more outtakes from the 2nd edition of Blueprints, most likely destined for the long anticipated yet never begun book on search. Please Note: this has not been edited in any way
The speed of search causes other problems for teams new to it. Normally, to figure out how well your design is working, you might do a bit of usability testing. But that doesn't work well for search.
Traditional research mars results
If you wanted to see how a hummingbirds fly, you'd have a hard time. They move so fast they become a blur. But if you asked them to slow down, they wouldn't stay aloft and you couldn't learn anything. .
When you are watching people try to complete a task in usability testing, you often use "think aloud" protocol, in which the user speaks their thoughts as they work on their task. But search is so fast, that if the user is forced to think, they slow down too much and behave unnaturally, thus creating false results.
With search, the physical actions are few... look, type, click. What you really want to know about is hidden, even from the person searching. Search requires different research methods
A few methods I've found useful:
Eyetracking: With other types of products, eyetracking is often a useless novelty. The data rarely tells you that much you can do anything about, since users consume multiple pages as they accomplish tasks. But because of Search's nature... a single page, with endless but constrained variations... eyetracking can teach you quite a bit about the mental process as revealed in the way a searcher scans a page.
For example, on a results page, the searcher always glances at the query box first. When asked why, the answers vary... "I wanted to see if I misspelled it", " I wanted to see if the search engine got it right" ... searchers often don't really know their own minds. The important thing was they look there. That allows us to return to our rule, "every pixel has a job to do" and decide that yes, you need to show the query in the search box on the results page, and no, you do not need a label like "You searched for X."
And that hotspot on the far right in the screenshot I've included? That's the user actually seeing an ad that uses the keyword they searched for. Proving ads are seen when they are relevant. Imagine that.
Bucket testing: Also known as A-B testing. In this scenario you release different variations of the same design to a portion of your users, to see what is the most effective. While this is invaluable in understanding what works and what doesn't in search, it's also quite difficult to get trustworthy results.
 |
| A good example of a A-B test candidate: you don't change anything else, but the color of the links. |
To get good data from bucket testing, you need
This means you have to get smart about what you are testing. Restraint is important, and trusting yourself and not putting every single design argument up for testing is required.
Field Research: This is most useful when you have some quirky users. Such as people who are searching in difficult environments, people who have a very particular idea about how search works, or people for whom is part of a larger, complex task.
Field research is a simple concept. You go to people's homes, or work environments, or wherever they use your product, and watch them using it in context of their real, daily tasks.
This information helps you decide what features you need, what expectations you must meet, and perhaps even who's design you should copy.
However, to do field research effectively, it's best to rely on a trained researcher. If you do it yourself, you can easy disrupt the natural flow, or ask leading questions, again delivering bad data. If you can't afford to hire a pro, at least buy Observing the User Experience: A Practitioner's Guide to User Research
To make it clear, our problem is not to find all near duplications. We just with to find near duplications in articles we serve, but it must be very fast. We might return 100 articles in one set. Comparing all of them to each other will take about 10K comparisons.
Some of conventional methods to solve near duplications I know of are using shingling and matching term frequency vectors. Shingling is great and most accurate, but is expensive. You can't take all the articles and compare them, not mentioning keeping all in memory. Creating the shingles takes time and in large documents there may be many of them. Vectors might be less accurate for these purposes and have similar caveats.
These observations are just that, observations. Not rules or laws. If it looks more like a rule than a observation, then add the rest of the sentence "or it won't get used" to turn it back into a observation. It's from an longer essay on search I wrote in order to publish it, I ended up folding most of it into the second edition of Blueprints for the Web. This section didn't feel "IA" enough to be part of that book, it may still go into a book on search someday. it's based on eight years of working with search in many contexts from web search to enterprise search. Some of these may prove not to be true-- I originally called this "ten facts on search interfaces" but who knows. They could be proven wrong. Somehow I doubt it. Search is remarkably rigid.
1. A search box needs to look like a search box. If you use css to make it look like something other than a form element, replace the word "search" with "go" or some other space saving/brand "enhancing" effort, the number of searches will go down.
2. Don't prepopulate the search box with text such as "Type search here." The number of searches will go down (because it interferes with people's ability to recognize it as a form element), and really, most people know how to use a search box.
3. Bold and Blue: keywords should be bold, and should always be in the result. Links should be blue. Good old fashioned 0000cc.
4. Showing URLs: This makes sense if you need to distinguish location. If you are searching one domain, it's not necessary. But the difference between shopping.domain.com and www.domain.com can help a scanning searcher.
Do not link the url; it creates unnecessary page weight and makes the page harder to scan.
5. Numbers. Yahoo numbers their results, Google doesn't (otherwise how would you know whose was figure 1 and whose was figure 2?). Numbers have been shown to be qualitatively better, but not quantitatively. This means users say they like numbers, but bucket testing doesn't show any measurable advantage. So you can ask yourself, is it worth the pixels?
6. Snippets vs. descriptions: A snippet is a bit of text scraped from the website that includes the keyword. A description is human written, and describes the website. What's better? It varies widely depending on the quality of your scraping technology, the quality of your staff writers, but if the description does not have the keyword in it, the snippet is better.
7. Graphics almost always make things worse.
Let's say you are considering adding an image to a best bet about the Beatles. Pick the image on the top, and you improve click-through, pick the image on the bottom and you damage it. Even though you may have been a fan of the Beatles in their latter years, the first image is far more iconic. But how many bands have easy, iconic images? How many concepts?
The same holds true for product shots; unless you perfectly match the searcher's mental model of what they are looking for, they will dismiss the entire result. It gets worse when you add big arrows, animations, or illustrations. In other words, the harder you try to draw attention, the more you become invisible. Graphics often get misread as advertising.
Conversely, small icons can occasionally help rather than hinder differentiation between best bets and results, providing honesty that algorithm purists often desire.
8. Balance your page. You can make sponsored results look more like "true" results, and their click-through will go up. But the true results click-through will go down, and your visitors will trust your search engine less and visit less. A short term pop in revenue may cost you later in retention; make design choices carefully.
9. The first result gets the most clicks. The second, the second most. And so on, right? Strangely enough there is a slight increase again in the last one or two results. This is true in search, it's true in navigation bars as well. If you want people to get good results (i.e. more relevant) it's a good idea to keep the default number of results on a page down to about ten. If you want more click through on a navigation bar item, make it first, second, third or last.
10. The bottom of the page is often forgotten. But it's the safety net for those searchers who haven't seen what they want. And searchers are more likely to abandon their search (to try another service, sometimes) than click to the second page. Encourage them to hang around with query helpers, another search box, spelling corrections, or even another search on your site (if you have one.)
is the innocent get hurt. But we all knew that, didn't we. Does it make it okay? Let's look at this adsense scenario.
Despite having some delicious metadata to pull from,
meta name="keywords" content="design, conference, event, workshop, information architecture, interaction design, meetup, cocktail hour, salon"
meta name="description" content="The place to list design-related events, from conferences to cocktail hours."
title Create Engaging Applications for Facebook - Boxes and Arrows Event Calendar
Google ignores that -- because it might be gamed-- in favor of what.. the URL? Even the title of the page has better information on the topic of the site than the name.
