Because my previous blog post was not a complete academic essay on the subject, nor indeed intended to really go any further than the few people that visit our blog, it seems that a few people are deconstructing the arguments and poking some big holes in the assertion that “The Demo is Dead”, which is fine, but the article is not at all complete, and contains no hard data (of which I have a lot). At the time, I just thought I’d pen some musings on the subject talking to people who already didn’t care (existing blog readers, who are generally customers and therefore unaffected by what we do with our existing titles).

Anyway, the internet sort of exploded in rage and disbelief that a tiny indie developer could become such a cruel, heartless, candy-snatching killjoy.

As a general reply to various comment all over the place, here are some further musings:

99 Reasons To Not Buy Your Game

This was clearly an exaggeration for literary impact, and if that’s not obvious to you, for shame. But instead of just asking me what those reasons are, maybe you could engage in devil’s advocacy, and think of some yourself. Here are some I thought of, spuriously:

1. I got my fill of gameplay already from the demo. (Our demos typically gave away 25% or so of the full game progression)
2. I’ve had 90% of the initial delight of the game for nothing. Paying some money for the remaining 10% is a waste of money. (Note disconnection between “delight” and actual content)
3. I can’t be bothered to pay for it when I can go and play another free demo somewhere else.
4. I’ve already got a bunch of games I’ve paid for but not yet even played. Maybe I’ll not bother getting this one yet.
5. I played the demo ages ago and forgot all about it by the time payday came because something else distracted me in between.
6. I only buy games through Steam.
7. I’m a poor student/waster/single mum and I don’t spend money on games especially when I can be entertained endlessly by demos for nothing.
8. I loved the game except for this one small thing that I didn’t like like I can’t remap the fire button to X and for that reason alone I’m not going to buy it.
9. I thought the game was too easy but that’s because the demo can only show the first 10 levels which have to be easy to not put off the 95% of people who find it too hard.

You’re Just Using Yourself As A Single Data Point!

Some have accused me of using myself as a single data point (“I’ve never bought a game in the last 5 years from playing a demo”) and drawing my conclusions based on this, which is fallacy. This is not the case; my own, singular experience was what got me to look at the data in the first place. It was just a hunch, that I got to thinking about actually a few years ago. It’s only in the last year or so that the data has become impossible to ignore (see below for some figures).

The Nature of Puppygames Demos

Few people were aware of the exact nature of our demos, or even our games, and it’s probably worth researching because our games are of a particular ilk and available only on a particular platform. We make desktop arcade games mostly, and that’s a pretty strange niche to begin with, which substantially effects the way demos work.

Our demos were “full” versions of the games, which could be unlocked by registration (no further download). They tended to let you play the first 25% or so of the game unfettered before expiring on a cliffhanger (eg. first boss appears, or you’re just about to see the next “world”, for example).

Claims that we’re “doing demos wrong” are from people who, I suspect, have not been doing this for as long as we have. The fact is, our demos were more or less no different from nearly every other demo I’ve ever seen. They weren’t even unsuccessful either – they converted at an industry-respectable rate, AFAIK. The problem is that rate is shit and the amount of money we can charge for a successful conversion has been eroded, which brings me to…

Context Is Everything

The context of pricing and market positioning, specifically. Over the last 10 years we’ve seen the average price of an indie game plummet from $20 (sold direct by the developers) to $5 (sold on Steam or BigFish in a sale) to about $1 (sold in a bundle of some sort). Steam pioneered the price slashing in the market – I’m sure you educated types with economics degrees have a special name for this manoeuvre. In the space of a couple of short years, direct sales plummeted to less than 1/10th of what they used to be (and they were never great). Almost overnight, the chances of being an actual indie developer – and succeeding! – have dropped from “you’ll be lucky” to “you’ve as much chance of winning the lottery”. Not only had consumer expectation of prices been eroded from $20 to $5, but consumers were also taught by Steam to buy on the basis of video and recommendation and, most importantly of all, discounts.

