Daily Elysium: Developing Demo(lition)

People seem to get their short hairs all twisted in a knot when their favorite game developers donÂ't follow what they see as standard practice, and Legend Entertainment is taking a few hits out there for their decision not to release a demo for Unreal 2.  Granted, a game demo can give consumers a better measuring stick by which to weigh their purchasing decisions, but IÂ'm not quite so willing to ascribe the certainty of nefarious motivations to Legend Entertainment.  After all, IÂ've seen more demos destroy a game before itÂ's ever out of the gate, than build a fan base. The problem is not, however, in the idea of releasing demos in the first place, but the approach many developers seem to take in producing their demo.  So, while IÂ'd love to see an Unreal 2 demo as much as the next guy, IÂ'd like to take a moment and suggest that maybe, just maybe, not releasing a demo is the better idea from a marketing and resources standpoint.

Read More to find out why.

LetÂ's begin with the recent example of Command and Conquer: Generals as our base of discussion.  For those of you who may not have been keeping track, an unoptimized multiplayer demo of Generals was recently released at Fileplanet for their subscribers.  The first thing youÂ'd have noticed in seeking this Â"˜demoÂ' out would be its jaw-dropping, system requirements: 2 Ghz CPU, 512 MB RAM, GeForce 4, etc.  Now, of the 16,000 + people whoÂ've since downloaded this clock-cycle gorged behemoth, what percentage do you suppose actually meet those requirement.  Go ahead and pick a number between, oh, one and fifteen, add a % sign to it, and subtract about three for good measure.  

Now, of that majority who downloaded without meeting the system requirements, what percentage do you think will be pleased with their perhaps jittery Generals experience when itÂ's all said and done?  Sure, Fileplanet kindly reminded us that this was unoptimized code, that the final release would not have such steep requirements, that many unresolved issues remained in this sneak peek, but the majority of consumers have no more memory for such qualification than a child with severe ADD trying to recall the quadratic equation.  What people remember is that they downloaded Generals, it ran like a heavy smoker at a 5k fun run, and their enthusiasm for the product is lessened as a result.  

So, these potential customers now have several weeks to muse on their experience before the actual optimized code can find its way into the fragmented memory locations of their waiting hard drive.  When these potential customers see that box on store shelves theyÂ're likely to recall the issues they had with the Â"˜sneak peekÂ', and be hesitant at the all important point-of-sale.

Worse still, this sneak peek needed not one but two dreaded demo patches.  While some might see this as evidence of the kind of dedicated support we can expect for the final product, I bet, considering that most gamers seem to sweat cynicism, more will judge this as evidence of shoddy coding.  A few will recall the initial warnings, that this was not evidence of the final code, but most will simply look at the shiny new C&C box on the store shelves, perhaps recall an experience – likely unpleasant – with Tiberian Sun, remember the TWO patches for the sneak peek, and then quite possibly put the box back on the shelf.

Now ask yourself this: has Command and Conquer: Generals been well served by releasing an incomplete sneak peek?  Would they be in worse shape had they just not released a demo at all?  

The problem here, of course, is not that they released a demo, but that they released a poor demo, and certainly they arenÂ't the first to do so.  The vast majority of demos I seem to run across are not particularly engaging, or present the game so badly that I hesitate to purchase the final product.  Even the much loved Battlefield 1942 demo, a demo so widely successful that it took a product with all the market presence of Emergency 2 and turned it into one of the most anticipated games of last year, still showed issues with graphical and internet latency optimization.  On the other hand, who remembers the BF1942 single player demo?  Better question: who remembers it fondly?  What saved Battlefield 1942, though, is that Dice ultimately produced a demo that was a reasonable approximation of the final product, that showed off a full range of engaging features, and proved to be fun to play in its own right.  This indicates that Dice took some time in carefully developing a demo that made people want/need to buy the final product, and they took this time because they needed that market presence.

Unreal 2 is a different beast, however.  It already has market presence.  It is a part of a successful franchise that has managed to build consumer interest through other media.  It already has fan sites, pre-reserves, name brand recognition, and consumer enthusiasm.  So, the likelihood for Legend is that any demo they produce stands a greater chance to hurt sales rather than help.  Unless they could produce an experience that genuinely increased customer anticipation, and that seems to be much more difficult with a single player only product, the best decision for Unreal 2 might very well have been to not release a demo, and let the final game stand or fall on its own merits.  After all, IÂ'm certain, as is Legend, that should Unreal 2 prove to be an outstanding game with positive reviews, most of those boycotting LegendÂ's decision not to release a demo will probably relent, and those who do not probably werenÂ't as likely to purchase anyway.

Ultimately, I find myself sympathetic with Legend's decision to shy away from a demo.  Certainly IÂ'd have liked to get my thick man-hands on the game before making my final purchasing decision, but if they could not or would not dedicate the time necessary to produce a quality demo, then I think itÂ's better to just avoid the whole mess.  A slapped together demo almost always feels just like that, and in the end that time might well have been better spent just fixing and optimizing the core gameplay code in the first place.  I think a lot of people have the impression that creating a demo is as simple as cut and paste, maybe kick out a few global variables, change a smattering of ones to zeros and then upload.  IÂ'm willing to bet itÂ's a significantly more involved process that can easily siphon away production hours better spent in getting the product out the door.  

In the final analysis, if Unreal 2 is a buggy mess, then we know why a demo never materialized, but IÂ'm reserving that judgment for the final code.  I accept the possibility that Legend made a reasonable decision based on the best marketing strategy for their product, and consideration of where their limited resources were best focused.  Maybe itÂ's a surprising moment of unabashed optimism on my part, perhaps just another glimpse at my inherent naivete, but I find myself unwilling to hold this decision against Unreal 2 or its developers.

With all that said, Legend has not left us with nothing.  TheyÂ've released a ten minute, 220 MB trailer of Unreal 2.  Again, this is only for Fileplanet subscribers.

- Elysium


Are you saying it would have been better if Westwood did not release a C&C:G demo, and thus not warned people that it needed a fast computer to play correctly?

Anyway, the article is entertaining, but I am still not convinced. I care about their marketing and resources about as much as I care for the snot on my dog. No, wait, snot is an important indicator of his health.

As much as I love Unreal, no demo still = no sale.

Mmm I don't think I really agree with what seems to be your conclusions. People should be able to down load and try a game on their respective systems, everyone has a different thresh-hold for performance vs. eye-candy. If the dev gets a lot of negative feedback about something like system requirements then it goes a long way to help them realize they might not be hitting that target market or perhaps they really need to buckle down and optimize that code. I do think demos should be released on the final code base, but that is sadly a rare thing, almost as rare as a full game actually improving between beta demo and release :). Personally, I have only very rarely bought a game that did not have a demo and ended up happy. I can count the acceptations on one hand, and most of those I was in the beta of anyway.

Pfft.  I preordered it anyway.  My computer can handle it.

Personally I agree with everything you said, especially about pre-release demos. I really do think demo's actually tend to decrease sales of a game. For one thing I tend to think most people are similar to me in that they don't finish a lot of the games they buy. So basically if I play all the way through a demo, I am likely to be burnt out on the game by the time it is released. Further a lot of demo's just plop you into the game with no instruction, and maybe it is just me but I am too impatient these days to read instructions. A good example of this is the Splinter Cell Demo, I didn't really get into the PC demo at all but on XBox when they actually had a training mission the game was awesome.