Optimal Game Demos

A game demo has a very specific purpose. From the point of view of the developer, it is to encourage players to purchase the game. From the point of view of an honest player, it is to help evaluate the game to see if it is worth purchasing. These distinct purposes should have a substantial common ground that offers the basis for a concept for an optimal videogame demo. But what is the nature of an optimal demo?

I believe too many developers misunderstand the role the game demo has in the life-cycle of their game, and perhaps some have even been put off by the not-too-surprising news that when you take demos as a whole they don't demonstrate an increase in sales for titles with demo versions, but rather a decrease. I want to advance the equally not-to-surprising perspective that the poor performance of game demos reflects the poor design of most game demos for their purposes. A good game demo will generate sales. Furthermore, a good game demo can generate sales that were otherwise impossible. But a bad demo will surely put players off your game, and in this respect, no-one should be surprised.

'Good' and 'bad' in the context I'm using here refers to the goals already picked out; a good demo is one that is closer to optimal, a bad demo is one that is far from optimal, in terms of both the developer's and the player's purposes. An example of a good demo from my own life would be Sega Bass Fishing for the Dreamcast (pictured above). There is absolutely no way I would ever have considered purchasing this game. I wouldn't even have picked it up in the shop. But it came on a demo disk for the Dreamcast, and my wife and I played it. It was a little silly, and kind of fun, but we didn't really know what we were doing. It intrigued us, though, so we played it again. And after we'd played it a few times we began to get the hang of what we were expected to do, and it started to become a lot of fun. Eventually, we came to the conclusion not only to buy the game, but also the custom fishing controller that went with it – and proceeded to rack up a great many hours of enjoyment with a game that without a demo we would never have discovered.

I play a lot of the demos that come down the pipe on the PSN service, and most are rather poor. Many are poor because the game itself is ill-considered – there's no hope of saving such a game with even an optimal demo. As the scatological expression goes, "you can't polish a turd". But many might be perfectly reasonable games, or at least reasonable for their price, but all too often the demo manages to annoy me, or exclude the possibility of getting into the play of the game, or otherwise act as a barrier between me and the game itself.

One sure way to achieve this kind of failure is a time limitation that is persistent e.g. you can play this demo for 10 minutes total, after which you cannot play it ever again. Such a demo is sure to be deleted as soon as the time limit is expired, and as such the chance of converting into sales is minimal. Time limitation itself is not problematic – the Just Cause 2 demo, for instance, offers a generous 30 minutes per-sitting slot of play which is more than adequate to get to grips with the game itself, and the demo has also been structured to encourage repeat play of the content. It serves to whet the appetite of the player, and encourages them to return to the demo multiple times. This, I suspect, is a principal factor in optimal game demo design – if the player returns to the demo multiple times, over any period of time, they are increasingly likely to purchase the full game.

But time limitation can also generate frustration in a manner not likely to lead to sales. Trash Panic, for instance, offered a time limitation of about a minute on a puzzle game which offers multiple levels. As a result, I was unable to really get into the play of the game because the only way to play the demo is to rush hell-for-leather through the content. Nothing was gained here. Since the final game contains multiple different levels, it should have been sufficient to let the player enjoy one of its levels at their own pace as often as they wish. The time limit on Trash Panic blocked the normal play of the game – it was no longer even a demo of the game, as such, in that the way one must play this demo is not the same as the way one would play the game proper. Compare Critter Crunch, which offered several levels of the game, with many more for players who purchased. I bought this game on the strength of the demo. (I had my issues with the full game, alas, but the demo was an accurate representation of the gameplay and the demo let me discover its play).

It seems inevitable that a demo must be limited in some way – it can't be the whole game, although there are titles that are close to exceptions. Tale of Tales The Graveyard is essentially complete in the demo form, and the 'full game' simply has one extra point of functionality. Some massively multiplayer games (e.g. RuneScape) use a similar kind of limitation – one which is cast as the paying player getting access to more of the content by paying, rather than drawing attention to specific limitations that "box out" the demo as a tiny proportion of the game proper. This "get more" approach has huge advantages since you're giving enough of the game to the player to get them hooked. And getting them hooked, I'm claiming, is precisely what an optimal game demo needs to do.