If they have to ignore what I, the site owner, say the site is about, why not turn to their own giant search engine monster. Surely it knows what queries Boxes and Arrows is relating to... and it's not archery. Nor packing companies.
Worse of all, you know who the real beneficiaries of this system are? The domain name squatters. Only they, the holders of "love poetry .com" and "cheap software .com" can guarantee that the adsense that shows up on their sites will match the searches well enough to get high click-through. This is so effective that one domain squatter I know (yes, gentle people, I have friends in high and very low places) had to give up his only genuine website because a squatter page made him so much more money.
I know relevancy is a sufficiently complex science to resemble black magic, but surely the wizards of Mountain View can do better.
remember altavista, and when you first started using google, you felt guilty? for abandoning altavista?
1st gen search engines: search engines "crawl" links to pages, they make a copy in something called a index, they find pages you are looking through, originally via term frequency. this was too spammable, because control was in the hands of the webmaster.
2nd gen search engines: use factors off the page that wemaster can't easily influence
Reading Astonishing! Spock, a New People-Search Engine, Thinks You're a Pedophile I can only imagine that Spock is so dumb (or potentially cleaver in a P.T. Barnum sort of way) they'll be successful if they can ever provide a valuable service. Or they'll drive a ton of traffic while their results are still crap, and those folks will never return. Hard to know, but with epically entertaining guffs like this, I'd consider a return visit.
I got an invite to Spock, and they have a ways to go yet. The invitation and the site itself gives you no sense of who they are, but they feel very comfortable asking you for passwords to your linked in, myspace, friendster and email accounts. And then they do it again. and guess what, having put in my linked in password twice, they asked for it again.
Okay guys, first I don't know you from adam, but thought I'd give you a try, and now you are proving insufficiently competent to keep that trust. Finally I run a search on my latest favorite author, Steven Gary Blank. No results. Search 101: there must always be results. At least roll over to Google guys, otherwise I need never. ever. never ever try you again.
Yeah it's private beta, but if you want to bet with my passwords, my network of friends and my trust, you aren't getting a second try. Go back and fix the messaging. Hire a writer.
They when you suck, at least you can show a cute cat and EXPLAIN what the heck is going on so I feel like giving you a second chance.
Results for "innovator." Pretty funny.
San Francisco, CA, USA - Google Maps catches a man with... er... lost keys?
Google's typical indifference to human beingss sense of privacy being upturned plays out in street view, but it's not all down side:
There is one shot from downtown San Jose, however, that should be preserved even if the images around it change. Taken as the Immersive VW passed the corner of Santa Clara and Market, it shows Cornelius Van Der Vies and his beloved dog Boo-Boo at their permanent spot. Cornelius was homeless by choice, spent his nights in a van and his days on the corner, but with his pleasant and polite demeanor had been sort of adopted by the office workers who passed every day. Cornelius died in late April after a fight with another homeless man who’d been yelling at Boo-Boo, and crowds came to his public memorial. Like many, I imagine, when I walk downtown, I’ll always picture the two of them on that corner, and it warms my heart knowing that image lives on.
Often I get frustrated when people say "search" as if what google does and what they do for their small eccomerce site should be the same thing. Other players are realizing that there are many search chores, and more than one way to skin a cat: FT.com / Home UK - How to free the slaves from the webmasters
"Mr Horowitz, however, has no doubt that social search is the future: "You don't need social search to find the population of London - you do to find a plumber," he claims. Now that could be a search too far, even for Yahoo."
Search has got to be the most masculine of the Internet business spaces, a fact revealed by the endless preening over size and speed more reminiscent of the auto industry than the traditionally female owned information retrieval space (madam librarian!). This is much evidenced in the latest show of peacock feathers by Google and Yahoo...
How Many Pages in Google? Take a Guess - New York Times
"In response, Yahoo issued a statement saying: "We congratulate Google on removing the index size number from its home page and recognizing that it is a meaningless number. As we've said in the past, what matters is that consumers find what they are looking for, and we invite Google users to compare their results to Yahoo Search.""
Sound familar ladies?
"Executives at Google, the rapidly growing online-search company that promises to "organize the world's information," announced Monday the latest step in their expansion effort: a far-reaching plan to destroy all the information it is unable to index."
Reading Google Anything, So Long as It's Not Google - New York Times I was impressed... for a company that values open access to information above all other virtues this seems out of character.
"Last month, Elinor Mills, a writer for CNET News, a technology news Web site, set out to explore the power of search engines to penetrate the personal realm: she gave herself 30 minutes to see how much she could unearth about Mr. Schmidt by using his company's own service. The resulting article, published online at CNET's News.com under the sedate headline "Google Balances Privacy, Reach," was anything but sensationalist. It mentioned the types of information about Mr. Schmidt that she found, providing some examples and links, and then moved on to a discussion of the larger issues. She even credited Google with sensitivity to privacy concerns.When Ms. Mills's article appeared, however, the company reacted in a way better suited to a 16th-century monarchy than a 21st-century democracy with an independent press. David Krane, Google's director of public relations, called CNET.com's editor in chief to complain about the disclosure of Mr. Schmidt's private information, and then Mr. Krane called back to announce that the company would not speak to any reporter from CNET for a year."
Apparently it isn't unusual, Apple and IMB have both punished companies for when they didn't like how their CEO was presented (or revealed) but considering it was Google itself that offered up this information, considering that they have one of the most cohesive cultures I've ever seen, in which a core value is information access, this is odd.
The article is worth a quick perusal, before it disappears into the deep archives, because it also raises a larger question. Just because information is out there, and can be brought to a larger circulation, should it be? While security through obscurity is not a wise policy, it's kept a remarkable amount of information safe for a long time.
Even if Eric E. Schmidt's personal contact information was always findable by those who tried hard enough, I'm sure he's feeling the difference between the number of folks contacting him after a search and those who are now enjoying "crimes of opportunity," dropping him a quick note to tell him "google roolz" or "sux". Then again, this is a lesson Google needs to learn as well, and those lessons earned through personal experience are those we remember best.
Notes from the last panel of Vertical Leap
Moderator barney pell, mayfield
ofer ben-shacher, RawSugar
julia komissarchik, Glenbrook industries
Paul pangaro, snap.com
bob syman, pubsub concepts
Julia: five words
don't search, question get answers
our company harvests fact form the web to give to users are answers.
bob, pubsub:
we monitor the weblogs real time + many others including FAA, earthquake warnings, specific focused streams
ofer, RawSugar
better search, courtesy of your friends
We define ourselves as a social search engine. you can see bookmarks of freinds, of community. many companies in the same area, our secret sauce is we can show tags underneath the first tags, but go beyond.
Paul, snap.com
an alternative experience for experienced searchers. all search is local search: me. now.
the idea is to put the power in the user's hand to control in real time the ranking process.
