $200 in the bank. Don’t laugh - I’ve seen developers lamenting the fact that they’ve had 7 sales all year! Unfortunately A.F. will have taken approximately 1 man-year of development, and we’ve got expensive wives and girlfriends to maintain, plus a tendency to buy titanium exhausts for our motorbikes.
Now it doesn’t take a very bright spark to realizing where to put the effort in to improve sales. The closer something is to 100%, the harder it is to get it any better - there is a law of diminishing returns. So you start looking at the lowest percentages and figuring out ways to increase them. Some of them are much harder to increase than others.
One of the hardest to increase is also unfortunately the most significant - site visitors. There’s a multibillion dollar industry involved in increasing traffic to sites so I won’t delve into it here. It will tend to grow on its own as your site gets older and more mature, and you release more games.
What about the things you can manage straight off?
Well, looking at the next lowest figures, it seems like we could entirely eliminate a 20% factor by not requiring a JRE download. It may not have dawned on you just yet what this actually means but I’ll spell it out: this will result in a five-fold increase in sales. Instead of making just $200 from Alien Flux, I could potentially make - hold your breath now - $1,000. It has to be said that the Jet version is Win32 only - but that simply means that my Mac and Linux customers will have to download a JRE and use the jarred version, and as they represent less than 1% of the PC gaming market, I can live with it 
The next thing I’d want to look at is customer feedback about the demo, and making alterations to the game to make it more what the customer wants. Be careful here! Vociferous customers are like cuckoos - the loudest biggest bird in the nest tends to get the worm at the expense of all the real nestlings. An example is the complaints I’ve had about difficulty playing A.F., and the control should optionally be via joystick. It took a very short length of time for me to surmise the difficulty was indeed far too hard, and I’ve toned it down to the point where it’s basically trivial on levels 1 and 2. The control method, however, is what makes A.F. the game it is - it’s unique, simple, elegant, and powerful - and once you’ve mastered it, it’s ideal for the game. So that’s staying. There’s a host of other little suggestions I’ve put in, and tweaks here and there. Maybe I’ve doubled that 20%? If I’d stuck to the 1.4 JRE distribution method that’d be 20 sales instead of 10 - nice! But because I’m doing both those things I’m going to be looking at a jump from 100 sales to 200 sales - that’s another $2,000.
And so on.
Nearly all the other changes you can make to your game are business related and don’t involve programming skills. Your business skills are considerably more important than your programming skills because there are considerably more business factors involved. It really puts your programming time in perspective and makes you realise the fundamental truth when the retailers claim that they spend twice as much on marketing as they do on the development of a game.
But remember - if any factor ever gets to 0%, you’ll have 0 sales. You can’t slack on any of them.
When we’re not talking business - well, it’s just toys and idle time, so who ca]
about the size of the JRE?
Cas
)