Week 3

Information Arch. Ch. 5 &6:

IA Chapter 5: “Organization systems”

This chapter starts with a quote by Hayden White, “The beginning of all understanding is classification” (Morville & Rosenfeld, 2006, p. 53). We cannot begin to understand the world around us without somehow organizing the information about it. The classification schemes we use reflect our social and political goals and points of view. The labels we use have great significance – is someone a terrorist or a freedom fighter? The answer reflects our beliefs and goals. Information architects try to create labels and organizational systems which will be logical and make sense to the users. The internet is so flexible that the same content can have multiple organization systems whereas the print world is more limited.

Challenges to organization: There are numerous challenges to the task of organizing the world’s information. Librarians used to do this almost exclusively but the internet has created a massive increase in available information and is forcing all of us to become librarians and take on the burden of labeling and organizing the information. Some of the many challenges include:

  • Ambiguity-some words have multiple meanings which can be very confusing in organization systems (ex. Pitch: to toss, sticky substance, what a ship at sea does, etc.). Concepts such as Alternative Healing can be hard to catalog (Philosophy, Religion, Health & Medicine, etc.).
  • Heterogeneity-object or collection of objects made up of unrelated or dissimilar components. This is the opposite of things which are homogeneous (made up of similar things) such as old library card catalog records of books which all provided the same information. Websites are heterogeneous in their offerings (text, audio, visual, articles, databases, etc.).
  • Differences in perspective-labeling and organizing is greatly influenced by the perspectives of the creators. To make the organization useful, we must find out what the users need. The use of multiple navigation pathways can help diverse groups of users find what they are looking for.
  • Internal politics-in any organization there can be many individuals or departments all trying to have the most influence. The choice of how a system is organized and what labels are used has a large impact on how a site, company, or products are perceived. Focusing on the needs of the users and compromise can help greatly.

Organization systems are composed of organization schemes and organization structures. Organization schemes define the shared characteristics of the content and influences the grouping of those items. Organization structures define the types of relationships between the content items and the groups (Morville & Rosenfeld, 2006). There are a couple types of organizational schemes: exact organizational schemes and ambiguous organizational schemes.

Exact organizational schemes: these divide information into sections that are exclusive and well defined.

  • Alphabetical-used for dictionaries and encyclopedias, indices, phonebooks, directories. Bookstores, department stores, libraries all use this type of scheme in some form.
  • Chronological-press release archives, history books, magazine archives all use this scheme. However in some cases users may also want to browse by title or topic as well as date.
  • Geographical-weather, politics, social & economic issues often depend on the location.

Ambiguous organizational schemes: divide information into sections that cannot be given an exact definition. They are subject to ambiguous language and human s. (is a tomato classified as a fruit, vegetable, or berry?). As difficult as these schemes are, they are usually the most important and more useful than the exact schemes. Library catalogs are the perfect example. The books are cataloged by title, author, and subject. Title and author are exact organizational schemes, but subject is an ambiguous organizational scheme. Studies have shown that more users are using the subject or ambiguous schemes than the exact ones. It is only when we are doing a search for a known item that we usually use the exact schemes (we know the author or the title). When we have a general information need on a topic we must use ambiguous subject searching.

  • Topic-subject or topic organization is one of the most useful but difficult approaches. Yellow pages, academic courses and departments, and newspapers use this scheme.
  • Task-content organized by processes, functions, etc. This works when there is a limited number of high priority tasks. An example is word processing programs which organize under categories : edit, insert, format, etc. Ebay has categories for buying or selling.
  • Audience-this type of organization is helpful where there are two or more defined user groups. College websites usually have categories or sections for current students, potential students, faculty and staff. Banks often divide their information by personal or business.
  • Metaphor-this is used to relate new topics/ concepts to familiar ones. The “desktop” of the computer has “files” and “folders” as a way to integrate computer concepts with people who have never used one before.
  • Hybrids-a pure organizational scheme helps the user form a mental model and it is not always wise to combine them. However, the main page or portals of many large sites successfully provide navigation in different schemes. Ebay has both tasks and topic on its main page. College websites often use audience and topic successfully on the main page.