Some demos offer basically a tutorial and no gameplay... this is ill conceived. The developer needs to give the player enough of the game to allow the player to try out what's on offer. Besides, most successful games do not reinvent the wheel – if your game is so complex that you need to tutor gamers in its intricacies your whole project may have problems (with, I accept, some allowances for attempts to move in new directions, such as Skate). By all means include a tutorial, what you don't want is to exclude so much of the game that players can't really see what's on offer. Another problem in this regard is that a tutorial is one of the hardest things to design for any game, so putting your tutorial on prominent display is something of a risk. It's not so much of a problem for a game that simply modifies existing tropes (such as almost any FPS), but it's something to bear in mind.

The other extreme is the game which throws the player right into the thick of it. This seems to be the norm for all manner of shooting games – I did not finish the demo for (say) Dead Space or Kane and Lynch 2 because I didn't feel in any way adequately prepared for the challenge delivered in the gameplay, and I thus quickly lost interest. Now in this respect, this might not have been a flaw in the demo – because these games might not be games that I want to play anyway. It might be that these games require a more challenge-oriented player than I am, and that having the demo present that level of challenge was the right way to go. But I wonder if it wouldn't be better to offer the demo in two parts – one easy and one hard – such that there is a chance for the player to adapt to the game before they are thrown into the thick of it.

An interesting trend in recent demos is to end just as the boss appears. Presumably the idea is that since players (it is assumed) want to fight and beat the boss, stopping at this point whets their appetite. For me as an individual, I love these demos but I will never buy the game that they are promoting. I love them because as a general rule I hate bosses, and these demos don't make me fight one. I hate bosses largely because they are usually poorly designed – I did not hate Shadow of the Colossus, for instance – but a demo ending on the boss tells me "this is a game that makes bosses have a prominent role", and discourages my purchase. But of course, vast numbers of games I have bought and enjoyed have had bosses that I simply put up with. While I frequently enjoy the demos that use a boss-foreshadow to close, I'm not going to buy the game because the boss is what's being emphasised, and I don't trust developers to make bosses I will enjoy. (I imagine other players, especially challenge-oriented players, have a different feeling about this).

In terms of determining an optimal game demo, the requirements will presumably need to match the form. As I say, I am not much interested in bosses, but other players are. There is a question here as to whether the optimal game demo for a boss-including game should be trying to reach me as a player or not. After all, if its core non-boss gameplay appeals to me, I might be in its hypothetical audience... I might ultimately put up with its bosses, as I do in Zelda, Castlevania or Metroid. But the boss-loving player has different requirements, and it might not be possible to devise a demo that would appeal to both of us. This being so, perhaps the optimal game demo in such a case is the one that meets the needs of the boss-lover. Perhaps what is optimal for a game demo is to prefer the larger audience over the inclusive audience; some players are excluded from any given game, after all, and no game can be all things to all people.

In this brief look at the idea of an optimal game demo I have suggested some key principles:

• the limitations on the demo should not block the core gameplay from being experienced
• the demo should include a sufficient slice of actual gameplay to entice the player
• the player should want to play the demo multiple times

It may be that these principles don't apply to all game genres – a computer RPG demo, for instance, might not encourage multiple replays but might simply be an adequate demonstration of what is on offer. Nonetheless, it seems to me that there is a concept of an optimal game demo, and that there are principles that developers should be considering if they want their demo to do its job and get players interested in their game. At the moment, there are far too many demos that have the opposite effect, and poison the well for any player who might be considering purchasing the game in question. A conversation between players and developers may be absolutely necessary if we are going to solve the problems associated with game demos.

What do you think? What annoys you in game demos? What do you think an optimal game demo should consist in? Please share your perspective in the comments!

Comments

[Clayton Hughes]

I'm a bit biased here, but absolutely the most memorable demo I ever played was the Final Fantasy My Life as a Dark Lord (WiiWare) demo.

It offered, untimed, the first chapter (say about 4 levels). After each level, you unlock a new item (traps, monsters, etc.) to use and you gain some currency which you can use to upgrade these things.

I assumed it would end there as it seemed a good stopping point, but it let you keep playing. After I completed the first level of the second chapter, I was thinking to myself "wow, is it really going to keep going?" Then, halfway through the second or third level, the characters, who had been popping up before each level to teach you how to play, appeared to tell you about the demo ending in a very self-aware way: "Hey, I was just getting started.", "You haven't even met me yet", etc.

It was marvelously done. I was very impressed. I replayed it a second time, spending lots of time replaying missions to max out my upgrades with the currency. I was quite impressed with how the demo handled itself.

