Raph Koster's Home Page

Good Design, Bad Design, Great Design

This was my talk given at the Social and Online Games Summit at GDC 2012.
You can also download a PDF or PPT of the presentation.

A Slide from the presentation.

What makes something good, bad, or great? And (in your more cynical moments!) does greatness even matter, when we are chasing our daily ARPU?

There�s a lot of lessons from other fields here� from business books, from marketing, from architecture and computer-human interaction and industrial design, from psychology and even the obsolete field of general semantics. And some of them are ones you have probably heard before. But they lead to an interesting place.

A Slide from the presentation.

The FAMILIARITY HEURISTIC also known as the �mere-exposure effect� says that there is a rule of thumb that we use automatically as shorthand: judging things as better/more common/important because they are familiar� particularly when under high cognitive load. It applies to things like interface approachability (ref Don�t Make Me Think by Steve Krug) as well as brand loyalty. We see this in operation in social game Uis for sure.

That said, we chunk things up. We take stuff for granted. And then we cease to see it. In fact, the principle of CONSTANCY says that we�ll even perceive things as unchanging despite changes in lighting or distance. Repeated ads lose impact. Babies prefer stripes over gray, moving over static, solid over flat, novel over familiar. The cognitive process is based on seeing a new pattern, increasing perceptual fluency, and then this fires rewards which increase positive affect. Humans love finding new APPROACHABLE patterns. But complex and interesting stmuli stengthen it back up again.

This simple aphorism then relies on human nature.

A Slide from the presentation.

That�s only one example of the way in which design is focused on human nature. Consider the classic example of affordances on doors� or on Legos. Or the desktop metaphor.

Exploitative design simply finds triggers like those listed in INFLUENCE by Cialdini, and attempts to make use of them. It does work, at least until people see through you.

But humane design has SYMPATHY for the user. It takes their side. Humanistic design is marked by a concern for the welfare of humanity. It is not antagonistic, and thus great design does not actively oppose the user�s best impulses.

I�m going to leave that there as a value statement, and not justify it right now� but let�s look at some consequences of this approach.

A Slide from the presentation.

Don Norman speaks of CONSTRAINTS � limiters on actions. These prevent users from getting into trouble. A classic example is a slider � it has an upper and lower bound. But a user can put the slider anywhere in the middle they want.

A lot of bad game design set the slider bounds at one. A social game example might be the game premised entirely on quest unlocks. Recently I wrote a post about how narrative is a feedback mechanism, not a game mechanic. It has more to do with controlling the user than it does with giving the user experssive power. But as we know expressiveness is one of the, if not THE, classic elder game.

Instead, you have to invite users in. Expressive power and choices are what make people learn, and then develop attachment. There�s a phenomenon called DEPTH OF PROCESSING. That says that users retain info that is analyzed better than info that is not. Your game will stick in people�s minds better if people think about it. And that will lead to greater brand identification.

In Disneyland, Walt called guideposts towards delving deeper �weenies.� In architecture and marketing they speak of �points of prospect� and �progressive lures.�

A Slide from the presentation.

The familiarity factor in cognition leads to forming habits. This leads to successful products. But habit can be created in many ways, and even coerced. Timed re-engagement, and gift reciprocity are two examples.

Needless to say, bad design causes friction, creating an opportunity for a better designed product to come in and steal your lunch. Consider what the design of the retail model is doing to the packaged goods industries, or how the iPhone disrupted the phone market. That said, mere vendor lock-in is not driving a DESIRE to stay with a provider. Given an alternative, people move.

Brand loyalty is defined as sticking with a brand even if the competition is cheaper or higher quality. The key drivers found by research include perceived value, trust, repeated purchase behavior � and it�s been shown (Dawes, in 2009) that long-term customers are less sensitive to price increases. Passionate customers � even to the marketing buzzword level of �lovemark� � drive evangelism, and word of mouth is the top factor driving purchases and trials.

A Slide from the presentation.

In educational theory there is ample evidence that strong guidance gives strong results. That is, you have to set expectations, lead students towards results through feedback and assessment, etc.

We often think of this as meaning we should control the user. We see this in many games today, where the user has an illusion of choice, and thus an illusion of learning. And games are, really, about learning, not passively experiencing.

