Unless you know the cost of porting + manufacturing costs vs. the royalties deal vs. projected sales estimate, you could be better off not porting until these factor weigh heavily into your favor.
So while you'd be increasing your potential customer pool by 4million (using your example) vs. 12million, the amount of resources spent to reach that 4million and capture sales from that pool, might not be worth. Of course this would vary game to game and studio to studio but you simply cannot use blanket statements when discussing such low volumes.
I disagree. The cost per disc in terms of printing and licensing is the same, roughly speaking, for the different platforms, unless a deal was cut - but that's obviously outside the scope of the argument. For the orignal XB360, there's the cost of the engine, the assets, the marketting, the printing, the licensing, and the distribution. Printing, licensing and distribution is a per disk value. It's, say, $15 per disc. Marketting, engine and assets are the development costs, which runs into millions, with development of the engine being the smaller part of that. To port to PS3, you'd need to create the engine on new hardware, which should cost similar to the original. Perhaps it'd cost more because of trickier tools to work with, and perhaps it'd cost less because you've got the engine mechanics worked out and only need translate them to different methods. The per disc costs will be the same, you save a fortune on assets, and the marketting depends on whether you release in a similar time-frame or not.
If you spend $10 million developing an XB360 game, and sell 300,000 copies to 12 million users, the returns are, say $50 per game, $15 million, and costs are $10 million + $15 per disc = $4.5 million = $500,000 profit.*
If porting the engine to PS3 costs $1 million, and you expand your market by 33%, That's a total cost of $11 million to 400,000 sales (assuming uniform sales) to 16 million gamers. The returns are, $50 per game, $20 million, with costs of $11 million + $15 per disc = $6 million = $3 million profit.
Yes, these are 'out of thin air' figures, though they're at least in the realms of what happens. And they're only for demonstration purposes. You can try out different figures to see how the earnings vary, and if you do this, you'll see that the only time the port doesn't make more money than not having a port is when the price to port is a higher percentage of development costs then the market size it adds. If the cost to port a title to PS3 is 50% of development cost, and the increase in market is 33%, it's not profitable.
Now the cost of porting a title
is low relative to the 20+% extra market size that a second platform introduces. You're talking some 8-10 devs for 6 months or so. ERP has commented elsewhere that a port can be made in 3 months, IIRC. You lose all the iterative development costs, have a finished design, have your finished data structures and full assets for testing. Obviously difference in architecture add complication, but not to incredible costs. Pretty much the only time it isn't economically sensible to port, barring incentives from the console company, is if the second console is expected to buy differently and you'd expect sales lower than the percentage market increase.
And this is why there's lots of cross-platform games coming out this gen! Factor in also that 3rd party developers want cross-platform engines, and the investment in the initial port carries through to
multiple other titles. Imagine Games X1, X2 and X3 come out exclusively on XB360 to an install base of 12, 20 and 30 million units respectively. Port the engine with X1 and with a 33% install base, PS3 fetches you 16, 26 and 40 million units unstalled base. That's 33% extra, for a relatively small initial outlay. That's why cross-platform is popular! That's why UE3 is big business! That's why devs are creating multiplatform engines! There isn't a massive economic barrier discouraging multiplatform titles. You don't need a console base to reach a large critical mass before you target it with your games, as long as it remains a worthwhile
percentage of the leader. The only exception, on economic grounds, for targetting the larger cross-platform userbase, is if you expect one platform to not buy. For example BGDA on PS2 was ported to XB and GC. It sold poorly on GC (poor port by accounts), so the sequel wasn't released to it. If it sold as well on that platform as the others, it would have received a port because GC offered an extra 20% or so returns on the investment. Sometimes a console appeals to a certain type of gamer, and the alternative platform lacks users interested in that genre. Porting 'Violent FPS' to 'Pink Fluffy Cutesy-Platformer-Favourite Console' doesn't make sense as it's a different userbase with different tastes. That's not likely to be an issue with XB360 and PS3 though, with perhaps the exception being regional cultural variations. You may well not port Pachinko to XB360!
One important factor to consider is the number of multiplatform users at this stage also. Like myself and many others, the people who bought a PS3 also own a 360, as we're still in the enthusiast state of purchase with people wanting the latest and greatest. The above is all assuming that the Port itself is comparable on both systems, something the PS3 has suffered from in these early stages.
Though that's a factor, I doubt it's high. Otherwise why are developers creating multiplatform games if the PS3 userbase is the same as the XB360 userbase? Why create Madden on PS3 if 80% (P.I.T.S. figure!) have an XB360? The PS3 userbase has to be considered a large enough proportion of the XB360 user base or it wouldn't be economical to port, with the exception of an investment for future games on the same engine...at which point there's no reservation to porting either way!
* And yes, I know licensing isn't paid per game sold. It'll still remain proportional to the platforms though, and can be ignored to keep the explanation simple, despite it already being long-winded!