Helping Us Think: The Role of Abstract, Conceptual Models in Strategic Information Architecture

By Andrea Ames | Fellow, and Alyson Riley | Member

Successful strategic information architecture is like alchemy—an almost magical blend of art and science. From a scientific perspective, information architecture involves disciplined rigor in the way that we:

• Follow repeatable processes to achieve measurable results
• Test and refine theories until we get them right
• Define variables and constants to arrive at solutions

And much like art, information architecture is about creating meaning through a deep understanding of the human experience—in particular, the science of human cognition. It involves a passionate pursuit of simplicity and elegance in the face of complexity and chaos.

For those facing complexity and chaos at the enterprise level, modeling the information experience in the abstract is an invaluable method for helping everyone to think: employers or clients (“the business”), users, technical communicators, and information architects (IAs). Models ensure that technical communicators (IAs, technical writers and editors, user experience professionals, and so on) can create or contribute to concrete information architectures that work across a broad diversity of audiences, user goals, products, and business requirements while maintaining consistency and brand recognition. Models also ensure that we capture and codify the best information relationships, delivery mechanisms, presentation formats, and content deliverables to address our clients’ business challenges and to meet the expectations of our users within their domains, leveraging their mental models, problem-solving paradigms, and information-processing approaches.

Abstract models of the information experience are key to an enterprise-level strategy for delivering consistent, scalable, effective information architectures. This kind of modeling blends art and science in the best ways:

• It yields artifacts based on tested theories of cognition and information processing.
• It leverages the scientific method by building on validated theories that produce consistent results.
• It borrows from the art-world frameworks for asking the right questions, discovering patterns, and tolerating the ambiguity necessary when dealing with human intent, behavior, and cognition.
• It encourages IAs to discover solutions by applying concepts in a systematic manner, nuanced by a vision for user needs—not by following rules and recipes by rote.

The paragraphs that follow define information models in general and introduce a handful of models for very specific purposes to illustrate how these models drive value in the business.

Models defined

For the most part, adult humans think in terms of models. We observe the world around us, conceptualize, and abstract into models to think and solve problems. To explore this idea further, consider the example of a model home. If you’ve ever looked at a model home while house-hunting, you know that the model is an example, a pattern that shows an ideal state, a depiction of how things might be if life were perfect, a representation of what’s possible if price were no object. In reality, the house you end up purchasing might differ significantly from the model—it might be smaller, include different countertops or hardware, or have a slightly different floor plan. The fundamental purpose, form, and structure remain the same—kitchens, bathrooms, bedrooms, etc.—while details vary based on home-buyer needs and capabilities. If you were to compare a collection of “real homes” built by the same developer, they would likely conform to the general pattern of the model home—you probably have felt this effect when driving through a neighborhood where all the houses look alike. While this kind of uniformity is boring in neighborhoods, it’s ideal for user experiences: consistency is predictability, which allows for recognizable brands and highly cohesive, usable information solutions across a broad diversity of design contexts (such as the environment that we’re in when working at the enterprise level).

Information models are a specific type of abstract, conceptual model. Using IBM as a case study, let’s take a look at a handful of “model homes” from the world of information. When used in concert, IBM’s Use Model, Content Model, and Access Model address all facets of an information experience and work together to form a comprehensive Information Model.

Use Model defined

An abstract Use Model consists of a complete, coherent collection of patterns that define ideal interactions between users and information—that is, the information they need, why they need it, what they’re doing when they need it, and how they will use it in that context. “Why?” is a critical question for good Use Model development. In a product documentation context, an IA can assume that the user’s primary purpose is not to use information—rather, the primary purpose is to achieve some other goal. In fact, in the world of “jobs” that the user performs, using information is “Job #3.” Consider the following:

• Job #1: the user’s goal—for example, to balance her checkbook
• Job #2: using tools—things like paper, pencil, calculator, or personal finance software—to accomplish job #1
• Job #3: using information to figure out how to accomplish job #2 to reach the goal of job #1
• Job #4: learning how to use and then using the delivery mechanisms in which we deliver our information (as painful as it is to admit, we often create job #4)

The Use Model standardizes and validates assumptions about users and the information they need to achieve their goals by helping the IA specify with precision the user’s principle objective and the information necessary for achieving that objective. A Use Model also helps an IA identify information needs based on user motives and skills. As a result, Use Models should describe an abstraction of how readers use information in typical contexts, such as:

• To do (accomplish a task, achieve a goal, etc.)
• To study (learn concepts)
• To locate (retrieve content by browsing or searching)

When an IA applies the abstract Use Model to their product or process, the resulting deliverable includes:

• Concrete user descriptions
• Product-, process-, or system-use scenarios that include the tasks that are the user’s primary focus (job #1)
• Descriptions of how the user’s primary focus areas are addressed by the product or process (job #2)
• Information scenarios that describe the user’s interactions with the information (job #3) necessary to ensure that the user can successfully complete her primary tasks

In short, the abstract Use Model, once applied, yields a specific use model describing a comprehensive set of use requirements for the information.