And indeed, it has been empirically shown that there are clear gains when you learn from a lecture � but it turns out that lecture�s effects fade, the gain is short-term only. In the end, all teaching is self-teaching (studies on online self study compared with traditional lectures). Studies show, however, that the best learning space characteristics are focused on adaptability: supporting people and varied activities, and change. Not rigidity. The next most important characteristic is social: collaboration, interaction, and engagement. (2008 Herman Miller/SCUP Subvey of Learning Space Design). All the recent trends in educational theory lean towards active student participation: something games are teaching teachers�. But many games are drifting in the opposite, more broadcasty direction.

A Slide from the presentation.

Design being invisible relies on principles such as CONSISTENCY� there�s visual consistency, but also consistency in behavior, particularly functional consistency, which is consistency between the action and the result.

In industrial design, there�s a well known premise around the tradeoff between flexibility and usability. HICK�S LAW says that the time it takes to make a simple decision increases as alternatives go up. That said, successful designs follow a hierarchy of needs � you might know Maslow? And lead you up the pyramid.

A great design is going to tell you about things that you can do that you didn�t know you could do. There are many design techniques for this, such as hierarchical organization of complexity, progressive disclosure, etc.

Arguably, one of the biggest issues in social games to day is that there aren�t any new capabilities in them that you discover. They don�t scale with the player.

A Slide from the presentation.

We provide feedback to users about their actions in order to tell them they did well or poorly. These are many best practices here� everyone played Peggle?

Our use of feedback in a game is based on not only providing feedback to the user on how they did but also in FRAMING. The use of excessive flashy colors and sounds to reward a player for actions that are not actually serving their needs is in order to create a particular perceptual frame on the activity. The thing about framing is that it is a propaganda technique. That has its place. But controversial I know! -- the point of social games is not to persuade users to play the social game more; it is to enable the user to have fun. In the end, because of increasing familiarity, we end up actually losing the impact of this icing on the cake, and that�s when users see the cake is actually sawdust. Because of the iterative tit for tat behavior, this actually results in actively alienating consumers, preventing later revenue from the same source. The other principles I already discussed get across why merely dangling another carrot is not the best long-term viable strategy for revenue or customer loyalty.

A Slide from the presentation.

The commonest reason for the above typical failures is failing to put oneself in the mind of the users. We think of a product as being bad or exploitative design when it exists more for the sake of the developer than it does for the sake of the user. Providing value to a customer is the heart of a lasting business.

It�s always a mistake in product design to not ask the customers what their needs are. But it is equally dangerous to accept what they say as the truth. They are going to give you an implementation as an answer, not the REASON why they are asking. There�s a thing called Performance vs Preference in ergonomics (Bailey) and usability (Jakob Nielsen & Jonathan Levy)� the Dvorak keyboard IS better. But QWERTY wins� people often prefer something suboptimal. This is why the dogfooding practice is so important, and it is why Steve Jobs led Apple not based on focus groups but based on user testing. They are not the same thing.

A Slide from the presentation.

People are resistant to the new, though.

The W3C lists four principles for accessible designs: perceptibility (can you perceive the design), operability (can you use it), simplicity (can you understand it, regardless of skills), forgiveness (minimizing the consequence of errors).

A best practice here is to build skills atop skills. Properly chaining skill learning in a game is an art: you have to relate the new challenge to something the user already knows, then specify what the new info they need to learn is, then provide that info. This is called �advance organizing�� you may also want to check out Dan Cook�s work on skill atoms. The goal is a positive feedback loop in player skills.

A Slide from the presentation.

The 80/20 Rule applies everywhere. As used in design practice, it says that �a high percentage of effects in any large system are caused by a low percentage of variables.� (cf UNIVERSAL PRIONCIPLES OF DESIGN)

Users have an aesthetic appreciation for this � and research shows (cf Norman, Asch, Kurosu & Kashimura) that less aesthetic designs have lower apparent usability.

In game design, I refer to this as ELEGANCE. The best games, longest lasting games have few rules, and few special cases, and achieve their replay through the core puzzle to solve. If you find yourself adding a rule to patch a rule, you may need to go back and redesign. If you find yourself making a game where you have to keep adding content to keep players coming back� that is a design failure. And think of the powerful long-lived games as a result of the opposite: Tetris, Go, Bejeweled, the FPS.

A Slide from the presentation.

You can�t accomplish all of this by accident. There is still a sense in the social games industry that design does not matter as much as metrics. But metrics are only good for optimization. Design is a discipline that sees beyond that. It is true that you can brute force your way with metrics if you have infinite tries and money, but it is more economically efficient to use these principles to land higher up the hill in the first place.

