July 26th, 2004


SearchTools Report: Search User Interface and User Experience

Search Engine User Interface and User Experience

There are two parts to a search engine's user experience: the user interface (design of the search forms and results pages) and the functionality (how well it matches and sorts pages). When you install a search engine, you should consider both aspects, design the interface carefully and test all aspects of the usability.

SearchTools User Experience Recommendations

While search engine pages share many interface elements with other parts of web site design, there are certain principals that you should keep in mind:

  • Put a simple, reasonably long search field on every page of the site.

  • Use simple words to explain the process: remove all jargon and technical terms, and make sure that any icons have labels.

  • Avoid inventing a new interface, which will confuse users: take the best of the formats of the large public search engines

  • Make the search forms and results pages fit into the overall design of the web site: they should use the same colors, fonts and so on.

  • Include site names and navigation links into results pages, so users can see the context and structure of the site.

  • Set up a special page to be displayed when the search does not find any matches in the index (see No-Matches Page Guidelines)

  • Avoid surprises: explain all automated search features, such as stemming, phonetic matching, thesaurus lookups and stopwords (see Glossary).

Articles, Books and Presentations

see also Overview Articles

  • Site Search can be flattened by Usability Demystifying Usability, as of January 29, 2004 by Frank Spillers
    Aspects of search solutions that are not considered during the enterprise search procurement process: search interface and search results.  The article discusses the need to be able to customize the results and simplify the interface, regardless of how the back end search works.

  • Designing a Search That Works for Users WQusability at wqusability.com, as of February 11, 2003 by Whitney Quesenbery
    Describes a Model of the user search process that goes beyond the user interface, and describes the overall user interaction with the search process.  Quesenbery identifies five approaches to information:  browse, find, query, structured, guided.  The purpose of the model of how people search is to design more "effective and usable search."  The user interaction is described as six steps: formulates a question, create the query, review the results, evaluate the likelihood of success, select and examine an item, and evaluate success. 
  • Employee Directory Search: Resolving Conflicting Usability Guidelines Alertbox at Useit.com, February 24 2002 by Jakob Nielsen
    Describes the seemingly contradictory requirements for a single search box on an Intranet main page vs. the value of a specialized search for employee telephone numbers and other directory information. Recommends providing two search fields, labeling very clearly and testing the interface design to reduce confusion.

  • Looking, Finding, Searching … How Users Do It STC Telephone Seminar, November 19 2002 by Whitney Quesenbery
    Presentation based on a study of users and health information sites found that about one-third of the users were adamant about needing to use search to find information. Study discovered that they had a hard time putting questions into words, worrying about language and technical terms, even on second tries. They tended towards broad short queries with some complete phrases, had trouble with spelling and typing, were confused about operators, and had a hard time even finding the search field. When results pages have too much navigation at the top, search box, multiple headers, links that don't look like links, users can't even find the results listing. When confronted with a result list, they read it as a page, assume the most important items are first, rely on titles as "headlines" and look for additional information. No one could tell the researchers what the stars meant (more relevant). Very few ever tried to refine a search when a results list did not provide relevant articles, these users gave up on the whole site. Quesenbery recommends testing search and results page designs, blending search and browse, and leveraging metadata such as page title, description, and general document category. She finds faceted metadata search a good solution for many sites, while visualization techniques can answer high-level questions at a glance.

  • Intranet Usability: The Trillion-Dollar Question Alertbox at Useit.com, November 11 2002 by Jakob Nielsen
    Results of an international usability study by the NNGroup on intranets finds that many are wasting employee time by failing to provide usable intranets. They found that "search usability accounted for an estimated 43% of the difference in employee productivity between intranets with high and low usability." They recommend that intranets make sure that the main search engine indexes all pages, shows results in relevance order with manual recommendations at the top, encourage useful page titles and descriptions. The NNGroup finds that tasks requiring 27 hours annually on a usable intranet could take as long as 196 hours on a less-usable one.

  • On-the-Job Research: How Usable are Corporate Research Intranets? Alison J. Head & Associates, April 2002 (non-SLA members: $135 print, $185 PDF)
    Report on usability testing of seven representative research intranets - those providing textual content for all employees, rather than services such ordering, billing or collaboration or a departmental site. Finds that most intranets are underutilized because they are badly designed and organized, and difficult to use. Common information needs include employee telephone numbers, offices and email addresses, current company news, media coverage of their company and competitive intelligence research. However, only 44% of participants, managers, administrative assistants, and researchers alike, were able to complete the test research tasks in these areas. Search engines had significant problems, such as requiring complex query operators and failing to provide context or filters for large search results. Study includes many details on testing and intranet content, design, information architecture and search issues. Alison J. Head & Associates, April 2002 (non-SLA members: $135 print, $185 PDF).

