Hi, I'm Adam. Email me about: CTOs, Mobile/iOS, Project Management, and Development

Micro-micro-payments via electricity bill

A new startup has appeared that essentially charges people a per-second subscription to play games that is added to their electricity bill.

On the one hand, I am personally against this kind of “invisible” charging because it can be – and indeed is being – mis-represented as “free”, when in fact every single user is being charged. I also feel it sets a set of very bad incentives for the lower-end game developers – essentially encouraging them to make lower-quality games. But that isn’t really a problem for me, it’s a problem for them, so … shrug.

So bear in mind my latent antipathy :) which I’ll try to keep quiet about – but for me the really interesting part is the use of this as an alternative payment system.

The Premise

On the face of it, they started with some assumptions like this:

Computer games don’t use much processing power, and yet game-players have really powerful PCs – just go look at the market for buying new PC’s: those marketed at video games players are the only part of the PC hardware market to have preserved and grown some decent profit margins on hardware, and very high prices.

Meanwhile, game developers have large amounts of extra content for their games that they don’t charge for and which they don’t allow the players to experience – it’s all there, but they can’t be bothered to let people play it. I mean, why would you? You’ve paid to create it, therefore you want to keep it secret.

Also, casual game developers, those making all those Flash and Java games you see on Miniclip, are all awesome experts when it comes to hardcore programming optimization. They might be a bit lazy, not bothering to do it a lot of the time, but despite this total lack of practice, they are all latent geniuses if you give them enough incentive.

…although actually I think they didn’t really think any of that, they just came from a compute-farm background and jumped on Games as an easy first target market.

Plura Processing is embedding an app in browser-based games that uses local CPU to run remote processor-intensive tasks – and then they charge the remote task-owner for the “rented CPU time”, and give a percentage of that back to the game developer in whose game the app was embedded. You can see a bit more on it on GigaOM, including a brief response from the company. Their VC backer, Creeris Ventures, appears to have an exec on the board who came up with this in defence of the product:

“At the end of the play session, the game sees how much compute time the player provided and rewards him with an in-game item (e.g., +1 longsword).

So what you get is something that the user loves (because he’s being directly rewarded) and a model for computing power that does work (because it’s using the PC as efficiently as possible). Both sides win. :)”

Yep. They are actually encouraging you to reward people for leaving their computers logged-in and churning up electricity by gifting them in-game items. I’m all in favour of making computer-games unfair, but I don’t see this as a sensible way to do it (from a design perspective).

A quick perusal of my LinkedIn network suggests that the company founders aren’t from a games industry background, so I guess that they thought of charging people for CPU time (an old business model from the late 1990’s that never took off, despite many tries), and then spotted games as just a good market they could take advantage of.

Something fishy

Plura is lightweight and secure

Plura applets use minimal resources on most computers. Our system is secure and each application’s integrity is verified before it can use Plura.

If you’d like to see Plura first-hand, just click here. We think you’ll see how harmless Plura really is.

…given that your business model is to use as much resources as possible, the above statement is surely only true via sophistry? (they’ll be going for “minimial resources” when it comes to RAM usage, clearly, but maximal *CPU* usage, which is what most people care about given how cheap and abundant RAM is today). Am I missing something here?

Anyway, that’s not what this is really about, IMHO, because:

Business Analysis

Finally, because Plura is an enabling technology, it allows other people to create new sites and web applications that previously could not be financially supported. Plura does this in two ways. First, it offers a completely new revenue stream to new and existing sites.

How much money?

Well, according to their website, they’ll give the developer $2.60 for each computer that spends a month devoting all its CPU power to their processing. Is that a lot?

Let’s put this into perspective.

The ratio between “users online” and “users” is around 1:100 for a free game, 1:20 for a free MMO, and around 1:6 for a subscription MMO.

So, with Plura, in a subs MMO you would be earning at the rate of around 40 cents per month per user.
At the opposite end of the scale, in a free web game, you’d be earning at the rate of around 2.6 cents per month per user.