In general, working in the abstract is a great way to never get the lay of the land. �The map is not the territory� � Alfred Korzybski. I�ve often said giant design docs are like making a movie off the director�s commentary. The biggest hazard of product managers who do not have design sensibilities is that they working purely on a map made of metrics, and not with the grain of the wood, the tensile strength of steel, the flex of the rules and the rigidity of player strategy. Yes, with Free to Play, we are more in a world where designers must understand business. But let us not forget that this is an entertainment industry, not ecommerce.

The best way forward is evolutionary prototyping, with regular periods of playtest. Metrics ARE a wonderful means of playtest. But as budgets rise, we are going to get pushed in the opposite direction.

A Slide from the presentation.

In the end, we need to make money. We provide value, users pay.

A hallmark of a market that has not figured itself out is one where we try to get as much of the user�s money upfront as possible, before they realize that there is insufficient value to be had. Think here of prepaying larger sums into wallets; box sale prices in retail games; push to preorders, etc. A beauty of our F2P model is that it is fundamentally pro-consumer � it forces the purchase to be made over and over again, with a fresh decision every time. The user is repeatedly assessing value. A prepay method betrays the fact that you don�t think the user will stick around.

This is probably the loftiest thing I will say all day, but you shouldn�t be doing this to make money. You should be doing it to provide value, and then the money will follow. All the greatest companies in the history of business have had their comeuppances when the users determine that the company cares more about its coffers than about its customers. Great designs spill off money because customers love and trust them to have their best interests at heart. This lies at the heart of the successes with 1000 true fans models in entertainment, of the success of companies like Zappo�s, and why Google used to not be evil � and today can�t seem to build traction on Google+. Cf Good to Great

A Slide from the presentation.

In the end, it�s not that you can�t build a successful company on bad design, inhumane design, exploitative design. You can. Just be sure you sell it quickly, or somehow change it.

The truly lasting designs manage instead to accrete around them loyal users who support them for years on end, through changes. From a business point of view this matters because they are then an easily addressable market of qualified users with high word of mouth factor, less price change sensitivity, familiarity with the product line. The acquisition and marketing costs are minimal. The product development process is dramatically simpler. The sign of truly successful design is that a community springs forth around what you have made.

A Slide from the presentation.

All design is an interactive conversation with the user, you see. Games especially, because they are fundamentally looped interactivity, aimed at improving the user�s mental models and mental state.

Failures in design are failures in interactivity. They are shouting though bullhorns, telling you what to do.

Great design coalesces tribes around itself. It becomes, in the end, less about the actual thing and more about the fellow travelers you meet on the road. Think of the Stratemeyer Syndicate, which took the idea of little pulp novels for kids � Tom Swift, Bobbsey Twins � and created a powerful connection with their readers, answering fan mail personally, etc, and ended up with Nancy Drew, which defined feminism for multiple generations of girls and women.

A Slide from the presentation.

Business realities may often force us to settle. Good design ships. But in the long run, as Miyamoto said, no one remembers that a game was late, they do remember that it sucked.

You have the chance to bring to the world something that people say �I know what that is, and whoa! It is more than I thought! And it makes me do things I couldn�t imagine doing� and have fun in the process!� You have to reach for the unlikely to do that. As budgets rise in this space, as marketing costs shrink our margins, we will have enormous pressure to be more and more conservative. But remember that disruptive innovation is what will drive that passion that creates truly great designs and businesses.

A Slide from the presentation.

It�s not enough for us to take this new space of social games and settle. Oh, we now know how to make a perfectly acceptable game that the audience says �Oh, OK� to. But the longer we hang on to the Gold Rush mentality that characterized the early space, the more we limit its potential.

Now is the time that this space can take risks, here at the moment of its greatest complacency, before it ossifies, before the door is closed.

A Slide from the presentation.

I can�t say that I have always followed these precepts. Many of them, I didn�t know when I started. Some I landed on through sheer luck. I have let bad design be shipped for fear of my job. I have let commercial considerations override what was right for the customer. I have been more caught up in playing with the delight of experimentation that fascinated me, rather than providing what the customer really needed.

I do know better now. It doesn�t mean I will be perfect. But I CAN say that I always WANT to do Great Design. And I want to know � stand up to answer, if you want

A Slide from the presentation.