Organization structures: The structure determines how a user can move throughout the system. The 3 major types are hierarchy, database-oriented, and hypertext.

  • Hierarchy-(top-down approach) a well designed hierarchy or taxonomy is the foundation of most good information architectures. It is a common structure we have used to organize our world (family trees, biological division in kingdoms, classes, etc.). It is best to lean toward a broad but shallow structure to allow future growth since users don’t want to click through a lot to get what they want. A medium balance of both depth and breadth is the best.
  • Database model-(bottom-up approach) metadata links IA to the database scheme. Relational databases help find items based on fields. A rolodex is done by last name. A company can have its employees searchable by name, department, campus, function, etc. Metadata, controlled vocabularies and the database structure allow for: automatically generated alphabetical indexes, see also links, field searching, filtering and sorting, etc.
  • Hypertext-non linear method of grouping information. The system has items which will be linked and the links between the items. These links can be complex and create user confusion as they can easily become lost. It is rarely used as the main organizational structure but is often a component that complements the hierarchy or database models.

Social classification:

This term encompasses tagging, collaborative categorization, mob indexing, ethno-classification, and folksonomy. It is the opposite of controlled vocabularies. With social classification users add tags to classify or identify the item. The websites flickr and del.icio.us are two places where the social tagging phenomenon has taken off with some success. There are those who would argue that with tagging capabilities, controlled vocabularies and metadata are no longer needed and cannot be used in places where tagging can due to costs. However, folksonomy doesn’t out perfom traditional organization approaches. There are loads of findability issues involved in tagging systems. With the help of information architects there may be able to be a successful integration of traditional organization with the new approaches.

Right now libraries are experimenting with catalog programs which allow traditional cataloging and controlled vocabularies but which also allow users to add tags and for those tags to be an added searchable field. It is still too early in the process to see if the tags will add significant benefits but some may find it useful.

Cohesive systems

All information retrieval systems work best when dealing with homogeneous content. It can be helpful to break down content into chunks of similar items and then work on what organizational system will work best for each group. You still need to keep in mind the big picture of a uniform system. Exact schemes work best for known item searching and ambiguous schemes are best for browsing and when the information need is not clear. If possible, use both types. Because language is ambiguous and individuals have different perspectives, it helps to provide multiple ways to reach the same information.

IA Chapter 6: “Labeling Systems”

Labeling is extremely important. These words represent categories of information that cannot be shown all at once. It is through the use of labels or signs that users find what is available on a website, and how to get to it. In a face-to-face situation, a person can ask or give non-verbal clues that they don’t understand a sign or direction and receive further help. On the Internet, there are no non-verbal clues. Either users will figure out what they want or they will leave the site. There are 2 types of labels seen on the web: textual and iconic. The textual are “text” based and iconic are visual symbols. Often textual labels are used along with iconic for added clarification. There are several types of textual labels:

  • Contextual links-hyper links to groups or sections of information on another page.
  • Headings-describe the content that follows much as print headings do.
  • Navigations system choices-these represent options in the navigation system.
  • Index terms-keywords, tags, or subject headings that represent content for the purpose of browsing or searching.

Sometimes labels filling multiple roles. A heading can be a link, or a contextual link can lead to a page with the same word as the heading. An index term could also link to a page with the term as a heading, etc.

Contextual links: you want them to make sense to your users. If you use a label that doesn’t make sense to most of the users, it doesn’t help anyone. If on a webpage you see a link labeled Shakespeare, and expect to find a biography of Shakespeare or a like to another website about him you will be frustrated to be taken to a travel site about Shakespeare, NM.

Headings: they should clearly describe what follows. Often this is hierarchical.

Navigation labels: these should be consistent because they are often repeated over and over again throughout a system. If they change from page to page, users will only be frustrated.