Q: what's the economics of creating a microvertical. How to extract the structure from millions of pages?
Julia-- six challenges
1. Dynamic pages versus static is a problem. Not even passworded. Did you have enough intelligence to ask the right questions to get a reasonable result? (Job or flight search) complex queries in dynamic web.
2. Once you've extracted those pages, precision becomes a critical issue.
3. Once you have a fact, how do you get additional information? Once you have a flight how do I combine with car, restaurants.
4. Long tail-- how do you find the smaller players? How do you get the local plumber's website?
5. You have to have flexibility; one vertical may not be enough. Maybe market size is too small, or maybe one vertical leads to another such as travel and local.
6. Speed of processing. Facts are harder than pages.
We do that.
Q: is it always pull
Bob: it's not always going to be pull. Pull just won't do it. Low latency requires publishers inform engines. Push based models are back. That model responds to the publisher’s needs. When a new job, a new flight a new product arrives, people can know about it. A lot of search takes advantage of temporary and anecdotal limitations. Like push-- we haven't had an infrastructure that supported the changing web. The offers that appear and are only good for a day or two. Big sites take weeks to get around. but big search engines have different index cycles for different sites... As we layer in IM, blogging comunity efforts, letting blogs tell search engines when the y update, we'll see changes in push. One place engines can own vertical is when they have structured information. The big guys are good with unstructured information. But structured data allows more. Technorati, semantic web will change things, in the future when someone wants to hire, they'll blog it and the search engines will find it instantly. Another bug in search engines is that they can’t' understand the content, but structured content will also change that.
getting tired. But in the last session. Must stay tough!
Q: will all content be decided by publishers
Ofer: we don’t think so. Sure you can have some publisher decide they want to go for the greyhound market, but more likely you'll have a bunch of enthusiasts who get together. If you look on a search engine you'll find directories made by enthusiasts, and you'll then have to go through them. But with searching on shopping.com you can get listings, narrow by facets. Think about taking all of that and bringing it to everybody. Suppose everybody has access to that. The greyhound and tag their sites and suddenly everyone has a better search.
Q: will people will go to verticals?
Paul: no people will have to spend too much time trying to figure out where to go. On the business side-- what does a vertical mean? But it's so important to everyday life, portals and search engines feel they have to serve it. News is common, real estate is valuable... there is a range of value to help choose. Where will the traffic come form if I am a narrow vertical? If many sites have to build their traffic, it's a problem. Snap thinks of itself as a general search engine, but verticals have promise-- digital drive shaft. Something like a camera is something we understand. As UI, my expectations are clear on travel, or news, I have a cognitive map. But I go somewhere and suddenly the dropdowns don't work and ajax isn't there, a common UI becomes desirable. imagine a world in which whenever I go to an interface, no matter what the vertical is, the UI is always be the same. If you can harness quickly, provide a uniform experience for the user, verticals become like horizontals.
Paul: when we are searching, we don't want to search. we are trying to comprehend things in order to act. when would we want different interfaces?
Bob: yes, genealogy... you have specific needs, and ways of talking. you don't want a list, you want a graphical representation, and you want to see relationships.
Paul: the richness is not in listings, listings are out.
Bob: listings are always nice for basic stuff, but there are places where you want more.
Julia: we can never get semi-structured information that is rich enough to get via tags only. you will always need another layer to help you extract information.
Ofer: it's amazing to me that we're talking about search which is supposed to be about helping us find, but in real life it's always about what people we know that is important. but in search we are not talking about that importance.
Q: why not rss, Bob?
Bob: we do use it. atom and rss, for those who want to pull. we ofer both for politics, since both will work with all the aggregators. we also use push feeds, we use the atom format. But it doesn't really matter. Atom is designed for push. rss is not, which is why we use it in the push environment, but in pull we allow both.
Q: advice of young companies starting vertical search.
bob: unless you have really special access to unique stuff, don't build a walled garden. walled gardens are going to be harder to defend, so build on a better experience not captured data.
Ofer: search is affecting every decision... searching the internet is like searching the world. this will grow in importance. one advice is that this market is still in infancy. think big. many things can be done that no one is doing, and in 0-20 years it'll be different. don’t' think of the next little improvement I can make on this o that engine, think of how you can change the paradigm.
Paul: watch out for tags. tags are not meaning. given all the disagreement, I say it's early and we're in trouble. don't presume everything we do now is everything. the tech is young, the biology is old. technology is disappointing because it's implicit and brittle, but biology is the opposite... start with the conversation.
Julia: find your niche.
first more rules for vertical entrepreneurs
rule 9 find your inner killer ap.
understand the core problems and built tools to solve those problems. make it sticky.
rule 10, let's get vertical.
community is a natural for verticals.
Steve Gilmore of zdnet moderating
tantek celik, technorati
index more than 12 million weblogs in real time, and sue that data for other things like top movies, etc
also trying to create microstandards for verticals.
Scott rafer, people will move to subscription model and there is a biz model for that
Jim pitkow, moreover,
enterprise and provides to big guys like MSN.
Chris tolles topix.net
categorizes news.
Q: Scott can you connect threads across the sessions?
Scott: I wasn't in all of them, i had investor things to do. but one thing I see, is how people see vertical search. vertical search is vertical market. but shopping search is functional in a certain way. and our work is more about a new standard rising. we're connected by rss, and know it screws up pagerank. here is a place where there is an inefficiency between gold standard of google and what some users want. So let's get that together and put a great ui around it and see what happens.
Jim: people read news online every single day, and it's tough, tens of thousand of sources, updating all the time. and end users want to know about it as it happens. search can be slower. and news is old, profitable and has standards around ti, so it's an interesting space.
Chris: in the shopping panel, a good shopping product isn't necessarily a search engine, with news, a big factor in relevancy is timeliness. if you want stuff from the last five minutes, the last hour, that's what ties together this vertical. Freshness.
Q: (Om Malik steps in biz 2.0 as moderator)how do you add context to the same news story over and over.
Chris: one thing we do, is cluster stories so we can cluster all the versions of the same Ap story. So the data structures is an event, not a story. so we don't' have 70 of the same story on top of each other. then we rank, and we have a story picker on the front page, and you can categorize by locality. We create a formula to act as editor.
Jim: relevancy is an interesting point, relevant to whom, when? We see it as a matter of metadata so users can drill down. provide context, so san joses make sense based on hierarchy. Also, add authority so you can create a feed that has locality, topic and authority. data isn't just data, we have metadata on top. you have flexible and agile, we use tech, we add human editors, and allow users to customize.
Scott: dedupe is important, but a lot of what we touch is two/three lines different. and that two/three lines makes the difference. we have to respect not pulling out the information provided. these are not dispassionate users, the energy is incredible. here is heartfelt work, and you can't just dedupe that, we've tried dialing it up and down and you have to decide what feeds are most important and go with that.. pagerank takes to much time to build. so the person provides the authority, that person has been blogging on this topic for a year and a half, and make them a micropublicaiton. it's tricky. deduping can loose publishers and searchers.
tantek: relevance is hard-- freshness, authority, SN? what makes it relevant? you can look at yahoo and see what's emailed and that tells you something, but someone who has blogged it, who has gone to the trouble of providing comments, looking at the hypermesh of bloggers can tell you what is relevant.