Content Model defined

An abstract Content Model defines the standard building blocks of content, from the atomic, most granular level to larger “deliverables”—those concrete, stand-alone objects that humans recognize and use. It includes the subject and structure (presentation) of the content, provides a content taxonomy, and relates metadata to appropriate content chunks. When an information architect applies the Content Model, the result is a set of concrete units of information and a structure that combines those building blocks in a way that reflects concrete user goals and tasks.

Product installation information, for example, is a unit of information or a deliverable (using our parlance above) that users typically recognize. The atomic information building blocks—a variety of tasks and concepts—come together in a relatively predictable way, ordered around the task of installing the product. At a high level, that predictable structure (order and organization) typically looks like this:

1. Prerequisites
2. Installation planning tasks and concepts
3. Installation tasks
4. Post-installation configuration tasks

The abstract Content Model enables defining patterns of information—building blocks, order, and organization—that meet customer expectation, removing from the communicator the design burden and eliminating from the user the cognitive processing required to parse information inconsistent with their expectations.

Access Model defined

An abstract Access Model defines a vision for how your target users will find your information. IAs use the Access Model to describe the organization, structure, and interrelationships of chunks of information and full deliverables, as well as the strategy for using a variety of access methods across an information space—methods such as search, browsable navigation, and so on. The Access Model provides a big-picture vision for overall navigation strategy, as well as patterns for specific access methods for individual collections of content or deliverables. When an IA applies the Access Model to the design of their specific information and user contexts, the resulting deliverable is a concrete navigation and access scheme, including the requirements for what information is delivered where and by what delivery mechanism.

Information Model defined

An abstract Information Model defines the relationships between and unites the other models into a coherent whole. In other words, the Information Model helps architects “put it all together,” standardizing the way we assemble the components of the other models in order to create an information experience for a specific kind of product or system. Like the other models, the Information Model is abstract; when you apply it to a product- or system-specific context, the result is a concrete information architecture.

Why models

While reading the model descriptions, you might be thinking, “Wow! That looks like a lot of work!” It’s true; it can be lot of work. For a large enterprise like IBM, it’s worth the effort. For smaller organizations, a careful cost-benefit analysis will help you determine if the modeling process has tremendous value for you, as well.

Models help businesses think. Models are good for business, because they adapt well and yield results that matter in the marketplace. At the enterprise level, information architecture must scale to handle increased complexity; diversity of users, products, or systems; flux in the market; and rigorous business processes. An enterprise typically can’t afford to update templates every time tactics change or strategy evolves, and no cookie-cutter recipe or library of templates can cover every situation. Models force a business to identify, prioritize, and design for those user interactions that are critical to achieving business objectives. Technology and visual identity may evolve over time, but those user interactions essential to business success do not.

While models can drive sea change, they do not change with trends. The abstract quality of the models keeps the architecture above the fray of trend and branding, aligning all aspects of the information experience, and ensuring that the focus stays on successful user interactions while allowing for interesting change at the presentation level. Because of this focus, modeling an information experience in the abstract helps IAs define and deliver an ideal, consistent, and strongly branded information architecture. Consistency, of course, leads to brand recognition, which in turn contributes to brand loyalty—something the business world definitely understands. In addition, user-centered information architecture ensures high-value and highly findable content; findable, high-value content generates highly ranked search results and priceless social capital, which contributes to the process by which mindshare becomes interest, then consideration, and ultimately a revenue opportunity.

Models help users think. Abstract models based on cognitive science and user-validated theories provide a robust, user-focused framework on which to build the specific architectures that our teams require to implement successful information experiences. Abstract modeling helps us to deliver intuitive information experiences that users don’t have to think about—allowing them the mental space that they need to maintain focus on the things that they care about—their primary goals and tasks.