I have to admit that I didn't purchase it (Square Enix does well enough on WiiWare that I felt they didn't really need my money), but it's one of those games I'm tempted to buy every so often

[Chris]

Clayton: thanks for sharing this! I downloaded that demo but haven't got around to checking it out yet... your account encourages me to give it a go.

[William Willing]

One of the problems with demos might be that they are more likely to be a quickly put together afterthought than a well-designed sales pitch. Too many developers default to just the first few minutes (or levels) of gameplay for their demo. Sometimes that works, sometimes you might need to be a bit more creative.

There's a lack of solid data on demo conversion, in my opinion. I used to work for a games company that did a lot of A/B-tests on their web pages, but I couldn't convince them to do A/B-tests on different versions of game demos. I have a distinct feeling that time-limited demos often hurt more than they help, for reasons you've mentioned in your post, but without the data to back this up, it will always be conjecture.

In short, you should consider the demo as an important part of your marketing strategy and give it the attention it requires.

[Chris]

William: "In short, you should consider the demo as an important part of your marketing strategy and give it the attention it requires."

I completely agree! The problem often lies in the crunch at the back end of production; the team just don't have the time to do a good job, and there's a sense that the demo has to be out there before the game.

I thought it interesting that Dylan's Q-Games put the demos out for their PixelJunk games sometime after the release of the game. Those demos have been quite optimal in my estimations - I couldn't believe how generous the PixelJunk Eden demo was with its content. But then of course, as a download game there isn't the rush for promotion during the narrow sales window.

The retail system is actually quite hostile to most games because of the short time allotted for sales, and the demos suffer as a consequence of this compressive effect. With the cost of development being what it is, I don't see this problem solving itself.

As for your idea to make multiple versions of demos to test effectiveness - it's a fascinating idea, but I don't see many companies willing to pursue that degree of investment in what amounts to research. Perhaps some intrepid academic will explore this with a small game title at some point.

Thanks for commenting!

[Mory Buckman]

Have you ever seen the early teaser trailers that Pixar makes for their movies? They're not scenes from the movie, but scenes specifically designed as teasers which introduce you to the characters in memorable ways.

I understand if it doesn't fit into production schedules, but I wish game demos were more like that. I don't much like game demos as they're typically made, because it's always taken from the final game. But in the final game you feel like your efforts are building toward something. It's harder for me to justify the time investment of playing early levels (which are usually less interesting than the later ones) when my progress isn't saved or unlocking anything. So I think game demos ought to be separate from the game, self-contained little scenes that show you why the game is fun without making you replay levels: "Okay, now buy the game so you can do this all again, but for real this time!"

You know what, what I'm suggesting really isn't so unmanageable. A lot of games have prototypes. They make a simple level showcasing the gameplay, which demonstrates that it'll be fun, before moving on with the actual game. Why not polish that prototype up a little and release it as a demo? Granted, this needs to be considered right from the start of development, because whatever prototype is made needs to be complete enough to share. But the benefits seem considerable. I would think the demo would get a lot of word of mouth if it were a self-contained experience rather than a conventional demo. When people finish it, even if there's an ad for the game they'll still feel satisfied, which is a good association to make. It's also more likely to be played multiple times, if freed from the feeling that demos are a waste of time next to "the real thing".

So I think the optimal game demo is one which doesn't actually come from the game.

[Chris]

Mory: what your proposing is interesting, but it favours the player over the developer, and so might not be optimal in the sense I outline here (i.e. hitting the perfect balance between the two). It *could* be optimal - but only if creating the extra material significantly increased the conversion rate of the demo, which I am slightly doubtful of, and would need proving in some way.

Polishing prototypes probably isn't an answer, since usually prototypes are too rough for anything useful, and also tend to be based on material that already form part of the game.

To create new material adds to the development schedule, and thus to the cost of development - there would have to be some strong evidence of the benefit of this approach before it could be considered optimal.

Re: Pixar's teaser trailers - in the 1980s, this was a common practice in film promotion; making short films as promos. But now this has fallen by the wayside (barring Pixar) and instead we get a cross section of the movie pasted together with fast edit technique. I loathe this kind of trailer because it is often possible to derive the entire narrative of a film from the content in the trailer - spoiler rex, if you will.

It's probably the cheapest way to make a trailer, but is it an optimal trailer? I'm mindful of the fact that people such as myself who will view it as a spoiler are less prevalent than people who will take away solely a vague emotional impression of the content of the movie. This is probably why the trailers now contain all the "best bits" of the film.

Thanks for commenting!