We've seen some stories that pop first in blogsphere before mainstream media, such as the tsunami. they can match ads to level of profanity, they can match to kind of consumer (a.k.a. don't want to advertise nokia on sony fan). there were problems with democratic ads on republican sites.. advertisers didn't mind as much as site owners. (tantek... accidentally, right?)
Scott: rotten tomatoes has figured out if the overall is negative is positive or negative. it can be done technologically.
Tantek: it can be hairy also.... we've been working on open standard on publishing reviews. the problem is there are many dimensions, it's a challenge to do. But I want to return to Om's question. With technorati you can see more recent, or least which allows for story breaking. on authority, we-humans determine it many ways, political bias, etc. but we (technorati) we do it abstractly, by (something that sounds like pagerank).
Q: how do you deal with relevancy against the reblogging of things, how to take the most value raise to the top.
Chris: most people don't want everything. maybe when egosurfing. but people offer a different kind of relevance, editorial touch. If many people are writing about a story, such as tsunami, we can put the scoop at the top if it works... but sometimes the scoop isn't important. the story is a commodity-- it doesn't matter who did it first, you want the best story or someone's take/brand preference is bigger than scoop. if you are a fox viewer, you want the fox view
tantek: do we want to encourage that kind of siloing of viewpoint?
Chris: you want to expose people to as many points of view. it's not as important if it's NYT or a blogger.
Jim: Some people care about the source, some want the color commentary. We can ask what is authority, or we can ask to whom? can we give tools so users can choose who is authoritative? these are pieces of metadata and we can use it. We can offer tools.
Scott: mentions "who broke the story", it's hard to tell who did. sentiment, editorial bias... is important to CPC advertisers also.
Q: you all have an editorial slant, what kind of user interface suits that?
Scott: it's religious. We believe in RSS, in which there is no ultimate interface. we are a we service. 99% of data leaves in xml. I don't knwo all the interfaces in which we're being used. I probably wouldn't' like most of them. We keep having to struggle with TOS, because we want people to remix, but we want to get paid. there is no ultimate UI, there is a long tail of user interfaces. and most value is in the one I'll never see?
Tantek: what is the one resource I am short of? Time. I have five minutes to hear something that is relevant to me. the UI's that will be best for me will be the most valuable.
Scott: the interface has to account for other things-- people returning to your site and optimizing for people to return to your site.
Scott: there will be many interfaces with the same core data in the future (lost track, got unplugged)
Jim: for us the interface is the api. that's the beauty of the data layer. you have to have good data, good metadata, and and let the user choose (I wonder if he means the user or the provider)
Scott: if it bleeds, it leads (tantek, you just said the reason I turned off my tv)... we want to see what's interesting. we have a point of view. you build for you users. if you are building for someone whining about a rss standard, that's for you.
Scott: the interface of feedster is fine most of the time, we've got the white page, the box... but if it's red sox, it belongs on the Boston globe and that's the right interface.
Q: not what he meant... can you send me data in the way I'm used to, can you get me the information I need, that's what I mean as UI.
Chris: we're an agregator, so we can't reproduce the NYT way... depends what you mean.
tantek: describe your ideal news reading experience, om.
Om: there is a a ordering of stories, most important to least, but on the others it doesn't work that way. How do you come up what the UI-- i see stories i want to break my head against the wall, i can't find what I want to see.
Scott: here isn't' that much news out there, look at the %of stories in chron that are from the wire. Some days there is no news.
Scott: there are only so many MCI executives to convict
The technorati guy dos these beautiful graphs of story strength, peaks and troughs of attention. I have things that only go bright blue (popular?) every few weeks.
tantek: we talked about different ways of filtering. on NYT that's editorial filtering, humans choosing.
Scott: 12 or 15 hours ago
tantek: or you can show what is most important to the blogosphere. But what it sounds like you want is a persistant search, so that if you are into a specific topic, you can follow it.
Om: a newspaper chooses the story of the moment, so why don't' you do that.
Chris: you can do that with a top page, like googles.
Q; how well is tagging going?
tantek: we've seen amazing results. by allowing people to tag their blog posts, and bring in flickr and furl, you can across. We've seen some spam, but they stick out like a sore thumb. mostly it provides a lot of value. If you have shown interest in one tag, and there is new stuff with that tag, we bump relevance.
Scott: I think tagging your own stuff leads to spam. With stand alone tagging, we're trying to get less sophisticated people to tag, and that is beyond most people even those who have blogs.
tantek: everyone is learning from everyone else-- we've seen so much interest. tagtuesday.com
Q: Micropayments?
Scott: we think google is already a massive micropayment system, it's called adsense. Micropayments to read an article is going to be limited.
Chris: everyone has tried it and it's never works. there is no success story from micropayments. people don't buy that stuff in volume
tantik: I have to agree-- there is an explotion in content, the question is not how can I pay for good stuff, it's how do I find the good free stuff.
Scott: of course I was completely wrong in regards to itunes.
Om: I can tell you for a fact, no one will get rich form adsesne.
Chris: people are making nice money form adsense
<>>dissolves into madness<<>
Scott Jampol-- travel search/farechase
Some people are very interested in (misssed the rest)
Vajid Jamfri -- Cfares-- huge, lots of opportunity, the next biggest is illegal. Very fragmented, hundreds of distribution channels. How do we collect this information? You can buy a ticket to London for 2000 dollars, but you don’t know you can buy it wholesale for 500. As search engines create information for the consumer, that information is also invaluable to the supplier to dynamically find price. Try to move form UI's to saver system to a new system (? same as all maps being on top of navtec?)
Q: how do you deal with confusion in marketplace over guarantees, etc?
Vajid: never go to sleep thinking you got the best price... no one can give you that. 20% of billions of travel are in wholesale, and you can't get that? Can search engine provide you that? Only a search engine can revolutionize in this way.
Q: how do you blah blah (the moderator's questions are really oblique. I think he's talking about the relationship between airline sites and the comarison sites)
Scott: we don't crawl without permission. Right now we look to provide value to travelers and companies.
Q: how do you sum up the relationships (between airlines and the comarison travel sites)?
Scott: right now the conversations are really positive. Right now we try to stay unbiased, we do not charge for listing, rankings are not affected. We're working through issues around that.
Vajid: it's all about value-- the low cost carriers are beating the big guys. People are looking for expedience, they are looking for richer content, and they are looking for comprehensiveness.
Beatrice Tarka, Mobissimo-- we've been successful working with carriers who have not wanted to be crawled before. The market evolves. Some people will change-- they are in period of review. They may perceive us as a threat, since we are a new model that is beyond expedia.