Index terms: since these represent the meaning of the content, they can provide more precise searching and browsing.

Guidelines for labels: We must think of users and try to avoid ambiguous labels. Narrowing the scope helps in reducing ambiguity-focus on a more defined audience, fewer subject domains, a narrower business context with clearer goals-all these help.

We must be consistent in our labeling especially in the following areas:

  • Style-consistent punctuation, etc.
  • Presentation-consistent color, typeface, font size, etc.
  • Syntax-don’t mix verb based, noun based, and question based labels, choose a single syntactical system.
  • Granularity-labels should be relatively equal in specificity.
  • Comprehensiveness-be as comprehensive as possible (a clothing retailer who lists pants, shoes and ties but has no label for shirts is not comprehensive and users will be confused).
  • Audience-don’t mix terms for two widely different audiences, (lymphoma vs. tummy ache). If necessary put the terms for different audiences in separate areas.

Some sources for labels include: your own site, similar or competitive sites, controlled vocabularies and thesauri, and folksonomies or free tagging. If you already have a website, look at what labels are there. Take a look at how similar sites use labels and what terms they use. A comparison of 4 different pc companies showed that they all used essentially the same terms on their websites (home/home office, business, government, education, etc.). Clearly those terms are important categories in that industry. There are websites such as Taxonomy Warehouse (http://taxonomywarehouse.com), and ThesauriOnline (http://www.asindexing.org/site/thesonet.shtml), which provide indexes of terms for various fields. Some sites allow for user tagging (flickr and del.icio.us) and those tags can provide some insight into what labels many people use for some items. Another source is the server logs from a current website. Search analytics can determine what terms are most being searched for. It can also determine what terms are yielding no results. This information can help with providing terms for labels.

Websites reviewed:

Rosenfeld Media/Search Analytics http://rosenfeldmedia.com/books/searchanalytics/

Wikipedia/Contextual Design http://en.wikipedia.org/wiki/Contextual_design

Rosenfeld Media/Search Analytics http://rosenfeldmedia.com/books/searchanalytics/:

This is Rosenfeld’s website which is actually his company which is a publishing company devoted to publishing books about user experience (UX). He wants to focus on short practical books rather than long books trying to make a case for good design. There are many different sub-categories that all fall into the UX area: user experience, experience design, etc are all topics that Rosenfeld deals with.

Under publications, Rosenfeld (along with Wiggins) has a book in progress called Search Analytics. Search analytics is a way of measuring what users of a website are searching for. It is an excellent tool for designing a site or redesigning it – improving its searchability and findability. Websites (servers) have logs that capture key strokes so one can see what search terms users are typing in. If there are terms users are searching for that are not finding results, it may be time to re-examine what might match that and add the new term. This is one tool that is used to understand the needs of the users and to improve a website’s navigation and search options. An example is given in the blog on search analytics in the NY Times. Apparently the website Spafinder.com used search analytics and noticed that lots of users were searching for nudist spas. They created a separate category now for “nudist spa vacations” and have found that searches for that type of trip average 720 a month (Rosenfeld, 2008).

Wikipedia/Contextual Design http://en.wikipedia.org/wiki/Contextual_design:

Contextual design is another tool (like search analytics) used to find out what users need in a website (how they use it, what they use it for, etc.). Unlike search analytics, contextual design uses ethnographic techniques for gathering data. Instead of simply looking at logs of key strokes without any feedback, contextual design involves observing users and talking with them about what they mean by certain terms or what they were looking for but didn’t find.

Week 3 references:

Morville, P., & Rosenfeld, L. (2006). Information architecture for the world wide web (3rd ed.). Sebastopol, CA: O’Reilly.

Rosenfeld, L. (2008, April 28). Site search analytics in the NY Times. In Rosenfeld Media. Retrieved September 8, 2008, from Rosenfeld Media Web site: http://www.rosenfeldmedia.com/books/searchanalytics/blog/site_search_analytics_in_the_n/