Marianne Sweeny's "SEO and IA: The Makings of A Beautiful Friendship"

Presentation Transcript

Second Life, Info Architecture island, 08-03-07

INTRODUCTION

Erich Hesse: Hi. – and welcome to this, the third presentation in our IA Summit Redux.

Erich Hesse: In First Life, my name is Eric Reiss and I’m president of the Information Architecture Institute.

Erich Hesse: Here in Second Life, I’m a completely inept avatar, so please forgive me if I fall off the stage.

Erich Hesse: I’d like to thank Stacy Surla from the Board of the Institute for arranging this series.

Erich Hesse: In fact, Stacy is the reason we have this island at all.

Erich Hesse: Many thanks also to Naomi Malone for managing the series.

Erich Hesse: Thank you so much for your efforts.

Erich Hesse: Moving on to tonight’s “main attraction”…

Erich Hesse: We are incredibly honored that Marianne Sweeny could join us this evening.

Erich Hesse: Marianne’s presentation on search engine optimization was one of the highlights of the recent

Erich Hesse: Summit.

Erich Hesse: And remember, we were in Las Vegas…

Erich Hesse: So Marianne was competing with Tom Jones, Celine Dion, and Tony Braxton.

Erich Hesse: For those of you who have already seen this presentation live…

Erich Hesse: …you know that Marianne is incredibly culturally literate.

Erich Hesse: Who else could work Saint Teresa into a presentation description?

Erich Hesse: These days, Marianne is an information architect and search specialist at Ascentium.

Erich Hesse: Ascentium is an agile consulting company with a slew of sexy interactive solutions.

Erich Hesse: In an earlier life – er…what’s that in relation to Second Life…?

Erich Hesse: In an earlier life, Marianne co-founded the Microsoft Information Architects.

Erich Hesse: This is a 300 member community dedicated to bringing the philosophy and best practices of IA to

Erich Hesse: the company.

Erich Hesse: She has also studied Library Science at the University of Washington…

Erich Hesse: …and is one of the world’s leading experts on search engine optimization.

Erich Hesse: So, without further ado, I’m delighted to present to you, Ms. Marianne Sweeny.

PRESENTATION

Searcher Shepherd: LOL...many thanks Eric. Toni and I remain friends in spite of that competition.

Searcher Shepherd: As for the world, it is in sad shape with me at the search optimization [SEO] helm.

Searcher Shepherd: However, like the chicken in the road trying to hold the sky up with its bird legs, i do what

Searcher Shepherd: i can

Searcher Shepherd: and my mission is to make sure that IA always has a place at the table when information

Searcher Shepherd: finding is being discussed

Searcher Shepherd: The title for this presentation comes from the movie "Casablaca" when the hero and his

Searcher Shepherd: former arch rival wal into the night together with a renewed understanding of their

Searcher Shepherd: relationship in light of a common enemy.

Searcher Shepherd: This is how I see IA, Interaction Design, User Experience Design and search technology.

Searcher Shepherd: New friends in light of a common enemy that is finding what we are looking for in a vast

Searcher Shepherd: sea of information.

Searcher Shepherd: Up to now, we have worked on a parallel track, the technology developing around us

Searcher Shepherd: while we create paths through the information space.

Searcher Shepherd: I believe that we can and must work with the technology to protect our ability to contribute

Searcher Shepherd: to the means by which our clients and patrons find information.

Searcher Shepherd: And I believe that it will be a beautiful friendship.

Searcher Shepherd: I would like to take your questions as they come up in this presentation

Searcher Shepherd: i have my chat window open

Searcher Shepherd: please preface your question or comment with MARIANNE: so that I will notice it as a

Searcher Shepherd: question and i'll answer it directly

Searcher Shepherd: and so to begin

Searcher Shepherd: First,I want to thank Stacy Narayan, Naomi Malone and the IA Institute for setting this up

Searcher Shepherd: and for working even harder to guide us in our ability to be here today.

Searcher Shepherd: I am Searcher Shepherd, have been an IA for 8 years and focused on search for the last

Searcher Shepherd: 3 years.