Then, just as things didn’t seem they could get more crazy, along comes the Humble Indie Bundle, and we’re now becoming accustomed to picking up titles for a dollar or less. Again, demo unseen. We’re conditioned to buying stuff because it is cheap not because we necessarily want it. I say “we” – yes! I am one of you. I am a consumer. I’ve got a hundred games in my Steam library. I am doing all these things. I won’t buy a game if it’s not on Steam any more. I won’t buy a game if it costs over $10. And so on. This reminds me of an anecote many years ago when a friend of mine came bouncing into the room full of glee because she’d bought some mint essence. When I enquired what was so amazing, she told me that it had been 75% off so she just had to buy it. I can’t recall her ever before or since actually making anything with mint essence in it, but it was a bargain!

In this context, what we now see is that 95% of our income – any developers income – comes not from conversions of demos, but from sales via gatekeepers and bundles. What the focus of my original article was really about is that there is a case for simply dropping prices through the floor and not giving anything away for free. There is “free” stuff everywhere, already. The differentiator we now have is that if you want to sample our stuff, it will actually cost you. Otherwise it is simply unavailable. It is out of reach. You can look through the glass into the shop but you can’t touch it until you spend a (paltry) amount of money. Just like with mostly everything else in the world these days.

Are We Right?

There’s no harm in being wrong. We can be wrong. We’re going on what the data tells us, and we have a lot of data. We’ve sold 481,529 games in the last 3 years, and 30,246 of those have been to people who played a demo. That means the other 451,283 sales were made without anyone ever seeing a demo. If you want percentages, that’s 6%. We’re quite happy to be proved wrong! If the data tells us we’re wrong, we’ll go back to using demos.

Our hypothesis is, we’ll make a bit more money if we ditch demos and drop the prices. As you can say what you like about the 97% of sales being without demos and argue till you’re blue in the face that you don’t buy games without playing a demo first, go on ahead. Argue away – you’re arguing that black is white. You’re not making us 97% of our sales. The bit you need to argue over is this:

6% of our sales are to demo players, direct, and they have made us $72,000. We think that if we drop our prices hugely, and ditch demos, that we’ll continue to make 6% of our sales direct, but that we’ll make a bit more than $72,000.

The Sands Shift Beneath Our Feet

And still that’s not the whole story. The thing that most beginning developers – us included – fail to take into account is how the markets change over time. As I said, when we first started, we sold conversions on demos for games that cost $20. We started just at the tail end of a golden era in independent game distribution (typical bad luck, huh). The internet had just revolutionalised developing games and the gatekeepers were just about to move in, along with a flood of other developers who suddenly discovered they could do it too. It is suprising in hindsight that so many developers clung to the $20 price model in the face of what was happening.

Things came to a head in about 2008 or so, when we released Droid Assault. Droid Assault was released to the sound of tumbleweed. No-one was even the least bit interested. It’s a great game (IMHO, haha), but when it was released, nobody wanted to buy it. Customers were already thoroughly in the pockets of Valve and BigFish by then. If you didn’t have a game on a portal, it simply didn’t sell. DA must have shifted literally a few hundred copies. By contrast on Steam, now it’s finally out on Steam that is, it’s shifted thousands of units.

And so we must realise that the market is changing, all the time, imperceptably slowly. Let’s look at those figures I just mentioned above, and instead, let’s look at just the last 12 months:

In the last 12 months we’ve sold 77,224 games, of which just 725 were demo conversions. The demos weren’t suddenly any different. The prices weren’t suddenly any different. Suddenly, after just 2 years, we’re only making less than 1% of our sales via demos. Nothing else changed except the entire rest of the market.

So actually what you really need to be arguing over is this:

1% of our sales are to demo players, and they have made us $5200 (yes, really). We think that if we drop our prices hugely, and ditch demos, that we’ll continue to make 1% of our sales direct, but that we’ll make a lot more than $5,200.

Put that in your pipe and smoke it.

I recently acquired a Raspberry Pi, the £25 computer which is the brainwave of various UK industry luminaries, most famously including David Braben, the author of 8-bit classic 3d-space-shooting trade game Elite. I have a few thoughts about how it’s delivered, what it costs, who it might be aimed at, and what might be done to improve things for level 2.

The Raspberry Pi arrives as a naked circuit board about the size of a box of what we in the UK call fags. This always amuses Americans, but Americans are of course easily amused, albeit for an extremely short duration. The circuit board has as many connectors on it as you could reasonably hope for in a computer – two USB ports, stereo jack, Ethernet port, micro-USB power input, HDMI out, and composite video out. There’s even something called a GPIO which is a bunch of pins which you can connect “things” to that do “stuff”, but that’s for very clever propeller head types who understand electrickery, and not programmers like me.