Phil Carpenter -- sidestep
You see changes, some people who have pulled out of expedia who will work with sidestep, or others who will do both. Looking for partners that are very partner-centric. They also want partners who have traffic-- you have to be big enough to provide interesting volume levels.
Q: who are competing with? Expedia, yahoo? General engines or online agencies?
Phil-- the easy money is shifting from online agencies. Companies don't really like traditional companies, expensive way to sell. Search engine drives directly to website, allows them to build their own brand, it's attractive to them.
Beatrice: the majority of travel spending is still offline. Radio, TV, newspaper. The first shift is to online. Companies can track directly roi. (Insert 50% of my advertising budget is wasted joke here) Right now Google's and yahoo's of the world are attracting a lot of that money.
Q: when you outline your offer, what kind of cost differences are there?
Beatrice: general search engines are 2-5 dollars a click, but travel engines are paid only on sale and it's much less.
Scott: for ytravel, it's on a search monetization model. So we surround the free results with paid results. It's quite different. When you are talking about trust, you protect that but it's an extremely good value proposition to suppliers.
Vajid: airlines pay 20-30 dollars for acquisition of clients traditionally. From the search engines it's $5. A major spread. It is enough a reduction to make them profitable.
Q: traditionally online is per click, now with verticals it's per acquisition. What does that mean... is it a temporary situation? Immature market?
Beatrice: it's a young market. some clients want CPA (cost per action) some others want cpc, because they have advertising so every view is valuable, still others are driven by brand, so cpi may return. It's hard to get the feeling across in a Google text ad, but an image or a video can send a rich sell, differentiate.
Phil: we think it's very important to be flexible with advertisers, so we offer all types. Who are we as an emerging sector to tell them how to do business? The budget lies in different groups with different metrics so we work with them.
Q: How do reconcile web crawling with real time.
Beatrice; in our industry crawling is the only way to get real time-- we don’t have the luxury of getting a feed. A feed would be a bad experience because it would be a day ago, things move too fast. Fare jumping happens... sometiems you can't get the info fast enough. Maybe xml feeds but she doesn't trust them.
Q: will screen scraping continue?
Vajim: xml feeds are starting, but the industry infrastructure is not able to, it's built around saver system. Some progressive airlines are working on it. It may happen in the next year or so.
Q: Asking about travel planning not just selling
Scott of yahoo: we know travel planning is a very complicated and multiweek planning system. We think of it as a travel funnel, starting with inspiration, research comparison, purchase and post trip. You can offer products are all. Farechase is vertical, but yahoo travel is being integrated... say, a mapping tool that shows you what you can do near that hotel. It's important for travel search engines to get to helping people plan and book travel.
Q: why can't you do searches around the dates you want, + or - a few days.
Scott: we depend on what's offered us, some airlines have it and some don’t.
Q: what incentive do suppliers have to work with one or more of you versus everyone else? How do you differentiate?
Beatrice: we are most comprehensive. Half the queries are international. We don't spend money on ad sense etc, so we aren’t competing with hotels on search engines. We provide rare information. And the query management system. But we drive traffic-- you can't just send a query because many of the companies have old systems and queries are expensive, but we take on the cost and send only qualified leads.
Vajim: the value to the suppliers is important if you want to build a large company. The suppliers feel the expedia model is broken, too expensive to difficult. They spend too much, they have unsold inventory, and the customer is owned by the travel sites, not the supplier. They want to own the customer. Any travel engine that solves that problem will get the suppliers.
Phil: you have to provide both qualified leads and enough of them.
Q: what affiliate programs are useful?
Vajid: today there is no search engine that is leading yet. For them to get traffic, leverage they have to team up, they have an aggressive affiliate program.
what is your vision for local search
Paul Levin yahoo-- yahoo's vision is FUSE: find, use, share, expand all knowledge (go john, Michelle and Nate!)
find is basic
use, share, expand is more... we're in the business of serving users. adding reviews, integrating with mobile, concentrate on use. 360 is the share component... you want to know someone when you are getting a recommendation on a dentist. and expanding the amount of content. users are users but also participants.
shallesh roa, Google
creating delight, surprise. experiences that do both. no documented vision beyond Google's... local as an extension of that mission. stay connected to users, don't stay with a static model, stay nimble, maps, local, Google earth is part of that.
Daniel read-- ask jeeves
looking beyond one set of data. local search is more than business listings. local weather, community listings, and more. how can you remix those types of data with an intuitive user interface. and monetization, but believe it will follow if you provide the best product you can.
brady forrest-- MSN
"i agree with him" . Working local, virtual earth, map-hybrid, user can create an interniery on a scratchpad, save listings, add data (sounds very yahoo) immersive environment.
Q: what's missing?
Paul: the human element-- so many constructs but it's the tip of the iceberg.t eh valuable piece is under the surface of the water, which is people's heads, what you use in their daily loves. User contributed content, and user personality.
shallesh-- vertical is a proxy for the word deep.. how do you get depth across a variety of domains. Engaging users is one tactic. Find a way to be a catalyst to index to connect uses with the content. be fully aware it's the early days, so be of a mind toward innovation, look for feedback, don't be comfortable, and ask how do you connect globally?
Q: how do you differentiate when everyone can index the same stuff?
shallesh-- blah blah. quickly innovate, laser focus, we dont' in advance look for a way to do something no one can, just try to lead.
Daniel-- data consistency across different data types is important. community is important-- Craig's list has broken new ground in community and in business model. use of local wikis, blog postings need to be knitted into the mix People still trust web search more than local, and that's crazy.
brady-- no one has mentioned mobile. we need to bring data to the user, when they are out there are need directions. avant go was a start, but phones are better, and there are areas that don't have database.. it has to change.
Q" what chance does a small company have in local, what are the sweet spots to go for a small player, something you might want to aquire?
shallesh-- depth. there is always room for people who have a different notion, know a different space, have a deep knowledge. local is about getting it right. the bar is higher.there is always someone with a very deep knowledge of areas, comapnies,
...like yellow pages...
yes, and also functional expertise.
Brady: also international, there is lots of room for other players. MSN will go international, but not right away
Paul: Lots of room, if only innovation comes fromthis table, I'll be upset.
two areas 1. depth. you can think of local as a vertical, but also as a horizonatal hat goes across subverticals such as restaurants, etc. we want to work with comapnies that have a head start there.
2. advertiser side. we're excited because of the huge amount of advertiser dollars in local... it's hard for local businesses to understand how to use local search, so advertiser tools and services is a great area.
Daniel: destinations and agregators. to fight the big guys, you have to have something unique, something the big guys don't have. on data aggregation side , it's easier, all search engines are looking for sources of data (navtac model, sort of?)
Q: dodgeball's relation?
shallesh-- interesting intersection. recent aquisition, not sure.
Q: What about events? and local journalism?
brady: yup, we missed that. I'm sure you'll see something. (what is this guy smoking?)
Daniel-- yes, lots of opportunity with things like evites.. another subvertical. local search is everything to some degree-- you always have a local modifier. on journalism, another set of emerging data to give users access to. micropubishing, it's our responsibility to bring it forward.