Searcher Shepherd: I am here today to talk about organic search engine optimization that is constructing and

Searcher Shepherd: amending sites to map to the functionality of search engine technology so that sites can

Searcher Shepherd: appear higher in the results sest for client search terms and phrases.

Searcher Shepherd: Our clients are using search rather than the elaborate pathways that we build. This

Searcher Shepherd: requires that we make room in our thinking for designs that acknowledge the search

Searcher Shepherd: tools.

Searcher Shepherd: Good SEO does not have to interfere or dictate IA – it is a composite of actions and best

Searcher Shepherd: practices that can ensure maximum visibility with peaceful coexistence.

Searcher Shepherd: Search puts wayfinding directly in the hands of the client - that is why they like it so much.

Searcher Shepherd: It is immediate. Revelatory by showing a lot of intepretations at once.Efficient, for known

Searcher Shepherd: item searches it cannot be beat.

Searcher Shepherd: Users can easily hop back and forth from the search results page to the site until they

Searcher Shepherd: find what they are looking for or iterate their search.

Searcher Shepherd: Ask.com saves them the effort of a click by showing a preview of the site.

Searcher Shepherd: Worst of all, search is indulgent. Google made it all look so easy now our clients expect

Searcher Shepherd: efficient, quick and plentiful results all of the time.

Searcher Shepherd: Close enough is often good enough. And, if it isn't, they turn to the most common

Searcher Shepherd: information finding tool...the person in the cubicle next to them.

Searcher Shepherd: And for this we can put the blame squarely on Google. Page and Brin creted PageRank

Searcher Shepherd: to combat spam. It seems that many were doing a poor job of honestly describing their

Searcher Shepherd: content. So, the system would figure out what is relevant to the client's query by counting

Searcher Shepherd: links. Because, links are the result of human thought plus human action...votes so-to-

Searcher Shepherd: speak.

Searcher Shepherd: However, PageRank has nothing to do with the client query. It is a pre-query value

Searcher Shepherd: calculated at indexing and recalculated periodically.

Searcher Shepherd: We also discovered that it is not flawless. A bunch of folks who use the same link and

Searcher Shepherd: target can make the #1 result for the search term "miserable failure" the autobiography

Searcher Shepherd: of American president George W Bush.

Searcher Shepherd: Here it is, the famous, to some infamous PageRank algorithm. This is its most stripped

Searcher Shepherd: down state. Rumor has it that the algorithm now has in excess of 27 components. We’ll

Searcher Shepherd: look at some of these extensions in a few moments.

Searcher Shepherd: When Google appears in 1998, it is the underdog to search giants like Alta Vista and

Searcher Shepherd: Yahoo! Its simplified relevance model with the foundation of human mediation through

Searcher Shepherd: linking [each link was at that time the product of direct human endeavor and so viewed as

Searcher Shepherd: a “vote” for the page or site relevance and information merit]

Searcher Shepherd: I read an interesting story about Larry Page and his PageRank -not named after him-

Searcher Shepherd: algorithm

Searcher Shepherd: he was evidently quite taken aback that other search technologies leveraged this

Searcher Shepherd: important finding after he and Brin published their initial paper.

Searcher Shepherd: This accounts for Google's "close to the chest" research and development.

Searcher Shepherd: One thing Google has given us is a base standard for relevance.

Searcher Shepherd: This has come at a price. One that we'll realize in a few minutes...depending on my

Searcher Shepherd: typing accuracy.

Searcher Shepherd: I believe that it is mandatory in all presentations to have a "dot oh" slide.

Searcher Shepherd: Here is mine, a gentle observation on what search has claimed as its own.

Searcher Shepherd: It harnesses the collective intelligence through linking and optimization.

Searcher Shepherd: It capitalizes on the architecture of participation through development of new an

Searcher Shepherd: innovative tools. Jimmy Wales is developing an open source search community to

Searcher Shepherd: liberate search from advertising dollars.

Searcher Shepherd: It takes data from a variety of sources and mashes it together to answer questions before

Searcher Shepherd: they come up. Google maps will show you what restaurants are near your movie.