Of the USB, Ethernet and stereo jack, we shall have nothing much to say. Or indeed the GPIO thing. They are as they are and I dare say no improvements could be made on them. The other outputs represent some rather odd thinking.

HDMI Output

HDMI is the future! HDMI is awesome! VGA is dying! … but unfortunately, HDMI is only available on the very latest monitors and newer TVs. So we’ve got a £25 computer, and the theory is that you’re supposed to pick up and use some existing mouse, keyboard and monitor that you’ve got, probably gathering dust in the spare room, to save you a pile of cash. Except, of course, your monitor only has the usual D-shaped SVGA input. Everyone I know has a spare SVGA monitor lying around. Not a single one has a spare HDMI capable monitor. A few people quite likely have monitors with DVI-D input, but I have a feeling that like mine it is already in use as the main PC monitor. The solution in my case was to buy an HDMI cable and a brand new HDMI-capable monitor. £115 vanishes.

Then there’s this thing about HDMI not exactly being reliable, like SVGA is. In my case, all I got was a black screen on startup. It could have been any number of things but fortunately I’ve got quite a diagnostic mind and I’m not 12 years old, like many of the prospective owners of the Pi are purportedly meant to be. Suffice it to say it took me an entire evening of Googling around to discover that you had to create a text file with HDMI configuration parameters in it to get the monitor to actually display anything. This is fun… for various values of “fun” which unfortunately lie somewhat outside my comfortable parameters. You see, I spend all day, every day, trying to figure out why the fuck stuff doesn’t work properly. I was rather looking forward to plugging in the Pi and tinkering with it straight away, but I didn’t even get that far. I had to wank about fiddling with it just to get a fucking picture. Yes, that made me cross.

Composite Output

A nice touch but pointless. Almost all composite-input devices likely to be found around the home are now rightfully in landfill and those that aren’t are generally massive and not the sort of thing you generally want cluttering up a spare corner of a small room. Not only that but they tend to be Jurassic power-guzzling dinosaurs and prone to going on the blink. The other day I spent a happy afternoon coding in CRT distortion effects in the new version of Ultratron. So anyway: composite output – might as well have saved the money on providing this output, no-one in their right mind needs or wants it who is going to own this device. Yeah, even you. You’ve got a spare old SVGA monitor in the garage too haven’t you? Throw that old black-and-white portable telly in the skip.

Micro-USB Power Input

Er… why? I’ve got about 20 different adapters lying around my house, all with sturdy jacks, providing voltages anywhere between 3v and 12v. I literally have a box full. I expect mostly everyone else does too, because over the last 30 years, nearly every widget you’ve bought came with its own. I bet amongst them all you’ve got a 5v DC input you could have already used. Well, it doesn’t matter, you can buy the flimsy microUSB input instead for another £5.

Supplied without an SD Card or software

Now, here’s probably the most controversial thing I’m going to say about the Pi. As it comes, it won’t actually switch on or do anything, even if you’ve got the spare keyboard, mouse, and shelled out £85 for a monitor and cable. You also need to get yourself a 4GB SD card to put on some firmware and an OS. Although this process is trivial, and the cost is nothing to worry about (literally, the price of a bag of peanuts), it’s completely unlike my first experiences of computing.

My first real computer was a Vic-20, back when they first came out. It cost quite a lot of money – a lot more in fact than all this Pi stuff has come to in real terms – but: I plugged it in to the telly, switched it on, and tuned the TV using the twisty analogue knob. And there it was: a BASIC interpreter, 3583 bytes of RAM free (though 2 went missing immediately somewhere). It came like this out-of-the-box. I could get coding on it within seconds. It’s this plug-and-play appeal that turned so many people off of PCs for gaming and on to games consoles in the first place and represents exactly why software engineers are so derided by mostly every other engineering principal.

Built to a Price, But What Price And Why?