Paul: we consider this interesting and vital. from a yahoo perspective, there is so much to do here, but you will be seeing more of this sort of content (hmm, another opportunity?)
look at northwest something for citizen journalism they are building a group newspaper.
Q: with adsense/overturn anyone make money, how can you get good local content supported with ads, local commerce?
Shallesh: from a Google perspective, it's about delight that creates a virtuous cycle, you will figure out how to monetize. we think of it a couple ways, reaching out to local community as we do now or third party relationships. but also thinking of how do you think of an experience that allows folks to engage but not feel intimidated. if we can create a compelling experience to the end user money will appear. for smaller players, monetizing through adverting has proved to be viable.
Q: what are the ad models?
Paul: new model very much from an ad perspective. you have to build up over time, and one day the dry cleaner realizes 30%+ of his cusomters come form teh web, then it starts flowing, but that's not now. small biz is familar with yellow page listings, so it's comfortable to small biz. Pay per call is starting to emerge, but we are at the beginning of a long evolution. it will take a long time.
Q: d quality i.e. dead businesss, bad results..
Daniel-- data quality is critical. and data quality an monetization is related. highest quality and consistency makes a difference.
Brady-- as well as data traditional, blog and other user data will help improve things.
Q: user ratings? how do you get ppl involved?
Paul: critical and yahoo is an early leader. ratings are important but reviews are even more important. Knowing it's a date or a business restaurant is real value. concentrating on richness. it can't be just restaurants, has to be businesses, even landmarks. it's not as big a challenge as you think-- once they see it and it's easy and comfortable, they do it. the more visible it is the more they do it. we were integrated into an episode of apprentice, to raise visibility, once people see it and start thinking about it?
Q: incentives?
Paul: we dont' do that now. right now I'm pretty impressed with user's willingness to participate.
shallesh: our approach is to help them find the right answer, by offering the broadest access provided across all the reviews on the web.
this is really throwing into sharp relief the difference of the Y & G perspectives
Q: Who's truth? you are moving into the dangerous place that journalists fear to tread.
Shallesh: a great question. form Google perspective, we try not to be intrusive in the relationship of the user and the content they are accessing.
Q: how do you determine authority?
Paul: it's a great question. we want the sf chron reviews, we want authoritative reviews. but also part of it is scanning through as a user, and determining. One tings we want to do is track authority over time. in connection with 360 we are doing this, you can see all the reviews that people have done. sense of legitimacy. but it's still better to have 20 reviews mixed, than none.
Daniel: it's important to have those, and then you can judge yourself who is trustworthy.
Q: can you transfer authority from bloggers to reviews? how can reputation transfer?
Paul: yahoo is doing a bunch of this. You can see how you are connected to other people, if shallesh wrote a review, I can see we have a mutual friend and that transfers authority. is this useful bubbles to the top.
Shallesh-- as a point of difference, if you look for Indian restaurants, we'll show the full gammit, and we won't intervene across those review sets. helping users ascertain by giving them everything. good god
panelists
mark kvamme - sequoia
chris moore- redpoint
theresia ranzetta - accel
andrea stavropoulous - draper fisher
mark, whose group funded yahoo and google-- search has been good to us. plans to fund more search. david filo wasnt' wearing search when the visited-- there was broken glass in the street. Filo said he wanted someone on a 14.4 to find the website he wants in 2 seconds or less. it's the same now-- yahoo did a great ob of human based search, google came along as did great algo-- the future is int eh combination of the two. the experience has to get better, consumers vote with their mouses. it has to be significantly better, visibility better.
one investment.. kayak... travel search engine. travel is unique,a nd one things kayak has... had a deal with AOL for distribution. people forget that google got their lift from being yahoo's search engine. you need distribution. If you have distribution, you have clicks and the advertisers will find you.
Chris-- ask jeeves, mysimon
thoughts on Internet 2.o? fatlands, oddle?
some good wins, not as nice as sequoia (laughter). Google, yahoo, jeeves... good at a certain class of queries: navigational queries. you know what you are looking for and you go there. the next set of challenges is a class of queries that are more complex the the ones you think of today. lots of room for improvement in the search expereinces. All the big guys re tryign to fill in the user expereince, goobling up the niche engines. the question for us is how big can these oppurtunities be?
oodle aggregates classified listings newspapers, watch out! and fatlands is comparison search (another?!?!)
andreas suggets all successful Internet products understand/use/focus on search. I can't think of any business that doesn't center on the search concept that's successful. in terms of moving forward, we're looking for... business model. volume and how much you can charge/earn for each result each action. you don't need both. enterprise with small market but high value 20-50 cents a page, as opposed to a few cents on a page for commercial site. a couple area's I'm focused on, new devices-- mobile device search. become the eyeball bandit. looking at web 2.0 stuff. distributed real time information if you can't assume static web pages is the way of the future. rss, pings, podcasting, blogging, etc.
teresia social networking/search.
see vertical search in focused media. accel has always been focused on rich media. see vertical search as focused media property... mixing human with algo. use humans as filters. not only editors and staff, but end users. facebook, brightcove.
facebook is college students, but the real value is community which mirrors physical community with verification (via .edu) and the end users make their space. good demographic, college folks. with web 2.0 tools, the end users can be more involved.
if it's a niche technology, it can be a problem to scaling, but a niche audience can be value, lasting value. One thing you can't beat is a loyal audience.
Mark suggests get live fast, get as many users using it as soon as possible. It's hard to judge size of market, put it out there to start learning.
he's said he's seen a lot of verticals that should not be VC funded... too small for VC, but still profitable (hot or not)
Andreas reminds you to do the math, make some projections.
Chris.. if you can get 4-5 scary engineers together who will go with no salary, you can do fine. you can sell vertical search for 20 mil, with no VC. you can get these demonstrated on the cheap. adsense makes it easy to monetize.. demonstrate your idea. take the money when there is a good reason to take it, such as scale the engineering team to solve a complex problem, or marketing to blow out on the traffic side.. be selfish and things when and IF to talk to VC. good money in sole proprietorships.
Andreas will do sub-one million investments to get into certain companies, do the bridge... such as technorati.
chris says do the small deals.. they incubated oodles because it could be money. VC can't make money on 100000 investment, but you can. myspace-- we could never had predicted it. we couldn't predict that was the mix. so we invested late.. it depends.
Q: are ads the way to go, or are there other ways to go?
teresia-- what are the ways to make money: ads, transactions and to a lesser degree, subscriptions. greatest rev is when you have two. shopping.com is 50/50 ads transactions. Sometimes we can't see what it'll be in the beginning. we want to see an interesting user experiences and yet entraprenuers dont' hire designers.
andreas... take it a level up. why do people search? these actions-- find a site, a product, etc-- have a value. consider it as a auction model, let people bid them up to their market value. What is the value of the action once the result appears? H5 knows the value. Regarding expert advice, a big issue is that it's hard to prove the value of the result before I get it. How do you avoid taking a discount on the uncertainty of the answer?