Searcher Shepherd: And it liberates search from the desktop and brings it to mobile devices

Searcher Shepherd: Search 2.0 is about returning control of search to the user. PeopleRank, social search,

Searcher Shepherd: P2P, AJAX search applications are centered on the user interacting with the system in a

Searcher Shepherd: way that extends beyond the solitary box on the big white page

Searcher Shepherd: Next step is “meaning”

Searcher Shepherd: What I find most fascinating is that all of the roads seem to be leading to a form of human

Searcher Shepherd: intervention in serach results presentation. Jimmy Wales of Wikipedia fame is developing

Searcher Shepherd: a peer-to-peer search solution that sees inviduals ranking search results and t his

Searcher Shepherd: information feeding display position moving forward. Now is an excellent time for us to

Searcher Shepherd: participate

Searcher Shepherd: Librarians wake up screaming when they dream of the emerging artificial intelligence of

Searcher Shepherd: search.

Searcher Shepherd: As the hardware gets cheaper [Google’s index is reputed to live on cheap PC servers

Searcher Shepherd: running open source software…many, many cheap PCs are needed to host a 9 billion

Searcher Shepherd: page inverted index and cache of all pages], the software becomes more sophisticated

Searcher Shepherd: with asynchronous as well as synchronous query calculations taking place.

Searcher Shepherd: The purpose of this consumption of processing power is to formulate enough brute

Searcher Shepherd: strength to actually derive contextual meaning through rapid and multiple comparisons.

Searcher Shepherd: Machines do not have a sense of “place”, emotion, or awareness, all of which feed our

Searcher Shepherd: contextualizing of language into personal meaning. Windows Live search does not know

Searcher Shepherd: the difference between dog as it applies to the canine, the bounty hunter in Hawaii or a

Searcher Shepherd: bad investment.

Searcher Shepherd: Search technology uses computational methods to apply a growing ontology based on a

Searcher Shepherd: vector space comparison.

Searcher Shepherd: i see a hand raised? is that a question?

Searcher Shepherd: Guess not and no worries.

Searcher Shepherd: Here is a brief look at what drives the technology.

Searcher Shepherd: I love the Hilltopo Algorithm and am delighted that technology has finally advanced to a

Searcher Shepherd: point where it can come a reality instead of a theory. It was developed by two computer

Searcher Shepherd: scientists at University of Toronto. Jon Kleinbert at Cornell has a similar theory with his

Searcher Shepherd: Hypertext Induced Topic Selection

Searcher Shepherd: Both of these are performed on a small subset of the corpus that best represents nature

Searcher Shepherd: of the whole Web.

Searcher Shepherd: Results are ranked according to the number of non-affiliated “experts” point to it – i.e. not

Searcher Shepherd: in the same site or directory Affiliation is transitive [if A=B and B=C then A=C]

Searcher Shepherd: A non-affiliated expert is a page that is not in the same domain or related domain.

Searcher Shepherd: From all of this Authority sites are derived with HUBs being pages that contain navigation

Searcher Shepherd: links to authorities on a particualr subject.

Searcher Shepherd: A beauty of Hilltop is that unlike PageRank, it is query-specific and reinforces the

Searcher Shepherd: relationship between the authority and the user’s query. You don’t have to be big or have

Searcher Shepherd: a thousand links from auto parts sites to be an “authority."

Searcher Shepherd: Once designated as an authority, links from these pages are afforded more weight.

Searcher Shepherd: Google’s Florida update is rumored to be an extensive deployment of Hilltop and saw a

Searcher Shepherd: lot of sites that had developed links from extraneous sites fall far in the results sets.

Searcher Shepherd: Hilltop is a solid step forward in the direction of "relationship"; the relationship between

Searcher Shepherd: pages that link to one another.

Searcher Shepherd: Those list-o-link pages are just taking up space on the Web now.

Searcher Shepherd: Topic Sensitive PageRank takes the PageRank vectors and combines them with a topic-

Searcher Shepherd: based Web. IT measures importance within the confine of a particular topic. Semantics

Searcher Shepherd: influencing PageRank.

