by Peter Morville, Louis Rosenfeld
I, Michael Parker, own this book and took these notes to further my own learning. If you enjoy these notes, please purchase the book!
- pg 5: Structuring mean identifying information "atoms," organizing means grouping them into categories, and labeling is naming those categories.
- pg 5: An information architecture must balance the needs of users with the goals of the business.
- pg 10: Experience design is an umbrella term for information architecture, usability engineering, graphic design, and interaction design.
- pg 12: Information architecture can make money, save money, improve employee or customer satisfaction, or make the world a better place.
- pg 21: If the lone information architect on a site, work with outsiders to help understand the needs of the user better.
- pg 22: Ideally, the IA is responsible for only the site's architecture, not its other aspects, so IA is not ignored for more visible and tangible stuff.
- pg 24: Effective IA relies on understanding the intersection of context, content, and users.
- pg 26: For context, your IA provides a tangible snapshot of your organization's mission, vision, values, strategy, and culture.
- pg 27: Your IA must reflect the ownership, format, structure, metadata, volume, and dynamicism of your content.
- pg 28: Your IA must account for user behavior, like who uses your site, how they use it, and what information they want from it.
- pg 30: An IA must not just understand the user's needs, but his information-seeking behaviors.
- pg 34: A user needs a particular item (known-item seeking), a few good things and iterates (exploratory seeking), or everything (exhaustive research).
- pg 35: Searching, browsing, and asking are information-seeking behaviors, and good sites integrate them, or allow transitioning between them easily.
- pg 37: The "pearl-growing" model assumes you start with a document you need, and allows finding similar ones, such as through tagging.
- pg 38: To learn about the information needs of users, practice search analytics, or reviewing the common queries in the search logs.
- pg 44: Top-down information architecture is where a site has been designed to anticipate users' information needs and is structured to address them.
- pg 47: Bottom-up information architecture is where the implicit IA in content enables answers to users' information needs to "rise up."
- pg 48: Bottom-up IA is important because users are likely to bypass top-down architecture, landing at a page using deep links or search results.
- pg 51: In content, headings, embedded links, metadata, lists, and identifiers (e.g. a logo) are all IA components.
- pg 55: Ambiguity leads to difficulty in organizing information -- e.g. ambiguity of what a label means, or ambiguity of what label to apply to something.
- pg 57: The different levels of granularity of documents, and differing characteristics of formats, makes organization of heterogeneous content hard.
- pg 59: Exact organization schemes are unambiguous, which allow known-item searching, which are: alphabetical, chronological, and geographical.
- pg 61: Unambiguous organization schemes are subjective, but are useful because you don't always know what you're looking for.
- pg 65: Audience-oriented schemes break a site into smaller, audience-specific mini sites; if open, users can go between mini-sites.
- pg 67: Be careful of blending together different organization schemes (topic, task, audience, metaphor), which clouds a user's mental model.
- pg 70: Be aware of, but not bound by, the idea that hierarchical categories in an organizational structure are mutually exclusive.
- pg 73: New sites should favor the broad-and-shallow hierarchy, because adding items to secondary levels is less problematic.
- pg 74: Metadata allows applying the power of databases to the heterogeneous, unstructured environments of web sites.
- pg 78: Free tagging creates a bottom-up categorical structure, called a folksonomy.
- pg 83: Labels should speak the same language as a site's users, and educate them on new concepts or quickly identify familiar ones.
- pg 87: When creating contextual links (hyperlinks), ask what information the user would expect to find if clicked to ensure they are representational.
- pg 92: Navigation systems have a small number of options and appear throughout the site, so their labels demand consistent application.
- pg 94: Navigational labels can be augmented by brief descriptions, or scope notes, when introduced on the main page.
- pg 95: Index terms can be generated by metadata, such as the META tag in the HTML, and can provide a view across different information silos.
- pg 98: Only use iconic labels if your site has patient, loyal uses willing to learn your visual language.
- pg 99: Labels should have comprehensive scope, use audience terminology, and be consistent in style, presentation, syntax, and granularity.
- pg 102: Professionals with subject-specific backgrounds may have created controlled vocabularies and thesauri you can use as labels.
- pg 108: Free-listing is selecting an item and having subjects brainstorm terms to describe it, which are potential labels if common across items.
- pg 109: Analyze search logs or folksonomic tags to indirectly generate labels that account for users' jargon, acronyms, and tone.
- pg 116: The global, local, and contextual navigation systems are the embedded navigation systems, wrapped around and infused with site content.
- pg 119: The navigation system should be consistent, and should always show the user's location within the organizational hierarchy.
- pg 123: On the main page, the global navigation can be expanded or supplemented with other navigation options.
- pg 126: Hypertext links inlined in the content can be replaced with external links, which occupy a specific area of the page, that are easier to see.
- pg 132: Supplemental navigation, like sitemaps, alphabetical indices, and search, are for when the taxonomy and embedded navigation fail.
- pg 135: When designing an index, use term rotation, which uses synonyms or rotates words in a phrase so it appears more than once in the index.
- pg 139: With personalization, we guess what content the user wants, and with customization, the user tells us what content he wants.
- pg 140: Effective personalization can be hampered by the user's desire for privacy, or lack of time to teach the system.
