Problems with UserCentric Document Design


Section 1 - Why Is It so Hard to Write and Evaluate Writing on the Internet?

In 1997, I began studying a problem I saw with content evaluation among my graduate students in technical and professional writing classes. Before pointing out the problems, I should describe the students in question. It was our policy to offer our Master’s degree, online, only to people already working as professional writers. The students in question were all working professional writers. I showed these professional writers two webpages that had serious problems. This is the content from one of the webpages:

Students thinking about majoring in English inevitably confront the question: "What are you going to do with an English major?" Contrary to popular belief, however, career opportunities for English majors are quite favorable because English majors are adaptable. They have the critical thinking skills to adjust to a variety of different career paths, and in a world where workers can expect to make major career changes more than five times during a lifetime, adaptability is no small asset. English majors have found job opportunities in financial institutions, insurance companies, federal and state government agencies, the hospitality industry, universities, museums, and service organizations. They are employed as personnel and planning directors, administrative associates, marketing directors, technical librarians, wage and salary representatives, service correspondents, claims adjustors, and insurance agents. The English major is also an excellent undergraduate major for those who wish to enter law, medical, or dental school; complete post-graduate work in literature, film, creative writing or library science; or enter sales, management, and marketing programs in large organizations.

The above paragraph was on a page in a Tech. Comm website. It was supposed to attract sophomore level students into the program. Knowing the intended purpose and audience, you should have no trouble seeing the problem. The paragraph was originally designed to attract English majors (i.e., literature majors) and not technical writing majors. Between 1997 and 2007, I never found a single student who could identify the problem. These working professionals (all of them) were unable to connect the fact that none of the jobs listed were technical writing jobs to the paragraph’s purpose or audience. Moreover, again and again the copy mentioned English degrees, and they were unable to make that connection. Hundreds of professional writers over the years . . . and none could see the obvious problems.

The second example follows:

Students take classes in two areas: first, they build a theoretical foundation in rhetoric so that they can assess any writing situation and adapt their writing to the context as audience-aware, self-aware, self-confident writers; and, second, they learn about writing in a variety of contexts using the most up-to-date tools of technology so that they know both how to write and why they are writing, thus preparing them for the ever-changing job markets of the twenty-first century.

This paragraph is a single sentence that just goes on and on. Once again, its purpose was to attract new sophomores into the technical writing program. Obviously, it was copy-pasted from somewhere else. In 1995, our university was visited by full professors from Northwest Accreditation, come to evaluate the university for reaccreditation. This was written to impress full professors but was being used to attract sophomores. Just as with the previous paragraph, not one of the professional writers who looked at the paragraph was able to point out the problem – 10 years with a 100% failure rate.

So what was the Problem?

When people look at a specification sheet or proposal or love letter, they look with different conceptual filters they can use to evaluate the documents. A proposal has a job to do and most everybody who writes proposals knows what it is and can judge whether the proposal is doing its job. Specification sheets and love letters also have their specialized jobs and will be judged through different filters. When people see a webpage, they only have the webpage filter for judging the content. People do not recognize the many new genres involved and do not know how to judge them. Often, a genre will be divided into sections which appear in different parts of a webpage.

Consider an Amazon product marketing page. On the top left is about one-third of a spec. sheet. On the top right is a menu. Down below are more menus, more of the specification sheet, a product description, and a bunch of reviews. Those are all different genres and demand different filters for evaluating them. The biggest problem is that when we write or evaluate material for the Web, we often fail to understand the genre we are creating or evaluating. Earlier, I mentioned the product description box on the Amazon marketing page. That makes an excellent example of what I am saying:

Woven Poly Tarp, Waterproof, Mildewproof, Tear Resistant, Acid Resistant, 8 oz per square yard, 16 mil thick, 16x16 weave count, grommets approximately every 18", reinforced corner grommet, finished size approximately 2% or 6" shorter.

The sentence is the product description of a tarp sold on Amazon by an unnamed company that does not understand what you are supposed to do in the product description box. The name, “product description,” is misleading. This is where companies are supposed to sell their product. The above describes the product but does almost nothing to sell it. Writers, who know the real purpose of the product description box, fill it with content designed to sell their product. Sometimes they put so much in the box that if printed out, the content would reach from ceiling to floor.

In short, the problem is most writers do not understand how genres work in websites. There are old genres that can be dissected and placed on different parts of the page, and there are new genres that writers aren’t sure how to handle. Another problem can be found in how we describe ourselves as we write.