So the Raspberry Pi commands all sorts of cool headlines like “a computer for just £25!” mostly because it sounds, well, cool. Except I’ve spent £200 on mine in total, which coincidentally is about what my Vic-20 cost me albeit in 1982 money. That’s clearly beyond typical pocket money for today’s cash strapped youth who of course absolutely, completely need mobile credit and, er, whatever it is that teenagers spend money on. Drugs I expect, as they can’t easily get booze any more.

About £165 of that cost was the monitor, keyboard and mouse; the keyboard and mouse were really just me being slightly extravagant (I have an awesome tiny Cherry ultra-compact keyboard), but the monitor… well, that was annoying, really, as I could have used one of several other devices kicking about the house if the Pi ditched its HDMI and composite outputs in favour of something more befitting its status as a cheap toy that is meant to be attached to stuff you have lying around gathering dust.

So with the Pi only about 12% of the total cost of the whole setup, why exactly is it designed like this? Why is it built to meet a £25 price point when a few bob here and there would barely change the overall cost but vastly improve the whole experience? Who buying one of these things actually gives a fuck that it’s £25 and not, say, £40? I really have no idea.

It would appear that there have been engineers involved in the market research. Oh dear.

No-one buying computers cares that it costs £25 or £35. Especially when you have to buy a bunch of other things to make it work anyway, and then waste an evening trying to get it to boot. There are of course a few electronic engineering types having gentlemen’s accidents over the GPIO port and that’s great, but I seem to recall that the Pi was all about getting a generation of kids into computing as we used to know it, back when we had Spectrums, 64s and Amstrads. The problem is that price was only a small factor in the choice of whether we owned one of those home computers back in the day, yet it seems to have been absolutely the driving factor in designing the Pi.

What I’d Do To Make Me Happy

Were I to think about the successor to the Pi, apart from the usual guff about making it a bit faster and giving it more RAM through the inevitable march of progress (nabbing the chip out of a Galaxy S II would just be incredible), I’d ditch HDMI and composite out in favour of a technically obsolete SVGA connector. I suspect the cost would be as near as identical as to make virtually no difference to the prospective buyer’s financial situation – as we’ve already established, the cost of a Pi is only actually a fraction of the total cost of actually using a Pi. Not only would this mean you can use that old Dell monitor, it’d also work, unlike the HDMI port, which doesn’t.

Secondly, I’d supply the Pi with a 4GB SD card plugged in to it already, with an OS on it, that boots. In fact I’d probably consider wedging the thing inside a keyboard casing with a trackpad and then you’ve literally got something just like one of those home computers of old, the spirit of which the Pi is attempting to capture. If you sold the bundle of things ready assembled for under £100 you’ve got a no-brainer for any prospective parent thinking about getting one for his or her geeky kids.

Lastly I’d switch from using microUSB power input to a standard 5v DC power input. And then I’d supply the device with one anyway, in the box.

The choice of Linux as an OS is unfortunate but a bit of a necessity given there’s bugger all alternatives yet. Linux is incredibly complicated. Just look at the BASIC interpreter command prompt startup of a Vic-20 compared to the Pi. Eek. This isn’t really going to help a new generation of geeks get in to programming; it’s likely to make them think, this is just too much effort for too little gain. Well done for creating the sort of barrier to entry that filters out all but the most heavily bespectacled and introverted propellerheads but is that really what we need? I believe that the barrier should be as low as absolutely possible in order to hook people into fiddling before people need to get into the nuts and bolts. Look at Mac OS. It’s got Unix underneath but even an idiot can use it after switching it on. I think the Pi is going to need something like that.

Anyway, more ramblings about the Pi later, as I get to grips with the horrors of Linux and C programming (Python! You must be joking). Eventually I’m hoping to get some Java code running on it.

Afterword

You might think I’m being a little negative about the Pi so far. Actually I’m having a reasonable amount of fun tinkering with it; my fears are really more meta-fears; I suspect that the Pi will turn off more people from software engineering than it will turn on, supplied as is. It is the very raw roots of modern computing. Unfortunately those roots are ugly, messy, nasty, clunky things; and one of the chief reasons I don’t use Linux as my main OS, and also one of the main reasons I use Java as my weapon of choice.

Talking of Java, seeing a bit more of an effort to support Java on the device wouldn’t go amiss. After all, it’s very, very likely that Java is what they’re going to be actually exposed to in higher education. Not Python.