As a profession, we follow in the footsteps of Steve Krug and his first law of usability—“Don’t make me think!”—when we deliver excellent information experiences that remove significant user-cognitive load from the equation. That means that we don’t make the user think about our information (Job #4)! Abstract models that are based on sound cognitive principles and user-validated theories work in our users’ cognitive favor. As Krug says, our “job is to get rid of the question marks” so that it’s obvious, without thinking:

• What to do next
• Where to go next
• Whether the information answers the question
• How to find more or different information that will answer the question

Note that “what to do next” and “where to go next” can refer to the product or process that you are documenting, or to the information system used to display your content. Work to make job #2 and job #3 more self-evident for the user; in so doing, you will eliminate job #4 and you might preclude the need for a lot, if not all, of job #3.

Models help communicators think. Abstract models that define access, delivery, content, and presentation remove much of the “guess work” with which our colleagues struggle—especially those teams without experienced IAs. By enabling a framework for thought, technical writers, editors, and IAs (particularly those who are new or have little formal training) can more easily and reliably produce high-quality and consistent concrete architectures.

Experience is still the best teacher. By providing abstract models—including education, examples, and best practices for applying those models in the context of a specific product or system—we build information architecture capability among the other technical communicators on our teams. Removing the need to study the details and theory behind the models enables teams to think less about the general, underlying principles and more about understanding the specifics of their users and how to best apply the models to deliver successful information experiences to those users. In this way, they learn by experience to design and deliver successful architectures while maintaining consistency and integrity within an overall corporate brand, or look and feel.

Models help IAs think. Abstract models encourage an IA to keep user needs and business strategy in the forefront of her thinking, rather than confining an information experience to the boundaries of a template. Applying the models allows IAs the flexibility needed when working across an enterprise. Models enable a trust in the IA to determine where the application of the model can be adapted to support specific business or user requirements while maintaining the integrity of the overarching experience. Models keep the focus on consistent, high-quality outcomes that map to business and user success: results rather than rules, as rules often require deviation and the IA’s time and focus to ensure that those deviations are appropriate and necessary for user success. Models provide the IA a method to ensure that the fundamental purpose, form, and structure of the information experience remain the same, while allowing freedom in the details as dictated by user need.

Models: supporting strategy and reducing cognitive load

Modeling an information experience in the abstract can help IAs develop a strategy for delivering scalable, effective information architecture. While wireframes and templates are important, abstract models yield concrete, enterprise-level information architectures that will stand the test of time. Building models and applying variables to get high-quality, consistent results—that’s science. How you adapt the model based on user need and business strategy—that’s art. And that’s the art of the IA—that magic place where user needs and business requirements get synthesized in unique ways that templates cannot fully articulate or replicate. Models help IAs capture the benefit of a scientific process while leaving room for the art that happens when you innovate on the user’s behalf. Salvador Dali captured this concept elegantly when he said, “I am a carnivorous fish swimming in two waters: the cold water of art and the hot water of science.” To our fellow fish, we’ll add the words of Dori from Pixar’s Finding Nemo: “Just keep swimming.” Give abstract modeling a try—for those of you swimming in turbulent waters, it just may help keep you afloat.

Want to learn more about IBM’s abstract information experience models and how we create, validate, and apply them?? Visit Intercom online to see more and to contribute to the conversation!

Andrea Ames is an information experience strategist/architect at IBM, and she designed, coordinates, and teaches for University of California, Santa Cruz in Silicon Valley’s Technical Writing and Communication program. Her mantras are “I don’t write doc; I solve user and business problems”; “installation is not a user goal”; and “think more, write less!” She is Fellow and past president (2004–2005) of STC and a distinguished engineer of ACM (the first technical communicator to achieve this distinction). She has published two award-winning technical books and more than 50 papers and articles and speaks regularly at conferences and professional meetings around the world. Follow her on Twitter (@aames) and check out her blog (http://thinkmorewriteless.wordpress.com/).

Alyson Riley’s mother swears that her first word was actually a complete sentence, and she began her career as an information architect shortly thereafter by developing various organization schemes for her plastic dinosaurs. She now works as a senior information architect and strategist on IBM’s corporate strategy team and has over 17 years experience in technical communication. Alyson serves as the corporate lead for IBM’s information architecture council, drives IBM’s corporate-level efforts to define the next generation of user experiences with technical content, and consults with IBM content teams worldwide to develop effective information strategies. Alyson has BS and MS degrees in scientific and technical communication from the University of Minnesota.