Meta-metaphors we Use when we Write

As professional writers, we may see ourselves as craftspeople (wordsmithing) or mechanics (fixing the text) or sculptors (molding the text) or researchers (collecting information and documenting it) or even translators (making difficult material available to a lay population). In each case we are seeing ourselves through filters we construct of metacognitive metaphors. Often unique to the individual, these metaphors give us different vocabularies for understanding and discussing our writing processes. Typically, we do not actually notice our metaphors until they are pointed out by others or we take the time to carefully explore our writing processes. Think about your own different writing processes. How would you describe them? To develop a vocabulary, you will necessarily construct descriptions based on metaphors you have internalized, usually subconsciously.

You may be aware of the processes in your writing, although you might never have considered them in detail, and you no doubt have different processes for producing different genres or modes of writing. For what it’s worth, I see my theoretical and fiction writing as exploration. I seldom know, when I sit down to write, where my ideas will go. As I write about a subject, I come to know it better. In a sense, the more I rewrite, the better I understand my topic. For computer documentation, my metaphor changes completely. I become a watchmaker, polishing, tinkering, and repairing until the text is (hopefully) flawless.

Metacognitive Metaphors Drive How We Write

Not only do we all have metacognitive metaphors for our writing processes, but those metaphors drive the processes to such a degree that one can tell (to some extent) which metacognitive metaphors were used while the text was being produced. In “Tuning, Tying, and Training Texts: Metaphors for Revision,” Barbara Tomlinson (1988) argued, “Such patterns of figurative expression are an important part of our socially shared knowledge of composing . . . and may well influence our composing behavior” (p. 58). In her research, Tomlinson examined more than 2,000 texts where writers described their writing process, consolidating the writers’ descriptions into a list of eight metacognitive metaphors.

When These Metaphors Fail

The metaphor we use can lead us in ineffective directions, however. Suppose a writer’s metaphor is more linear. Perhaps the writer sees himself as something akin to a conduit, spilling out the text in a singular and unaltered process -- teachers often see this metaphor in the “midnight specials” they receive from their students, usually turned in with a mumbled apology. This year, I looked at a high school, science essay written by a nephew. He had written 10 pages about the impact of Copernicus, Galileo, and Einstein on astronomy, and he received a D. As I talked to him about his process, I realized he saw himself as a funnel of sorts. He poured a variety of resources into his mind and dribbled them one by one onto the page. He didn’t even filter the material through his own opinions. He simply poured it out with never a thought for rewriting. Needless to say, the work was D quality, or worse. Fortunately for my nephew, the teacher asked for a rewrite for a final grade, and he ultimately got a B.

I am not suggesting my nephew was thinking, “I am a funnel.” He saw himself as simply writing. But he saw writing as finding material on one page and moving it to a new page—basically, nothing more than cutting and pasting. Had he seen himself as working within a metaphor that involved digesting the material, responding to it, and polishing his product with recursive writing, things would have been completely different from the beginning.

I am not suggesting my nephew was thinking, “I am a funnel.” He saw himself as simply writing. But he saw writing as finding material on one page and moving it to a new page—basically, nothing more than cutting and pasting. Had he seen himself as working within a metaphor that involved digesting the material, responding to it, and polishing his product with recursive writing, things would have been completely different from the beginning.

The “mechanic” who sees herself as fixing problems has a definition of perfection that involves elimination of all errors. Part of being flawless for the “mechanic” is selecting vocabulary for maximum readability and effective reception, making certain all statements are complete and accurate, and making certain there are no mechanical problems in the text.

Imagine writing a National Science Foundation (NSF) grant proposal. In this environment you meet severe style and formatting restrictions; everything being mechanically correct and accurate is critical. The NSF receives thousands of proposals every year. Often, the first thing reviewers look for is any stylistic or structural justification for rejecting the proposal. Missing or misnaming a topic can be enough to get a proposal thrown out. Moreover, if the vocabulary is surprising or the mechanics are faulty, reviewers will characterize the author as unprofessional and/or outside the discourse community. The demands of documentation and NSF grant proposals shout, “Be a mechanic!”

