In response to a question from a post a few days ago–and a question I always get asked and seem to answer over and over–here is an effort to lay out the differences between opinion and judgment so that we can put to rest the false claim that “subjective” is always worse than “objective.” I have blogged about this before (e.g., this, this, this, and this), but I also understand that is not something that one can just hear once and then deeply get it.

That said, it is very, very important that you all start to understand this now, and understand it deeply. So if you have questions, concerns, skepticism, counter-examples, “what about …”s, and so on, please write about them or ask them. This is absolutely essential to understand! Anyway, here is my lengthy, example-laden response to that post:

OK, the problem here is that you are confusing “subjective” with “opinion.” And “opinion” is considered to be bad reasoning, totally worthless, and since you don’t distinguish it from “subjective,” then that seems like it must be worthless, too.

So let me start with a simple definition of opinion: We normally think of opinions as our non-reflective gut reactions, preferences, and tastes. Everybody has lots of these. They are informed by our life experience (so they are not groundless or worthless) but that informing happens in non-systematic and non-structured ways. It is usually not self-critical or reflective.

Now let me talk about subjective. “Subjective” refers to reasoning that has its origins in the subject, rather than in an object. But what is this subject? This subject is, of course, a person. And people vary: at one end, you have people who merely have opinions. And on the other end, you have experts. Now–and this is important!–these experts aren’t experts because they rely on external data. They are experts because their capacity for subjective reasoning and interpretation is particularly valuable.

In a word, these experts are experts because they have and can wisely exercise judgment. And judgment is, like opinion, a point of view or a personal perspective. But unlike opinion, it is reflective, self-critical, experienced, systematic, and grounded in experience. Now experience itself is subjective! We learned this in I502 Experience Design. Experience refers not just to “what happened” but to the way that the “what happened” is organized and made sense of in people’s minds, memories, and stories about it (and sensemaking, organizing, and storytelling are all fundamentally subjective). Experts have lots of experiences and lots of practice making sense of these experiences. As Nelson and Stolterman write on the concept of judgment,

Judgment is not founded on strict rules of reasoning. [Jeff: That would be a Cartesian approach, since these rules would transcend any actual reality.] It is more likely to be dependent on the accumulation of the experienced consequences of choices made in complex situations. [Jeff: Please note how phenomenological that sounds: choices = actions, complex situations = particular everyday realities.] … To put it simply, judgment is knowing based on knowing that is inseparable from the knower. By this we mean that judgment is based on accessing knowledge that is generated in the particularity or uniqueness of a situation; knowledge that is inseparable from the knower and is only revealed through the actions of the knower.

There is much more in their chapter on judgment, and if you have any doubts about what I am writing here, or if you continue to believe that subjective judgment is inherently bad (as you suggest in this post), I enthusiastically recommend that you read it.

The point here is that sometimes, the data speak for themselves, and when they do, objective reasoning techniques are great. Imagine that we run the food and concessions at a baseball park. We know that when it is hot, we sell more cold drinks and ice cream than when it is cold, when we sell more hot food and hot drinks. We can collect data about how much more hot/cold drinks/food we sell, based on the temperature, and develop statistical models to predict the sales. Then we can order and prepare the right food, based on data. That is objective reasoning, and in that example, it would be very effective. (Credit where it is due: This example was given to me by my friend and statistician, Bob Michael.)

But sometimes, data don’t speak for themselves. And you have chosen a field where this is usually the case. For if it is obvious why hot chocolate sales are low at baseball parks when it is 95 degrees outside, it certainly is less obvious why some fashions this fall are doing much better than other fashions. But fashion design is more complex. People buy clothes for an incredible array of reasons: cost, texture, color, time of year, self-expression, symbolism, silhouette, what else they have in their closets, brand loyalty, convenience, marketing, whim, and so on. You can try to model fashion tastes at a given point in time based on objective data (people do), but it is clearly a much harder problem than predicting that ice cream will sell better on hot days than cold ones, all other things being equal.

But as a designer, no one will ask you whether ice cream will sell better on hot or cold days. No. You will be asked to deal with situations whose complexity cannot be modeled, where data can inform decisionmaking but cannot be sufficient to make decisions obvious. And in those circumstances, you will have to make decisions based on your own interpretation, reasoning, point of view, and experience. And if all you have is “opinion” then you won’t be a designer for very long.

Again, remember the example of the National Geographic cover: they don’t do empirical studies and statistical modeling to determine which photo will sell more copies; they have editors with experience-based judgment who make those decisions. And they never put lousy photos on the cover. How does Vogue know which clothes to feature? Same thing. How does Kieslowski make an art film? How does Kickasola tell us the extent to which and the nature of his success?

All of these are grounded in an expert with substantial experience practicing judgment by interpreting reality in its full complexity. This of course includes objectively collected data. Vogue does reader surveys to “decode” tastes and hot topics. But ultimately their decision-making processes include–but are not determined by–the outcome of those objective approaches. Editors make decisions based on this kind of data; what they see today in runway shows; what they have seen in their lifetimes in magazines, runway shows, on rock stars and celebrities; in Old Navy and Saks Fifth Avenue; in Milan and Sao Paolo, Beijing and Martinsville; on the history of fashion; on cultural trends outside of fashion; on previous performance of certain design studios; on punks, goths, college students, tweens, and the sons and daughters of the wealthy. There is no objective model for all that!

I hope I have made my point.

Now, you ask:

I am thinking of writing the criticism of the interaction of Dental Restoration Sculpting Software. But I am not in the field of dental technology, and I haven’t use that software to do daily tasks as the dental technicians do. How can I write a criticism about it from something “subjective”? If I try to observe how technicians using this software to do tasks, as well as interview them, I believe I can claim something that is persuasive. But I guess that is the approach of “decoding”.

This might not be a great topic for you, if it is indeed the case that you don’t know very much about it. At the same time, you are already more of an expert on this than you realize. You are a graduate student in interaction design. You have read quite a bit of interaction design literature. You have been lectured in classes on design theory and practice. You have been exposed to dozens of key concepts in interaction design. You are surrounded by smart fellow students who talk about interaction design all day long. You have access to a rich variety of interaction design types. In short: you are already more of an expert on this topic than you give yourself credit for. The fact that you think that maybe you should observe or interview technicians is evidence that I am right, because that is one way that interaction designers collect the evidence they use for their decisionmaking. Last week I spoke to a grad student in CS who was working on a very cool project and who never even thought to study his or her users. The point is that you may be put into situations where you don’t know as much as you would like to; you have to mobilize whatever it is that you do know, and leverage it as well as you can (and that is a subjective skill, not an objective one).

This class is about cultivating judgment in interaction design. That is why we are here. No one is born with it. It is not a personal attribute, like a blood type. It is something one cultivates and develops over time. That cultivation must happen through practice (not just reading and listening to me go on about it). That’s why I’m making all of you do it. Not because I think you are already Kieslowskis or Vogue editors, but because “designerly ways of knowing” (Nigel Cross’ term) are centered on evidence-based, responsible, reflective, systematic interpretation of unbelievably complex situations, about which only a part can ever be known (and designers often have to decide which part they want to know). And because designers must act–they always act. And this action always happens in the spectrum between ignorance and opinion on one extreme and knowledge and expertise on the other. Good designers are able to push themselves toward the second of those two extremes, but they cannot ever actually get there.

Critique is a skill (and there are others) that designers cultivate in order to identify and mobilize the best of what they know and can know in order to act in a complex situation is a responsible and effective way.