Finding Information

Information & Interaction Design

Mick McQuaid

University of Texas at Austin

02 Nov 2026

Week ELEVEN

Finding Information

In the future, attention will be our most precious resource.

Search engines

  • Revelation of personal information when you search
  • What do you think? Should you be compensated?
  • Read Page et al. (1999) and Wikipedia on PageRank

Exercise: PageRank by hand

Four pages. A links to B and to C. B links to C. C links to A. D links to C. Nothing links to D.

  1. Form a pair.
  2. Open this Google Doc.
  3. Draw the graph. Give every page a rank of 1/4. Use the simplified rule from the Wikipedia article: a page’s new rank is the sum, over the pages that link to it, of each linking page’s rank divided by its number of outgoing links. Each of you computes two rounds alone, then compare. (6 min)
  4. Rank the four pages. Then guess where the ranks would settle if you kept going. (2 min)
  5. D’s owner wants a higher rank. D has no money and controls only page D. What can D do? What if D could create a hundred more pages? (4 min)
  6. Report to the class: your ranking after two rounds, and D’s best move. (4 min)

Exercise: follow the scent

  1. Form a pair.
  2. Open this Google Doc.
  3. Each person writes one task to be done on a site you can both reach, phrased as a goal and not as a page name: find out how to appeal a parking citation, or find whether this backpack can be returned after thirty days. Do not pick a site you know well. (2 min)
  4. Take turns. The navigator starts at the site’s home page and works toward the goal without using search, thinking aloud. The recorder writes down every link label clicked and, after each click, whether the navigator said the goal felt closer or farther. Stop when the goal is reached or the navigator gives up. (8 min)
  5. Together, find the click where the scent was weakest. Rewrite that one label so it would have carried the scent. (3 min)
  6. Report to the class: how many clicks, where the scent failed, and your rewritten label. (4 min)

Limits of search engines

  • Search by example
  • Search for things for which you have no name
  • Link rot

Discussion: rot in this deck

The notes in this deck cite these pages, most of them written down in 2015. With a neighbor:

  1. Open two of them. Does each still resolve? To the same page, or to a redirect, a paywall, a bot wall, or a 404? (3 min)
  2. For one that fails, look it up on the Wayback Machine. How old is the newest capture? (3 min)
  3. Report to the class: which links should I replace, and with what? (3 min)

PageRank paper, Stanford InfoLab
Tracking the scent of information, APA Monitor
Credibility in social media, Wiley
Gaming Twitter to troll, Ars Technica
Security maturity, Krebs on Security
Filers vs pilers, Forbes (week eight)
Filers vs pilers, The Economist (week eight)

Conversational interfaces

  • Asking an assistant now replaces searching for many questions
  • The answer arrives without the path
  • First mover advantage vs market dominance
  • Corrosive influence of OpenAI on StackOverflow

Exercise: the thing with no name

  1. Form a pair.
  2. Open this Google Doc.
  3. Each person thinks of a thing you can picture but cannot name: a part of an everyday object, a typographic mark, a gesture, a pattern you have seen on a building. Write a one-sentence description for your partner. (2 min)
  4. Swap. Find your partner’s thing with a search engine, text only. Record every query you type and stop after five failures. (5 min)
  5. Find the same thing with a conversational interface. Record the number of turns and whether it was right. Then ask it where its answer came from. (4 min)
  6. Report to the class: which channel found it, which query finally worked, and what the conversational interface said about its sources. If it learned the name from a forum, who writes next year’s forum post? (5 min)

Answers instead of results

  • A results page is a menu of patches; an answer is one patch
  • Scent was carried by labels; now it has to be carried by the prose
  • The user judges a paragraph that has no author

The engine is still underneath

  • Assistants with web access search first, then summarize
  • Ranking decides what the model reads
  • Generative engine optimization is SEO’s second act

Exercise: check the citations

  1. Form a pair.
  2. Open this Google Doc.
  3. Agree on a factual question with a checkable answer that neither of you knows: a date, a number, who did what. Avoid anything from the past week. (2 min)
  4. One of you asks a search engine and records the top three results. The other asks an assistant with web search on and records the answer and every source it cites. (4 min)
  5. Open each cited source. For every sentence in the answer, mark whether a cited page says it, says something near it, or does not say it. (6 min)
  6. Ask the assistant again with web search off. What changed? (2 min)
  7. Report to the class: your counts of supported, near, and unsupported sentences, and whether the cited pages were among the search engine’s top three. (4 min)

Who pays for the answer?