Such a metaphor can lead to competent work that communicates well. Still, the metaphor, itself, may keep the “mechanic” from seeing (or valuing) the possibility of going beyond a document that is correct or accurate to a document that could be called “aesthetically exceptional.” For example, sonnets, plays, and works of creative nonfiction should be something more than just correct and accurate. This is not to imply the metaphor “mechanic” is flawed or that computer documentation should be sublime. The “mechanic” metaphor is perfectly appropriate for the right documents, but it (as do all metacognitive metaphors) establishes a relatively specific vision for what an excellent document should look like—a vision that will never meet the needs of all documents.

In short, as we write, the processes we choose can be described using a variety of metaphors that can help us understand them. The different processes we choose, and the metaphors we use to describe them, can lead to excellent works, but they can also blind us to interesting and different possibilities and direct us into unfortunate choices.

Why We Have So Much Trouble

Just as we write using metacognitive metaphors as our filters, we also read with similar metaphorical filters. These, however, are genre-based filters. When we read a genre, we know what its purpose is—what it is supposed to be doing for us and to us. We know what the social interactions are. We know what the structures should be. Even if we cannot name the genre, we know what to expect from it, and we have no trouble seeing when it fails to do that. But what if we cannot identify the genre? What if all of the metaphors lead us in wrong directions? What if we are evaluating a text, but think we are evaluating a place?

If, when we design a website, our metaphor is “place,” we may fail to notice we are designing a document. If, when we evaluate a website, our metaphor is “place,” we may pay attention to navigation and atmosphere and forget that we are also evaluating a multitude of genres. We have so much trouble evaluating content on websites because we have so much trouble seeing that we are evaluating genres and not shops.

People Who Visit Websites and Contractors Who Construct Them

While computer documentation and NSF grants might demand we “be a mechanic,” sometimes documents are wrong when they seem to shout about how we should approach their publication. The common metaphor used in web navigation, design, development, evaluation, and discussion is “site” (a “place” metaphor). We go to distant sites and navigate them, and we evaluate sites based on the quality of their navigation and our sense of the design of the place. When we have tested these sites over the past 10 or so years, we have used usability studies. In these studies we look for breakdowns in navigation and failure to bring the users to whatever they need. We might ask test subjects to look for a particular piece of information or ask them to perform some task. What we will be measuring is how well they can do that. Designing a website that does that is called user-centered design.

Writing Theory Evolving out of Usability Theory

Writing Theory Evolving out of Usability Theory Within the metaphor of “place,” and in the tradition of usability theory, when we describe appropriate writing, our descriptions tend to be based on usability prescriptions—“keep it short,” “chunk it,” “use bullets,” “try not to write below the window,” “use subheads.” The list of prescriptions is quite long; it almost always involves structural terms and virtually never includes descriptions of style or rhetoric. p.4, describes writing for the Internet. The book begins with the following questions, “What did you do on the web yesterday? Were you just browsing around . . . or were you looking for something specific?” Redish answers, “Most people say ‘something specific’” (p. 1). Of course, she is absolutely right. Most people look for specific things on the Internet. But Redish moves from that statement to a claim right out of usability theory “Most people skim and scan a lot on the web,” and “Most web users are very busy people who want to read only as much as they need to satisfy the goal that brought them to the web” (p. 2). She also describes her own habits when using the World Wide Web (WWW) “Yesterday,” she said, “I downloaded a file, ordered a book, compared prices on a new cameras, read a few of my favorite blogs, checked the Wikipedia entry for usability, looked for information on a health topic for my elderly aunt ( p. 1). These statements do describe the processes of a huge number of Internet users, so it is reasonable for her to ask, “What makes writing for the web work well?” and answer, “Good web writing is like a conversation, answers people’s questions, lets people grab and go,” and “Think of your web content as your part of a conversation—not a rambling dialogue but a focused conversation started by a very busy person” (p. 4). That is good advice, but only for one community of WWW users. The truth is actually much more complicated. The WWW is made up of an infinite number of genres, and good writing is always writing that is appropriate for its genre. It is carefully considered and crafted by its author, and it meets the needs of its exigency. Good writing is not some prescriptive, one-size-fits-all (brief and friendly) style.