  • Why search is not a technology problem: Case Study, BBCi Search (follow link to 3.4 MB PowerPoint file) ASIST IA Summit February 2002 by Matt Jones
    Describes how the British Broadcasting Corporation added a search engine to their web site, offering access to both all the information in the BBC site and web-wide information. The group created some "use models" to show examples of information needs and find ways to add context to search results. They chose to create a taxonomy based on the most frequent searches, and test the process over the course of several months, starting with paper prototypes. They found significant user sensitivity to wording and layout, and to functions that break with their expectations. Showing manual recommendations was successful once they integrated with the other results, and people loved the search zones tabs once they found them (after several visits). Recommends early testing and use models to learn how people really use search engines.

  • Search Relevancy Display CHI-WEB Discussion, January 20, 2002
    Several answers to the question of whether users find relevancy ratings helpful. Consensus was that they are often confusing, especially when expressed as specific numbers. Some feel that providing the general range is useful. One relevance percentage system is described in detail.

  • Paging vs. Scrolling: Looking for the Best Way to Present Search Results Usability News, January 2002 by Michael Bernard, Ryan Baker, & Marissa Fernandez
    Results of a usability test comparing displays of search results of 10, 50 and 100 links. Study found that locating a specific link with in the results list is fastest in the fifty-link condition, that users rated the 10 and 50 link condition easier to find information, several of them felt the 100 link condition had too many choices, most seem to think that the 10 and 50 link conditions look more professional. Overall, the participants prefer middle range, although this may be due to the study setup which defined some required search results after the first 10. In any case, recommends shorter and midrange lists over very long (100 link) search results.

  • In Defense of Search Semantic Studios December 7, 2001 by Peter Morville
    Responding to Jared Spool of User Interface Engineering, who claims his research shows that search is a problem, rather than a solution. Morville counters that, while taxonomies are useful, they can't keep up with growing sites and linguistic ambiguities. Search engines are provide a valuable form of content access using a wide variety of mechanisms including parameters, manual recommendations, synonyms and context. Recommends investing time and resources to make search rise to its potential.

  • Getting Them to What They Want User Interface Engineering Report, October 2001 by Erik Ojakaar and Jared M. Spool, $24.99
    Provides interface recommendations for sites to help users find their desired content. Recommends categories and taxonomies over search because it can be more efficient and provide better control. The find that significantly more users continue browsing after locating their target by following category links than by searching. Recommends basing taxonomies and category labels on popular terms from search logs.

  • Search Interface Standards Bobulate.com Analysis, November 16, 2001 by Liz Danico
    A taxonomy of search forms for web sites. Divides them up into standard search forms (search field, button, perhaps a link to advanced search); surfacing: those with search zones or filters based on site taxonomy; and qualifying: with other filters such as date, or local vs. global. Passive search interfaces are just links to search forms, rather than live fields.

  • Can Navigational Assistance Improve Search Experience? A User Study FirstMonday, September 2001 by Mazlita Mat-Hassan and Mark Levene
    Evaluation of the NavZone search engine, which shows where a user has been and the context of the current page in the site. The test compared it with the ht://Dig search engine ("Compass") installed on the site and the Google service indexing the site with 24 people and five fairly complex questions. Results show that NavZone was more successful, with significant reduction in the number of clicks required and testers getting more correct answers. While testers, especially novices, began by being intimidated by the interface, they became used to it and appreciated it after some training.

  • Web Usability - Search Capabilities section of Designing Web Usability (1999) by Jakob Nielsen, excerpted on the Web Designers Virtual Library, June 2001
  • Search: Visible and Simple Useit.com Alertbox, May 13, 2001 by Jakob Nielsen

    Both these pieces provide simple rules for search interfaces, with results of research showing that users are impatient and quick to give up searching when they encounter problems, type very short words, and rarely look beyond the first page of search results.

    Note: If you buy the book from Amazon using this link, SearchTools.com will get an affiliate fee.

  • Developing Schemas for the Location of Common Web Objects Usability News, Winter 2001 by Michael Bernard
    A formal usability test on 346 people about where they expected to find various standard elements such as home page links, internal links and advertisements. Figure 4 shows that both novice and experienced web users expect to find the search field in the upper center of the page, followed in preference by the bottom center and upper right corner.

  • Search Survey Results: "Sex" Popular on the Web, Many People Inefficient at Reaching Their Online Destinations (content removed, see summary) Alexa Research, February 14, 2001
    To no one's surprise, this comprehensive study of 10 web search engines over two years finds that the single most common search is for the word "sex" (0.3289%), as well as "porn", "nudes" and "xxx". More interestingly, web users get confused by the search field, because it is quite common to find them typing in URLs, such as "hotmail", "yahoo" and "ebay".

