His talk concerned the next generation game company. As a would-be game investing venture capitalist, Harrison expressed that he felt uncomfortable with the old school software development and business model used for games. He wanted companies that could be flexible, who could use metrics and agile development to create minimum viable products and test their assumptions. His ideal, as he explained it, was a company that could potentially add or remove game dynamics quickly from its games, be able to measure user behaviour immediately, respond to audiences, and do so at low cost with the prospect of high profitability.
Harrison was advocating a message of customer validation, low cost, and iterative process as the secret to game development in the 21st century. This is a whole school of thinking that the Silicon Valley people call the lean startup.
What is ‘Lean’?
Lean is an approach to building products and companies that involves several phases.
The most important part of those phases is customer validation, which means proving that something works in the market by earning even a few dollars with a low-cost version of your product before throwing significant resources behind it. With the attitude of quick building, measurement, learning and meeting customers in the real world, you can avoid many of the traps that starting companies fall into, and get to the heart of what your business is really about fast.
After validation comes scaling. To scale, you take what you have learned so far and figure out how to deliver it to your whole market. Whether it's razor blades, detergent, televisions or search engines, scaling is what turns a startup into a billion dollar company like Groupon.
It has become a very popular method for starting software companies in particular. First written about by Steve Blank in his book, Four Steps to the Epiphany, and later by Eric Ries on the Startup Lessons Learned blog, the idea of lean has taken root in the wider business community. In these tough economic times, lean is especially attractive because it promises low cost before success, not as a precondition to success.
Lean has become something of a movement in the business world because its message is simple and pragmatic, and promises a template for success.
What is a Secret Sauce?
What the lean idea is mostly about is discovering a secret sauce. Customer validation is the equivalent of making lots of test sauces and getting people to try them to find the best one. Then, knowledge in hand, you can invest in a factory to mass-produce it.
The idea is that if you can create a product that has a hard-to-copy quality that people love, and is also easy-to-manufacture, then you can build a business on that. By being hard to copy, a secret sauce is a valuable asset. If it’s easy to manufacture then it is easy to scale.
Google's search algorithm and Facebook's social graph are two examples of secret sauces. Both can spread all around the world for free and generate huge advertising revenues, but at the same time very few people know how they actually work, so they are protected.
Here’s the kicker though: What actually makes the secret sauce work is not the sauce itself. Lean assumes that validation through purchase (or some other kind of transaction) indicates objective proof that the sauce is the right one, but this only gets a business so far. What actually makes a secret sauce explode in business terms is the marketing story of the sauce.
The successful secret sauce has a story that attracts investors or customers. Their funds allow the company to spread its story further. This in turn attracts even more investment or purchases, which allows further reach, and so on and so forth. Revenue then comes either from the gigantic number of users that the company has acquired, or from the customers that the company has built.
Of course the sauce cannot just be based on a fairy tale. The Google search engine is actually really good, for example, and it is carefully managed to remain friendly and uncluttered. Google works at many innovations in search all the time and releases product after product with which the public can play and love.
These products are often whimsical (Google Earth, for example) but their power is to reinforce the marketing story of Google as an extremely smart and nice entity that gives you the best software on the web. Their search engine is thus trusted. It’s an on-going tale that shapes attitudes and makes people choose Google over Bing, even though in reality there’s little or no difference between them.
This story vs reality aspect of secret sauces has been proven repeatedly by example and experiment: A developer does a test where he swaps the front ends of Google and Yahoo so that a search in one produces results from the other. Test subjects say they prefer the 'Google' (actually Yahoo) results. Why? Because they believe it's better, not because it actually is better.
This is not the sort of quality that is easily validated, which represents a problem for lean. Marketing stories often need time to build up authenticity over several cycles. Google is not really a lean startup because it didn’t validate itself with earnings until quite a long time after the search engine had been taken up and funded.
The marketing story is not just something that you tack on. It has to be lived and breathed by the company, and this matters more in some industries than others. In games, for example, the marketing story behind Blizzard, Valve and Popcap has been long in the making. It is born of a track record of success, over-delivery and making brilliant games.
Entertainment Is Not A Problem/Solution Industry
Some kinds of company are not based on secret sauces at all. They do not have a hard-to-replicate process that drives the story of the business forward. There is no Blizzard algorithm or Popcap social graph. They have built a marketing story as artisans, companies who only release their games ‘when they’re ready’, and the marketing stories of successful game companies are pretty much all the same.
Artisanal companies don’t have a hard-to-replicate process. What they have is hard-to-replicate people. Such companies include high fashion, music, acting, the visual arts and literary fiction. It also includes games. Unlike most lean-oriented companies, artisans are not trying to find a problem to solve.
Solution industries, such as most software development, are trying to find a provable, replicable secret sauce that will scale. Google solves the problem of finding things on the internet. Facebook solves the problem of finding your friends online and keeping tabs on what they are up to. McDonalds solves the problem of hunger.
Zynga does not solve any problems, nor do Bethesda, Valve or Bioware. Instead they are trying to entertain. The kind of thinking that Phil Harrison evangelised at the Games Invest event is thus rather misaligned.
Useful Aspects of Lean
While using metrics to build aspects of a game and tune it are entirely appropriate (and more companies should be doing this), Phil’s root idea of adding and removing dynamics on the fly can quite easily result in a game that is fragmented or simply rubbish.
In order for a game to entertain, it needs a coherence that creates gameplay loops. When writing about CityVille the other day, for example, I talked about how the game is structured to create many open loops. Zynga are a company that many of the would-be lean advocates in games look to as an example, but in reality Zynga is a massive operation with hundreds of people that produces one kind of game very well.
Their games are still fully-formed however, and they don’t add or remove dynamics willy nilly. What they actually do with their metrics is customise what they have, adding or removing items of inventory, adjusting prices and balance. Like every other successful game studio has ever done in fact, only they do it with billions of data points from millions of users.
What they don’t do is suddenly change how collecting objects in the game works though, or how game tiles function. So we have to be careful about taking the lessons of lean too far in games. There is still a baseline of functionality and entertainment required in games for them to work before you can learn anything useful from their metrics. This baseline is a lot higher than the minimum viable product ideal of lean however.
That’s why I think it’s smarter to talk about games in terms of clean development rather than lean. Lean is all about building that killer function that will turn into a sauce, but games don’t have that. Games are not expressible as single functions. They are, however, expressible as dynamic loops. So the secret to clean is developing an elegant, extensible game dynamic that provide enjoyment over many generations.
Some of these already exist. Can you tell what they are yet?
(If you found this article insightful, show your apprecation with a tweet!)