In Contrast, How I Use the Internet

My descriptions of metacognitive metaphors and genre filters above lead to results completely counter to the advice suggested by Janice Redish. Ironically, I also did many of the things she described. But I also did things she did not mention. I read an essay about Tiger Woods. I went to ABCNews.com and read several news stories and the reader comments that followed. I went to CBSNews.com and watched 60 Minutes in its entirety. I went to YouTube and watched a number of video reviews of video cameras. I marked up several student papers in an online class, using Word’s remote editing tools (I also spent time reading comments from students in that online class). I read several of the latest reviews on the new BMW 330 diesel. I read all of the research I could find on biofuels. And I read a 25-page scholarly paper about genres. Moreover, I downloaded a good book onto a computer (a Kindle) and took it to bed. In other words, I treated the web exactly as I would a library. I read a variety of different things written in a variety of different genres and in a variety of different styles—some casual, some exceedingly dense, some short, some very long. In all of that, I read from three large categories of writing: user-centric, persuasion-centric, and quality-centric.

Redish and the other usability experts are describing only one audience—“the user.” I suggest there are many purposes for web content for a large variety of different audiences. Redish and other usability gurus discuss only the first type of writing. She is right that many, perhaps even most, people on the WWW are users. Even so, we still have to write and evaluate content for the others, and the usability model does not work for that.

Persuasion-Centric Content

Consider tourist sites. Of course you can do an information “smash and grab” on the site, but the state of New Mexico would rather see you linger on its site. They hang out bait, and the longer you stay on the site, the more likely you are to take the bait. They have contests, and if you win, you get to go to New Mexico on an amazing ski vacation. They hold events: “New Mexico restaurant week returns for second helping” (New Mexico Tourism Department, n.d., n.p.). You will not likely buy anything on the New Mexico tourism site, but if they can put an idea in your head about how they can give you whatever it is you hold most dear (skiing, art, camping, history, birding, fishing, the list goes on and on), the site will have met its purpose. Maybe you will choose not to go to New Mexico right away. Instead, if they can stay in the back of your mind, you might be in a meeting where you are considering a company retreat . . .“I wonder if we could time the retreat to New Mexico’s restaurant week.” The point of this site, and thousands like it, is to get you to linger. This is a persuasion-centric site, and its task is not accomplished by efficient navigation, but by quality of content.

Quality-Centric Content

In a different environment, readers are encouraged to linger and learn. Lynda.com offers a good example. This site offers instruction by subscription. You can spend dozens of hours learning in any one of hundreds of instruction sets. As with the tourism sites, the more time you linger at Lynda.com, the better it is for them—and for you. The same thing applies for NASA.gov. You can find thousands of pages of games, videos, instructions, descriptions of planets, moons, stars, galaxies, descriptions all the way to the furthest object ever found. For years I have tracked the adventures and misadventures of the Mars Rovers until they have anthropomorphized and seem to have personalities, Spirit being the unlucky one. Again, quality of content and not navigation is the critical key

Image to come.

In Figure 1.1, it is possible to see three distinct classes of content. These classes will overlap to some extent. For example, one might argue that in one case the content might be expected to sell a product, although it might be quite long—Canon sells its cameras on Amazon.com with content that would run for pages if printed out. In a different case, the content might consist of a table that offers the details of a variety of solar panels in only a few words and a click-to-buy button. The difference is not so much what the content is supposed to do as what the author expects from the reader. If the audience is expected to go to a page and pick up a piece of information or do something, the page is user-centric. If the audience is expected to linger and learn, the page is quality-centric. If the reader is expected to linger in the face of persuasion, the page is persuasion-centric. I will discuss these all in great detail in the second section of this book.

Infinite Variations in Digital Genres Demand Different Approaches to Writing them.

Within each of these three categories are a variety of genres. Many of the genres were exported out of analog media, but many more could never exist outside a digital environment. I suggest the writer should fully understand the genres (exigency, purpose, audience and author expectations, in addition to structure) and write the texts based on that understanding. In other words, I suggest that genre, not medium (even on the Internet), defines writing style and quality.

User-centric texts

In Letting Go, Redish continued a tradition begun by Jacob Nielsen in the early’90s. He did a series of studies on how people read on computers (beginning before the web even existed), and accurately concluded that people tend to scan rather than read. Based on that, he devised a series of important and valuable design rubrics. They have become the standard for web evaluation, because in the beginning they seemed to work. But he and others extended those rubrics to describing the act of writing and evaluating writing properly, creating a one-size-fits-all approach that is proposed to this day by him and the others: “Don’t require users to read long continuous blocks of text,” they say. “Instead, use short paragraphs, subheadings, and bulleted lists.” In a nod to rhetorical traditions, they also sometimes suggest, “Know your audience.”