Q: Stanford kids starting a search company ask how big investment do you want and how do you find angels who are savvy in search.
mark: We'll come find you (heh)
andreas: you only do the small investments if you believe it will be huge. but if you believe, you do the small investment even when it's not typical-- it's a cheap investmeent.
teres-- we dont' think of dollars in, we think of dollars potentiall out
mark: we never say, put money to work, we htink of ROT-- return on time. VC can only do one or two early stage, you do 15-18 in general. you have to choose. If i invest 100k, you are my investment for the year. I have to believe.
Q: barrier to entry?
Chris-- nontrivial. the big guys are watching. it's easy to get people to find you, but you need a order of magnitude better user experience and distribution.
Andreas-- they are not winning today because of quality of search results. there are problems in google's results. people can game it. google aims for microsoft.. it's hard to fight on different fronts. large companies are slower to react to new threats... by the time they figure you out, it's too late. the current darlings can be the next victims.
Q: duplication of efforts, do you seen any strategic technologies everyone needs a start up could provide?
mark-- advertising networks, and others.. plenty!
chris-- fishing frontier, drive traffic, montezation services.
I'm at Vertical Leap, a great idea... hopefully a great conference.
David Hills of LookSmart is speaking. I've got the ipod recorder going... maybe he'll let me post...
NOTES
his definition.. providing essential search content and tools for people who have a passion, need or repetitive task.
I'm not sure this really captures it's nature... Ill try to think on that further.
two types of search
categories where it works today
jobs, travel, shopping, health, blogging, entertainment, local, homework/education help
by content or audience... provides vertical results. Part technology and part human.... !! interesting. I'd like know more about what that means to him
verticals offer consumers the ability to automate the drill down process, save time, assemble and meet like minded people.
shift toward complementary verticals
focus on essential not exhaustive (!)
more tools... document saving/bookmarking
more alerts
new level of consumption/engagement
advertisers will love verticals... better reach, increased frequency. bester targeting. better segmentation.
he brings up the classic cycle... research, winnowing, decision, purchase... trackable in vertical search? possibly, though need deeper tool integration like Y! shopping has.
he mentions advertisers need and desire alternative mediums.
Dollars will follow. (cable tv took 18 years to get dollars to match share)
all media markets have same path.... online will be no different (pattern of consolidation then diversity... network to cable. david is a ex-tv guy.)
good question: are we providing value to consumers and advertisers. balance!
participants will develop
a affinity with consumers in a category of need
a trusted brand
the ability to market yourself
ability to service advertisers looking to reach a discrete audience (very excited by power of viral/pr/SEO/promotions rather than traditional marketing)
other distribution options
partner with established companies
distributed on branded/white label
provide content and technology plumbing to companies with recognition in the space you desire-- build both brands. (TV model again... cable companies promote each other's shows. when the tried to create a walled garden, they failed) again, there are no walled gardens... it's a failed model. why can't people learn that???
be willing to provide advertiser hand holding-- be aware this is a hard medium to understand and use.
looking around... it's guyville. 1-7, roughly looks like
room for many companies
Innovation usually comes from new entrants into a market.
more than enough advertising money for now
cable and web prove consumers can handle a lot of choice
plenty of fnding
no shortage of smart and motivated people to pick up the challenge
consumers
wil have both broad and vertical needs
can assimilate many brands into their life
want to save time and increase productivity
want additional tools
adertisers
wnt multiple chocies
will support alternatives
wnat increased oppts to reach consumers
will adapt to buying deeper nto the medium if it's easy and productive
participants/innovators
will creat services wiht a laserfocus
Q how do you pick markets?
is there a need, are their adverisers, do you have something to offer (in short)
Q missed question, but he's now talking classic innovators dilemma... talking about why network tv resisted getting into cable, via fear of canabalizing their own audince, plus taking wait and see attitude.and there was regulations
follow up.. why coudln't google just copy anything that succeeds? He says every comapny has ot make choices... no resources are unlimited, google will always choose what they want to do, it may not be what you do. you can't get up every morning worrying about the competition. You can choose to be their partners. and speed can be as important as exclusivity.
Q annecdote about how CBS couldn't take HBO advertising. hurt themselves.
A he liked moving into cable, coudl do business with competitors
Q: vertical challenges-- real time data understanding . relevence is hard, because no link analysis. how can you do relevence? pubsub/technorati relevence issues. How do you see the guts of search changing.
A: a little bit technoloyg, a little bit human. he brings up the directory! let technology updte human knowledge... he looks to see if directory is already there if he wants to move into vertical market. the combination works well togeher (he doesn't mention the joy of the short tail, but hey)
Bu if you create a company of directories, you have 500 ppl trying to go as fast as the web changes. not gonna work!
Q: B2B
A: a little more quirrely from time to tome.. verticals go up and down. but in teh end, B2B will do well, but selectively. some wil do better than others (well, yeah...)
Q: do you think the advertising money will come form the site itself, or partnership with ad company (adsense, oerture)
A: bit of both. any large site uses 3 or 4 products, advertising.com, adsense, etc AND their propriatary ad force. what are accounts geography you can put a person in, what are the marketplaces you can't afford to put a sales person against... when do you use ad networks to fill in the edges? a blend... most folks use at least two, large ones will use 5 adnetworks it's god to be a ad network 80/20 to choose. Some networks do NOT provide relevent ads... deadly to vertical search very true, relevence is king in search INCLUDING the ads. Especially the ads
Q: exhaustive vs targeted?
A: pick and choose... know what's needed in the vertical. is it going to be content, transactional, or community. plus what's the cost compared to value?
Q: vertical search with your own content, as opposed to web-harvesting.
A: your own content provides you with a personality and a brand. your own content can be a set of tools that help power a search service.
The problem with making stuff for the web is all things are temporal. This page I was once quite proud of, and has been replaced by a new design. sigh.
If the old saw is true, then MoreGoogle is great. Open in new window from Yahoo, screenshots from alexa/amazon and interface from google. I might ask "is there no shame" but I'm sure someone else is asking "where's my lawyer"
Seeing ResearchBuzz -- Cookin' With Google made me dream of an ironchef cook-off between yahoo and google....
The Onion | The Onion | Yahoo Launches Soul-Search Engine
"There are bound to be some bugs, but we're not too worried," Semel said. "We at Yahoo have a lot of experience in helping people navigate an environment full of falsehoods, random useless information, and truly horrifying pornography. I don't think the human soul will hold any real surprises for us."
A bit weary of the googlemania, I bagged on reading this issue of wired. But I finally took a look, and really enjoyed the google redesigns. They are fun.
They won't work at all, of course, at least not as a main interface. The three laws of search design are speed speed and speed.
yahoo! vs. google is a very cool visualization tool to compare results. I think the addition of numbers or a bit of instruction would have been helpful, but once you get the concept, it's sweet.
the two rows of dots represent websites in load order, left=#1. So you can easily spot divergence.
here you can see a search for "information architecture" in which a webmonkey index page on design topics is #1 at google, while Yahoo's #1 is jesse's IA resources page. However, I wonder how accurate this is, since when I do the search a second time on google straight, the #1 is argus, then the webmonkey tutorial on IA, while yahoo shows jjg then argus. Maybe slow on the updates?
