The Atari Vision

Atari is undoubtedly one of the most significant brands in the history of the videogames industry, but more recently the company has suffered more troubles than it has enjoyed moments of glory.

With a chequered history that has seen stakeholders come and go, and the shape of the business change multiple times, the announcement that David Gardner had joined the publisher as CEO raised a few eyebrows - but not as many as when GamesIndustry.biz broke the news that Phil Harrison had left his post as president of Sony Worldwide Studios to become president of Infogrames.

In the months since that all happened, the level of industry speculation about what Atari - now the unified global consumer-facing brand of Infogrames - is planning has been frenzied, with suggestions coming in from all angles.

Would it be a Flash-based games aggregation plan, to tackle the likes of Pogo and Yahoo? Was it an MMO using one of the company's IPs? Or maybe a proprietary hardware platform, merging mobile with web?

In fact, it's none of the above, but rather a bet on internet-based technology such as the Unity platform, which enables high-quality 3D visuals in a web browser environment, in combination with a general shift in the attitude of the games development process.

The revelation came in a hotel suite overlooking the city of Copenhagen, on the eve of Harrison's keynote at the Unite 08 conference in October, as he took me through his presentation and explained that it was bringing together much of his thinking on how Atari will approach the future.

The key is to use this new technology to bring extremely compelling, visually-engaging games to vast numbers of people - who have no need to invest in new hardware, because they're browser-based. It's beautifully simple, and Harrison's excitement on the prospects for success was clear.

To begin with, the possibilities afforded by the Unity engine, as an example - Harrison was clear in his assertions that Atari hadn't at that point signed any deals to use any one particular platform - are wide-ranging. The specific demonstration he showed me was of a tropical island, somewhat reminiscent of a Far Cry or Crysis level, and within seconds - on Harrison's own laptop - we were gliding through the richly-detailed 3D environment at very nice frame rates. The laptop, I'm told, had no special hardware to enhance its performance.

Of course there's a huge difference between a 3D tech demo and a full-blown triple A finished product, and that difference has yet to be proven - but then the next part of the Atari plan really involves changing the conventions of just what a triple A experience is, both on development and end-consumer levels.

In terms of development, GamesIndustry.biz already revealed some of the main tenets of Harrison's argument back in October - fundamentally changing the shape of the cost versus revenue graph, which then allows much more freedom for experimentation.

This "experiment earlier, fail more often, and fail more cheaply" mantra, with far more emphasis on early prototyping, is something which other industry leaders - including Rare's studio manager Mark Betteridge - completely agree with, but specifically, how is it possible?

Essentially the problem with current development cycles, says Harrison, is that they encourage companies to continually pile on risk and cost as the project goes on, so that projects hit their "peak negative revenue" - in other words the point at which the game's costs have hit their highest point - just as the game is released.

What's more, it must then be hoped that the cumulative costs of perhaps two or three years of development, leading up to that peak negative revenue, will be offset by a handful of weeks in the sales spotlight - a process which puts extreme pressure on the marketing and PR teams to secure as much visibility for their titles as possible, and one that leads to a make-or-break situation.

Harrison's belief is that this is an unsustainable model, and that the graph should look very different. The main thing is that cost should increase only gradually as time goes on, but that the risk with the project should effectively lessen over the same period - and this is something that comes from the process of development, and specifically following a set pattern which involves early prototyping.

"All of the mistakes I have made in software development have been based around one problem and one problem alone, which is accelerating through this pipeline without successfully and properly satisfying the requirements of each of the stages - and typically it involves going from concept to production in one jump," said Harrison. "That's pretty much the definition of why projects fail - because you don't know what you're building, you don't know how you're going to build it, you don't know who you're building it for, but you've got 60 people working on it and they're all running in different directions - that's how most games fail.

"Experiment earlier, fail more often, and fail more cheaply: This is the mantra - you want to fail early, to kill those poor ideas, but you also want to do it repeatedly and quickly so that you will eventually find those great ideas, but you want to do it as cheaply as you can so you save money."

Another facet of the plan is to involve users at a much earlier stage of development, more akin to a beta phase, but involving less of the early look 'pressure of performance' that current testing involves - and that will require a re-education of the consumers that may take part.

"I believe that a true 21st century business model is to do all of that [production] in the glare of the game-playing public, so you can expose that innovation to your consumers," he said. "You may not charge them for it at this point - but why not deliver your innovations and experimentations directly to your players, and let them be part of the process of deciding which games to make?"

It's a nice idea, although quite how that re-education that's required will take place is unclear, particularly as user expectation has - usually unhelpfully - risen beyond all recognition in the past few years, most notably for MMO titles.

But if successful the plan should mean that the peak negative revenue point isn't a horrible downward spike when the game is launched. Instead it's an undefined point in the future, that's far more likely to happen some time after launch than before it - and crucially the costs are offset on an ongoing basis too by ongoing revenue from consumers.

Is this consumer revenue based on micro-transactions, or a subscription model? So far that much is unknown, but Harrison is clear on where he wants to position the Atari.com web portal, and that's firmly in the space between today's casual sites, such as Pogo and Yahoo, and the premium MMO products at the other end of the scale.

The trick will be choosing the most suitable title genres for this development pattern, as well as managing the post-launch activity - although the traditional concept of launch is unlikely to be used, but rather the Gmail soft launch style of 'turning on the tap' and at some point monetising the process.

And of course the other crucial spoke in the wheel is attracting, supporting and developing a community to live around the titles, preferably with meta-services and potentially some tools to enable user-generated content as well.

At the moment none of this implies an abandonment of the boxed product distribution system, although the recent Namco Bandai deal has provided not only a significant level of funding to the company, but also the opportunity for complete divestment further down the line.

The biggest question is, of course, will Atari be successful in its endeavour? Time will tell, but one thing is certain - morale and motivation within the company are both very high, and the leadership and talent that's being attracted to the "best-funded, best-branded, most energetic start-up in the history of computer gaming" is certainly compelling.

The company is hosting a global two-day event in London today, where it will unveil more of its traditional boxed product titles for the next 6-12 months - with, I'm told, a few surprises - but I suspect that the genuine success and reinvention of the company will be a long term project that may not yield anything visible as part of this new plan until this time next year.

Either way, while the speculation surrounding just what Atari's big reveal would be can be somewhat laid to rest, the industry's favourite sport will now shift its focus to whether or not it will work out, and when.