- pg 141: Since corporate intranets have a captive audience of repeat visitors, customization has a better chance of being used than on public sites.
- pg 148: If your site has a substantial amount of content, users will likely expect your site to have search.
- pg 151: Defining search zones, or repositories of homogenous content, can keep the results from your search engine focused and relevant.
- pg 154: Common ways to define search zones are by audience, by topic, or by chronology.
- pg 158: Indexing numerous content components (title, summary, keywords, URL, body) for retrieval allows added flexibility in designing results.
- pg 159: High recall returns numerous results of varying relevance; high precision returns just a few high quality results.
- pg 163: When presenting results, display less information for users who know what they're looking for, and more for unsure users.
- pg 167: Given too many results, allow a user to easily revise a search by populating the search box on the results page with the original query.
- pg 169: Different ways to measure relevance work best for different content formats, so be careful when ranking heterogeneous content.
- pg 171: Showing relevance scores can aggravate a sense of ignorance in the user if the relevant results are not first, so it's best to omit them.
- pg 177: If the user is typically looking through hundreds of results, offer them a way to favorite the documents they like for easy recall later.
- pg 178: If the content of your site is very dynamic, allow the user to save search queries so they can be re-executed later, even automatically.
- pg 181: Users won't spend time crafting well-constructed queries, so simply give them a box for queries and a "search" button.
- pg 183: Put the search box in the global navigation, so always available, and well away from other boxes (like ones for passwords).
- pg 188: Integrate searching and browsing, where searching may yield a navigational section, and in that section search is restricted by zone.
- pg 190: Given too many results, allow the user to revise the search from the results page, and search within their current set of results.
- pg 191: Given no results, allow the user to revise the search, give them search tips, and offer a means of browsing.
- pg 194: Metadata allows us to focus on how to describe a document, instead of pondering where it belongs in the taxonomy.
- pg 195: A synonym ring is a set of words, which may include misspellings, that are equivalent for the purposes of retrieval.
- pg 197: Synonym rings can increase recall, as more documents are returned, but reduce precision.
- pg 198: Authority files are synonym rings in which one term has been defined as the preferred term or acceptable value.
- pg 204: A thesaurus models equivalence relationships (synonyms), hierarchical relationships, and associative relationships (related terms).
- pg 209: A classic thesaurus is used when indexing and searching, mapping variant terms to preferred terms in both stages.
- pg 211: An indexing thesaurus is only used when indexing, but can still allow creation of browsable indexes of preferred terms.
- pg 212: A searching thesaurus is only used when searching, and its development and maintenance costs are independent of content volume.
- pg 216: A hierarchical relationship emerges when one has the characteristics of (generic relationship), is a part of (whole-part relationship), or is an instance of something else.
- pg 217: The ANSI/ZISO Z239.19 standard for constructing English thesauri discusses many relationships for building related terms.
- pg 218: The standard encourages for preferred terms: nouns, in the plural if countable, and abbreviations and acronyms if the popular use.
- pg 221: Using compound terms in categories usually leads to polyhierarchy, which is challenging to represent in the navigational context.
- pg 222: Faceted classification uses multiple taxonomies that focus on different dimensions of content.
- pg 225: Guided navigation takes faceted classification further, allowing users to refine or narrow searches based on different facets of the results.
- pg 233: A balanced approach to research focuses on context, content, and users, and how they all interact.
- pg 234: Context focuses on goals, budgets, schedules, technology infrastructure, human resources, and corporate culture.
- pg 236: Use an introductory presentation to sell IA to authors, developers, designers, marketing, and managers and get buy-in.
- pg 239: Perform a gap-analysis, finding the disconnect between business goals, user needs, and practical limitations of the technology infrastructure.
- pg 240: A heuristic evaluation is an expert critique of a site against principles and design guidelines, learning what to keep of the existing site.
- pg 243: Content analysis extracts data or metadata from a representative content sample, and helps find patterns and relationships between items.
- pg 245: Competitive benchmarking compares different sites, while before-and-after benchmarking compares different versions of the same site.
- pg 247: Analytics tools can reveal effects of navigational changes to your site, but not why a user visits.
- pg 248: Search log analysis can accurately reveal what users are looking for, in what volume.
- pg 253: Focus groups don't work well for information architectures because people don't have the understanding or language to articulate about them.
- pg 256: Open card sorts used for discovery allow the user to write his own labels; closed card sorts used for validation use prescribed labels.
- pg 256: To explore alternate navigation paths, allow the user to copy cards, thereby cross-listing them in multiple categories.
- pg 262: Do research so that the site won't have to be redesigned later at great cost, and to find out what users actually want.
- pg 265: An IA strategy is a high-level conceptual framework for structuring and organizing a site, and is typically blurred with research.
- pg 270: Communicate your ideas with peers using metaphors, stories, use cases, conceptual diagrams, wireframes, reports, and presentations.
- pg 272: Wireframes are not fully-designed prototypes, but expose how users interact with the IA within the broader context of a web page.
- pg 274: Metaphors can leverage the familiarity of organization, of tasks that can be performed, or understanding of images, icons, and colors.
- pg 275: Write scenarios, or stories, to show how people with different needs and behaviors might navigate your site.
- pg 278: Blueprints show relationships between pages and content components, while wireframes show the contents and links of major pages.