Section II -- INCORRECT GENRE METAPHORS INTERFERE WITH EVALUATION OF INTERACTIVE MEDIA

I have already suggested that “place” is the first big problem with evaluating websites. Just as we write with metaphorical filters, we also evaluate with metaphorical filters. When we read, we read through filters designed to help us evaluate the text—honesty, quality of writing, quality of mechanics, design, etc. When we visit a place, we use completely different filters—ease of navigation, ambience, design, lighting, odor, background sound, etc. The two metaphorical environments (published documents and remote places) only overlap in the sense that they are designed. But even “design” means something completely different for each environment.

“Visitors” Tour Websites

Suppose you build hotrods. You can go to GMPerformanceParts.com and shop the Performance Parts store to buy your new engine. Or you can go to WhiteHouse.gov and do a White House tour of the West Wing (White House, 2011). The metaphor for the site is “you as a guest of the President in the White House.” Although all of these sites are different, developed for significantly different purposes, the common metaphor for all of them is they are distant places we can visit. You can also go to any number of airlines and purchase tickets. Usability studies guru Jacob Nielsen points out that some early airline sites imitated the ticketing desk, giving purchasers the ability to click on virtual objects on the desk for information. The trend continues unchecked. Virtually all sites selling a product call themselves stores and have shopping carts.

These, of course, are all valuable metaphors for the users of the sites, but because they distract authors and evaluators from the publication process, they are not so valuable for them. The publisher needs to recognize that while it is perfectly appropriate for the reader to see the site as a location, the publisher must publish the site as a document that only looks like a location (a complicated popup book of sorts).

These, of course, are all valuable metaphors for the users of the sites, but because they distract authors and evaluators from the publication process, they are not so valuable for them. The publisher needs to recognize that while it is perfectly appropriate for the reader to see the site as a location, the publisher must publish the site as a document that only looks like a location (a complicated popup book of sorts). Usability Studies as the Tool for Evaluating Websites For the past 10 or so years, the preeminent tool for evaluating website quality has been the usability study. The usability study can be a powerful tool for identifying structural problems in a hyper-document, but it is less effective when applied to some of the other reader experiences—reader cognition and reader preference, for example. Knowing exactly what the reader takes away from a marketing document in terms of attitude, knowledge, and intent can be very valuable. Knowing how much the reader enjoys the site is also valuable. Usability studies have few tools and no vocabulary for evaluating reader preferences or changes in the readers’ cognitive condition. This is not to imply that a usability study is a problem for evaluating websites. It is a perfectly acceptable tool, but we should keep in mind it is only one evaluative tool in a place where many evaluative tools are needed.

Usability really is useful, but there are other important tests

I suggest that the justification for Nielsen’s usability studies for evaluating quality of websites is valid, but only in evaluating quality of navigation, structure, and user-centric content. That said, usability is but one of many tests we should use for evaluating quality on a website. Others include

In short, in my experience, authors who develop interactive media within the metaphor of place (and evaluators who test them) have trouble seeing the sites as documents. They design, construct, evaluate, and discuss their sites as if the sites were structures or remote places. Because authors see the sites as structures, they lose sight of the purposes of the texts and write in a unique style prescribed by usability studies experts, and because evaluators see the sites as structures, they tend to evaluate structures and environments, not texts. On the other hand, authors see some sites as documents with perfectly clear genres (e.g., portfolios and online help) and have no comparable problem with them. Recognizing that all websites are documents containing at least three major writing styles makes it easier to move away from user-centric design toward reader-centric writing.

Conclusion

People who build places tend to be contractors or assemblers. As a consequence, the cognitive metaphor for producing websites tends to involve something akin to “assembly” or “construction” or “manufacture,” or even “putting together Chinese puzzles.”

This structural metaphor has not always been the case, however. Early scholars who examined the new, interactive genres being developed in the late ’80s and early ’90s saw them as textual—documents written with interesting new formats in interesting new ways offering us interesting new possibilities for (and freedoms in) communication. Online help and websites have evolved together pretty much throughout the history of interactive media. Although they are fundamentally the same (particularly with the advent of server-centered online help), no one has a problem seeing a help file as a document. This is largely because help file production has traditionally been under the control of documentation specialists. Major websites, on the other hand, were initially seen as programs developed by programmers; and so, although online help and the WWW are structurally the same thing at the fundamental level and although they evolved together, they took diverging metaphorical paths. One became a place while the other became a document