Searcher Shepherd: So, the good news is that the technology is coming around to a classification schema for

Searcher Shepherd: Web content and search results ranking. The not-so-good news is that we, as information

Searcher Shepherd: architects/interaction designers/experience designers and the like have the expertise in

Searcher Shepherd: human behavior torward information and information spaces and are nowhere to be

Searcher Shepherd: found in these developments.

Searcher Shepherd: I love America. Inventors must fully explain their inventions in order to get a patent and

Searcher Shepherd: lock up ownership. The United States Patent Office is a great resource on where search

Searcher Shepherd: technology is going.

Searcher Shepherd: There is an thesaurus-like search engine that has combined its index with human

Searcher Shepherd: mediated catagories. A query is matched to the catagories.

Searcher Shepherd: One search company wants to take the trouble out of selecting a search engine by

Searcher Shepherd: brokering queries and directing them to the engine with the best results.

Searcher Shepherd: Both MSN and Google are working on behavior search applications. These desktop apps

Searcher Shepherd: collect information on how the client segments information on the hard drive, what they

Searcher Shepherd: search for, what they click on, how long they stay on a page, what they have in their

Searcher Shepherd: bookmarks, how those bookmarks are arranged, how they iterate their searches, etc.

Searcher Shepherd: This information is then synthesized and used to sort results.

Searcher Shepherd: My favorite, predictive search where the system notes the user’s behavior with search,

Searcher Shepherd: results, time on page, etc and then produces content that would seem to be the next

Searcher Shepherd: logical choice were it present on the page.

Searcher Shepherd: All of these focus on the tool, few focus on the client. This is where we com in.

Searcher Shepherd: There are a lot of options available to IA with regard to the development of search

Searcher Shepherd: technology

Searcher Shepherd: We can continue to be disengaged entirely.

Searcher Shepherd: We can cooperate by developing best practices based on our knowledge of user needs

Searcher Shepherd: and our abilities to structure information spaces

Searcher Shepherd: We can initiate changes in the direction of serach technology to facilitate understanding

Searcher Shepherd: of spatial relationships through the modeling of information

Searcher Shepherd: I cannot emphasize enough how much I learned from the McCollough book.

Searcher Shepherd: His thoughts on translating building architect theories on space influcing perception to

Searcher Shepherd: information architecture are terrific

Searcher Shepherd: back to the script

Searcher Shepherd: Navigation blindness – don’t make me think, just show me to the search box. Considering

Searcher Shepherd: the sometimes hundreds of sites folks visit in a given session, it is understandable that

Searcher Shepherd: they get tired of figuring out the navigation schemas for each one. Especially when it is so

Searcher Shepherd: easy to go to the search box or pogo back to the search results.

Searcher Shepherd: Navigation fatigue – every site is different and sometimes that gets tiring.

Searcher Shepherd: Page paradigm – Mark Hurst gave us the Page Paradigm and the observation that the

Searcher Shepherd: locus of attention for the user is at the center of the page. The navigation that we so

Searcher Shepherd: painstakingly develop is confined to the periphery.

Searcher Shepherd: Transitional volatility – This comes from David Danielson's master's thesis at Stanford.

Searcher Shepherd: The key takeaway from this stellar research for the purposes of this "talk" is that subtle

Searcher Shepherd: changes in the navigation, i.e. transition to a navigation that is contextual to the directory

Searcher Shepherd: or area, nudged the client to notice it and use it.

Searcher Shepherd: A visionary site manager at TechNet, Pam Salle, used it in 2005 to great success in the

Searcher Shepherd: desktop deployment section. Client use of navigation and satisfaction went way up.

Searcher Shepherd: I am not suggesting here that we do away with the traditional tools that we've been using

Searcher Shepherd: so far. many of them have and will continue to withstand the test of time.

Searcher Shepherd: I am suggesting that "findability" is not almost fully in the hands of search technology and

Searcher Shepherd: that we need to acknowledge that to ourselves

Searcher Shepherd: so that we can bring the locus of attention back to the client and the information need and

Searcher Shepherd: the information that will service that need

Searcher Shepherd: and away from the speed of processing and the complex math required to build predictive

Searcher Shepherd: engines that might understand