  • Search paid publishers in visits
  • An answer keeps the visit
  • Licensing deals, crawler blocking, and paying per crawl
  • The compensation question, turned around

Where did the questions go?

  • Help demand used to be visible: videos, threads, search logs
  • Now it is a private chat transcript held by the AI vendor
  • The application vendor loses the signal, and the health proxy fades
  • The clickstream became a transcript

Exercise: sketch the results page

  1. Form a pair. Sketchbooks out, pencil.
  2. Each person sketches the page you would get, five years from now, after typing a question. Decide four things: list, answer, or both; where the sources live; how the page shows how sure it is; and what one click costs the user. One page, no erasing. (5 min)
  3. Trade sketchbooks. On your partner’s page, mark everything that is hidden and when it would appear, and any comparison the user has to make between two places. (3 min)
  4. Together, decide which of the two pages a publisher would prefer and which an advertiser would prefer, and whether either is the one the user would prefer. (3 min)
  5. Hold the pages up. Report to the class one design choice and who it serves. (5 min)

Information credibility

Information credibility concepts

  1. Technical knowledge, skill or expertise
  2. Consistency of actions, values, methods, measures, principles, expectations, and outcomes (definition of journalistic integrity from Wikipedia)
  3. Objectivity
  4. Pecuniary Interest
  5. Agreement with ideas and values held by the recipient
  6. Community membership
  7. Precision (variance)
  8. Accuracy (bias)
  9. Falsifiability (using scientific method)

Exercise: rate a source

  1. Form a pair.
  2. Open this Google Doc.
  3. Agree on one factual claim you have both seen online this month and are not sure about. (2 min)
  4. Each person finds one page that supports or refutes it, as fast as you can. Before reading closely, write one sentence saying whether you believe the page and why. (4 min)
  5. Now score your page from 1 to 3 on each of the nine credibility concepts. Leave a blank wherever you have no way to know. (5 min)
  6. Compare your one sentence with your scores. Which concepts drove the snap judgment? Are they the ones Rieh and Hilligoss (2008) found in college students, or the ones found in adults? (3 min)
  7. Report to the class: how many blanks you left, and the one concept that did the most work. (4 min)

Getting help

Exercise: help as a health signal

  1. Form a group of four and pick two browsers. Chrome and Firefox are the easiest because both are heavily customizable, so people need help with themes, user CSS, and extensions.
  2. Open this Google Doc.
  3. Split into two pairs. One pair studies the YouTube help for both browsers; the other studies the official user community for both. Each pair fills in its half of the table on the next slide. (8 min)
  4. Regroup and compare halves. Decide which browser is healthier, and name the one number that most swayed you. (3 min)
  5. Decide whether the vendor could fake that number. If so, which number would you trust instead? (2 min)
  6. Report to the class. (5 min)

What to record

YouTube pair, for the query 〈browser〉 custom CSS

  1. videos uploaded in the past year
  2. age and view count of the top result
  3. whether the maker has a channel devoted to the browser
  4. whether the maker is monetized: sponsor segments, Patreon, merchandise
  5. whether the maker replies to comments
  6. tenor of the comments

Community pair, for the official forum or subreddit

  1. posts in the past seven days
  2. time to first reply on five recent threads
  3. whether a moderator or vendor employee replied
  4. tone of moderation
  5. whether the rules and the moderation log are visible
  6. the three most common topics