That’s assuming 100% CPU usage devoted to Plura. In practice, an MMO would be sparing around 1% CPU (assuming this app is correctly configured), and even for a free web game, you’d have IM running, probably Outlook, etc, and you’d expect with a *well-written* flash game on a Single-Core sysem to be seeing at most 30% of the CPU going to Plura, probably more like 5% realistically (pure guess, based on the CPU usages I’ve typically seen of Flash apps on the few occasions they’ve been tuned for minimal CPU). NB: this WILL change as adoption of Flash 10/11 increases, assuming that the rendering engine efficiencies come in as expected (flash prior to 10 uses the CPU for graphics).

What about dual-core systems? So long as flash games are single-threaded, Plura could get 50% CPU to itself (well … fighting off Outlook, IM, etc – so perhaps more like 40% to itself). I’m not about to try it on my Core2Duo here, though, since this is on a laptop and anything that maxes both cores will usually overheat and crash the laptop (a problem with the Macbook Air and it’s Intel CPU. Sigh)

So, my guess is that a typical casual game developer will see around $1-$5 per thousand users, whilst taking away their users’ CPU power. At the bottom end, that really does NOT compare favourably with advertising revenues. At the top end, it starts to look worthwhile; all depends how much CPU power you give up, I guess.


As a developer, you are merely replacing one agency-based fluctuating-value revenue model (advertising) for another (compute-farming).

Even with the many agencies and advertisers and healthy industry behind advertising, rates soar and plummet unpredictably, which is enough to put game developers out of business all on its own.

It’s going to be a lot worse working within what is by comparison a tiny niche marketplace, and where you are so heavily disengaged from the purchasers.

I’m sure plenty of people will try it out since it’s simple to attempt, but it doesn’t currently appear viable as a source of revenue for any game studio that wants to remain in business long-term.

Cost of Goods Sold … to the user

How much is this costing in electricity? Interesting question…

For a modern PC, with an Intel Core Duo, the difference between 0% CPU and 100% CPU is around 60 watts, which works out to a little over 40kWh per month.

At the moment you pay around $0.2-$0.5 per kWh for electricity. So, the electricity used by each home user will cost them from 80 cents (cheap power, efficient CPU) up to $4 (expensive power, inefficient CPU – e.g. the Pentium 4 uses almost twice the watts!).

i.e. as a game developer, you are being paid at just around the cost of the electricity used. That’s how this could be seen as a somewhat odd (but nevertheless interesting) indirect form of payment-system for users without credit-cards etc.

Growth potential?

But it’s capped at very very very low ARPU. Ridiculously low. This is a poor business model for anything except lowest-common-denominator games. Strangely, the company advertises it on their own website as a “replacement” for advertising, although they also run testimonials from a game developer who tried it and said the opposite.

(that developer, by the way, was Paul Preece, famous for making just one game (Desktop Tower Defence), and I suspect he’s getting a much better rate than $2.60 per month given the huge marketing value to Plura of having DTD as a “client”).

Market: gamers with powerful computers?

  • Generally, gamers who have spent vast amounts of cash on powerful computers are playing powerful games – not casual web games
  • All games, everywhere, use 100% CPU. The basic, fundamental model of game development mandates this: it is far too expensive to faff about on optimization that is not absolutely necessary
  • A game that does NOT use 100% CPU does poorly in the market, because the players will condemn the noticeably poorer graphics, framerate, AI, sound quality, reactiveness (of controls), etc.

Market: casual gamers?

  • Typically have weak and crappy PCs
  • If you’re playing games in a browser, that almost always means you have other apps open at once
  • Generally speaking, you don’t want the rest of your PC slowing down because a low-quality game is killing it
  • Flash is already phenomenonally bad on some platforms, and will steal 100% CPU at the drop of a hat; this causes lots of problems in the game-playing community of casual gamers who don’t understand it, but condemn games for it and refuse to play them. If you played much on popular casual games sites, you’d notice this uncommon but not rare problem
  • Flash game programmers generally know little about how to write good code (this is why Flash is great – it opens up the market to more programmers, rather than requiring everyone to be a highly experienced, highly trained expert), they certainly don’t write efficient code; if you want them to start learning how, expect a long long uphill struggle to re-educate the entire worldwide base of Flash programmers

So, I don’t think this is particularly exciting as a revenue model itself.

But I do like this idea of the consumer’s electricity bill being converted into per-millisecond subscription to the game they are playing; in a world where you are not able to get direct payment from your consumers, or are somehow prevented, this could *in theory* (I think the lack of established marketplace invalidates this in practice) be a good poor-man’s replacement for micropayments