  • Roper Study on "Web Rage" while Searching (removed from the Roper site: see the ZDNet article on the study). Roper Starch Surveys, December 18, 2000
    Commissioned by the search engine WebTop, this reports the results of a study of people doing general web searches saying:

    On average, it takes 12 minutes of searching the Web for specific information before Internet users get frustrated. Almost one in five Internet users get frustrated if their search takes them up to five minutes to complete. About half of Internet users have the patience to search the Web for longer than 15 minutes.

    Danny Sullivan of SearchEngineWatch has termed this "search rage" and a total of 71% of people reported being frustrated at some time during searching, and over half get frustrated by irrelevant information in search results. However, more than 75% usually or always find what they're looking for.

  • Next Generation Web Search: Setting Our Sites IEEE Data Engineering Bulletin, Special issue on Next Generation Web Search, September 2000 by Marti Hearst.
    Thoughtful discussion of search and results interface features in existing and theoretical search engines. Covers the importance of "task" in searching on web sites, metadata and categories (using Epicurious and Yahoo! as examples), and proposes a way to apply that to more specific data such as medical papers.

  • Fitts's UI Law Applied to the Web Microsoft Developer Network, May/June 2000 by Scott Berkun
    Discusses Fitts's Law (covering how hard or easy it is to select an element on a computer screen using a mouse). One example is of search results with clickable links to the various pages in the results listing. The author points out that with several very small items in close proximity, it will be very difficult for users to select a specific one, and recommends a larger link to the next page only.

  • Searching on the Web, Two Types of Expertise SIGIR 99, August 1999 by Christoph Helscher & Gerhard Strube
    Describes searching tests conducted by Internet experts and novices, and by domain experts and novices. Includes notes on features used by the testers.

  • Cognitive Strategies in Web Searching Proceedings of the Human Factors & the Web conference, June 3, 1999 by Raquel Navarro-Prieto, Mike Scaife, and Yvonne Rogers.
    Results of a small study on web-searching behavior, including cognitive strategies and interaction with the systems.

  • User Interfaces and Visualization in Modern Information Retrieval, chapter written by UC Berkeley professor Marti Hearst, provides a valuable academic study of interfaces for information retrieval and searching, including graphical overviews and visualization. Get the book "Modern Information Retrieval" from Amazon or FatBrain.com and we'll get an affiliate fee.

  • Web Site Searching and the User Experience Bay Area Computer-Human Interface group, January 1999 by Avi Rappoport
    Slides from a presentation on how to learn from the experience of other search engines, including large public engines, to make a comfortable and rewarding experience for local search users.

  • Clarifying Search: A User-Interface Framework for Text Searches D-Lib Magazine, January 1997 by Ben Shneiderman, Don Byrd and W. Bruce Croft
    Excellent article about designing search interfaces, with a "framework" of four steps: formulation (defining the searchable data and search options), action (issuing the search), review of results (results formats) and refinement (adding relevance feedback). Also describes how principles of user interface apply to search interfaces and includes a case study of redesigning the search interface to the Congressional Record.

Web Sites and Listings

  • User Centered Information Retrieval - Marcel van Mackelenbergh's site full of short, informational articles about information retrieval, with a focus on metadata from a human factors perspecive.  There is an excellent article on faceted interfaces, as well as some interesting tips on how to measure information retrieval from a usability perspective and improving search interfaces.

  • Search Usability Workshop -Starting with material presented in a workshop at CHI 20003, this group of researchers will be publishing design patterns and other tools for designing usable search engines interfaces. Papers from the workshop are on the site now, and there will be more information soon. Avi Rappoport of SearchTools is coordinating a bibliography of search usability papers: to be included, send email to subib @ searchtools.com

  • Research-Based Web Design & Usability Guidelines - Large amounts of helpful information on all aspects of web site and page design, culled from the best available research.

  • Software Usability Research Laboratory, Usability News - Results of research on web usability issues such as link location, fonts, layout and shopping decisions.

  • Usable Web - a wonderful collection of links and articles on web usability issues, including searching, collected by Keith Instone.

  • UseIt - Jakob Nielsen's helpful articles about designing usable web sites. See especially the article Search and You May Find.

  • SIGCHI - ACM's (Association for Computer Machinery) Special Interest Group on Computer-Human Interaction
  • What People Search For - links to sites showing live or aggregate search term information for large search services. Extremely useful for testing hypotheses about numbers of search terms, natural language, use of Boolean and other issues.


Information Foraging Theory

Analyzes how people look for information based on concepts from biology and anthropology, mainly food-foraging strategies. This includes making choices in entering search terms and looking at results based on the interface and content presented to the searcher, as well as their expertise in searching. They have identified information scent as the perceived value of a document based on the information presented in a list, such as search results, and discuss ways to analyze how people choose strategies to select the best match from a set of items.