A sub-sub-sub-sub-section

Exercise: read a search log

  1. Form a pair.
  2. Open this Google Doc.
  3. The next slide is one week of searches typed into a university department’s own site: the query, how many times it was typed, how many results the site returned, and how often a result was clicked. Group the queries by what the searcher wanted. (4 min)
  4. Find the queries for which the site has no page, and the queries for which it has pages that nobody clicks. Those are different problems. Name each problem. (4 min)
  5. Propose three changes: a page to add, a label to rename, and a synonym to teach the search box. Say which rows of the log justify each change. (4 min)
  6. Report to the class. (4 min)

Analytics

Pictures accompanying analytics articles always feature a magnifying glass

Google analytics

The logo for Google Analytics in focus

Gaming the system

Metadata

data about data

Information Architecture

The Information Architecture Institute offers the following definition of information architecture in their pdf, referenced in the Wikipedia article on Information Architecture.

We define information architecture as

  • The structural design of shared information environments.
  • The art and science of organizing and labeling web sites, intranets, online communities, and software to support usability and findability.
  • An emerging community of practice focused on bringing principles of design and architecture to the digital landscape.

Findability now has a second audience (about half of all Internet traffic): the crawler and the agent, which read metadata, not layout.

Information architecture tools

  • conceptual
  • commercial

baby busy box

Information hiding

Information comparison

Discussion: what your app hides

  1. With a neighbor, open the app each of you used most today. Name three things it hides on its first screen, and say when each one gets exposed. (3 min)
  2. Name one comparison it makes you do across two screens, holding a value in your head. Could the two be side by side? (3 min)
  3. Report to the class. (4 min)

Organizational information maturity

Da real woild jist ain’t like dat!

—Someone in an information-immature
organization

Exercise: the maturity interview

  1. Form a pair.
  2. Open this Google Doc.
  3. Each person writes three questions to ask a prospective employer that would reveal how the organization treats information: its metadata, its tools, and who is allowed to change what. A good question is one where a vague answer is itself an answer. (4 min)
  4. Take turns. The interviewer asks. The employer answers as an organization at level one or two of the five levels in Spruit and Pietzka (2015), without saying so and without lying outright. (5 min)
  5. Swap roles. This time the employer answers as a level four or five organization. (4 min)
  6. Together, rewrite the one question that failed to separate the two. (2 min)
  7. Report to the class your best question and the vague answer it drew. (4 min)

END

References

Page, Lawrence, Sergey Brin, Rajeev Motwani, and Terry Winograd. 1999. The PageRank Citation Ranking: Bringing Order to the Web. No. 422. Stanford InfoLab.
Pirolli, Peter, and Stuart Card. 1999. “Information Foraging.” Psychological Review (Washington, DC) 106 (4): 643–75.
Proença, Diogo, and José Borbinha. 2016. “Maturity Models for Information Systems - a State of the Art.” Procedia Computer Science 100: 1042–49. https://doi.org/10.1016/j.procs.2016.09.279.
Rieh, Soo Young, and Brian Hilligoss. 2008. “College Students’ Credibility Judgments in the Information-Seeking Process.” In Digital Media, Youth, and Credibility, edited by Miriam J. Metzger and Andrew J. Flanagin. The MIT Press.
Spruit, Marco, and Katharina Pietzka. 2015. “MD3M: The Master Data Management Maturity Model.” Computers in Human Behavior 51 (October): 1068–76. https://doi.org/10.1016/j.chb.2014.09.030.
Viviani, Marco, and Gabriella Pasi. 2017. “Credibility in Social Media: Opinions, News, and Health Information—a Survey.” WIREs Data Mining and Knowledge Discovery 7 (5): e1209. https://doi.org/https://doi.org/10.1002/widm.1209.

Colophon

This slideshow was produced using quarto

Fonts are Fira Sans, Fira Sans Light, and Victor Mono Nerd Font

Math is set in Fira Math via MathJax 4