Still cool visualization idea.
I'm kinda Ms. Super-biased on account of I used to work on Yahoo Search, and I still manage the team that does and I'm a blogger who is into rss and I've never found a rss read I really loved and I think"my yahoo" is a fine substitute, but anyhow, biased as I am, i think this is darn nifty.
You get to judge for yourself.
Google - Yahoo Comparison: Compare Search Results
looks like it to me.
ongoing · On Search, the Series
"series of essays on the construction, deployment and use of search technology "
From Google Fans Fill Web With Buzz Over IPO (TechNews.com)
"As a public company, "if they are not developing a particular product and their competitor has it, they may feel pressure to go in that direction to please investors," said Danny Sullivan, editor of Search Engine Watch, a site that provides analysis of the search industry. "It's 'I've got to keep up with the Joneses whether or not the Joneses are doing the right thing.' Right now they have the luxury of avoiding that." "
Yup-- it's a lot more than just paying the piper....
The new Yahoo! companion search tool bar thingie offers a bunch of stuff like pop-up blocking, search within site, highlight search term and the usual Yahoo! stuff like access to calendar, bookmarks and mail, but what *I* like is the right click access to search. Now I search on everything like crazy-- love it!
(I feel like such a ho, but I love it!)
BTW, I never noticed I had google access in my rightclick until I took the screenshot. What a difference a bit of white space, and icon and a good call to action make. "google search" vs. "Search the web for term" are day and night apart in clarity.
"For logical reasons, Amazon seems to have designed "search inside" to help readers find text in books that they haven't bought yet. But there's just as much opportunity to apply "search inside" to books you already own. ...We tend to think of search requests as generally taking the form of "find me something I've never seen before." But real-life search is often different: You're looking for something you have seen before, but you've somehow mislaid or only half-remembered. "
An excellent insight.
However, I wonder if this has made the original purpose of search, finding things I don't have, less effective. I typed "product management" (sans quotes) into the search box as i always do, and was appalled at the results. i'm certain if they had stuck with their old system of title/subject indexing, the results would have been better. Did they do blind-testing for relevance? Is this another instance of the googlitis-- that the web search way is the best way to do search?
because it ain't. index size. context. known content. just three reasons off the top of my head why sites should not adopt a web search approach to site search.
i'm worries.. have they broken my amazon? Maybe I'll start doing my shopping at Yahoo, and just use Amazon for reference purposes. Hmm, don't think that will help their business model.
Cool new group, Yahoo! Groups : y-search-users, which includes the Yahoo! Search product and design teams.. come ask questions, lob ideas, and ponder that core Internet art, search.
From SearchDay - To Or is Human - 9 April 2002
"Perhaps no other "advanced" search technique causes more trouble than the incorrect use of the Boolean OR operator. Here's why this simple little world can wreak havoc on your search results."
Reading Wired 11.01: Google vs. Evil
"Newbies flocked to the site, grateful for a simple search engine that was both powerful and intuitive. More sophisticated techies came to appreciate Google's computational elegance and its willingness to shun the "portal" model that crammed ecommerce down their throats."
What's funny is everything I know (I do have access to some research data most folks don't) says the opposite. Early google users were hardcore techies who wanted nothing but undecorated data. Later on the newbies started coming, as the early-adoption "taste-makers" such as slash-dot got the word out, and the general media picked it up and spread the knowledge google existed to newbies.
I hate it when the news presents what appears like a logical conclusion as fact.
As for the rest of the article, it's really very interesting. I love and admire google-- they not only power Yahoo, but they are our competitor and thus inspire us to innovate (and interesting conudrum, the in-bed-with-the-enemy syndrome). They also make a lot of good fun stuff. But "Don't be evil" is overly simplistic attitude. some things are easy to recognize as evil, such as a KKK site. But if you censor those sites, then you keep people from being able to research them to form arguments against their ideology. Is censorship evil, making google evil when they stop evil things? Good and evil are for first graders. Grown-up life is far more complex.
Ryan points out The Google gods - Tech News - CNET.com
"Pragmatists in the industry even say its dominance in Web search gives Google a new responsibility to maintain fair access to as many sites as possible, leading some to suggest that it be regulated as a quasi-public agency. Last week, for example, an Oklahoma marketing firm filed suit against Google in federal court charging that Google unfairly began listing the company lower in search results.
"So many people are dependent on Google's free editorial traffic that it's like food out of their mouths to lose ranking," said Danny Sullivan, who runs Searchenginewatch.com. "Search engines are not in the business of supporting people's companies. But if they are going to provide editorial, they need to provide support. "
from O'Reilly Network: Google Needs People [Oct. 11, 2002]
"Partly, we fear the truth. Google News will not be the first nor the last software application to perform some work more efficiently and effectively than humans. Mostly, however, we fear the lies and the harmful ripple effects they cause.
Google's claim that it offers "a news service compiled solely by computer algorithms without human intervention" is misleading, at best. What about the programmers who wrote the algorithms? What about the designers and architects who structured and organized the templates? What about the thousands of reporters and editors who wrote and selected the articles?"
But someone hasn't given up on humans yet.. this same week Yahoo announced a new stretegy to blend human intellegence with computer efficiancy. So who will out in the end? Humans are still the most powerful computing machine in existance. But they are slow and fickle. Computers are swift, impartial and effective but uncreative and lack common sense. And one of the hairiest questions of search-- disambiguation-- is still something humans do best. We'll have to wait and see what the future of search will hold.
Despite the lurid title, From E-sex to E-commerce: Web Search Changes is a really facinating read. User behavior appears to be changing-- glacially, but changing.
Amanda Spinks really rocks. If you enjoy this article, you may want to dig up some of her academic work.
my favorite search term from this week's report
"busines okupacion"
Somebody clearly understands how I spell.
It occurs to me that people coming here via search are not happy when they arrive. Some search terms:
4: hack
3: browser statistics
2: penny for your thoughts myhrvold
2: website statistics sites
2: simscity
2: information about architecture geeks
2: whee squirrel
2: hack sim card
1: postnuke hack
1: anchient greece pictures
1: information and pictures on household gadgets in the early days
1: yahoo articles
1: concept architecture information product design
1: internet wide browser statistics
1: website modeling
1: browser statistics web
1: user centered design
1: books hack
1: visual design list
1: definition of personification
I want to meet myhrvold.
want to know how to design information architecture for search?
(via louisrosenfeld.com)
Madman Madhu Menon gets to be Amazon's guinea pig for a new search design. 
meanwhile, I'm still seeing the same old thing.
visit his site for more pics and commentary, including the tale of him telling others "look, it's changed" and everyone looking at him funny...
I have to say this is one of the smartest things Amazon does... testing new interfaces on small segments of thier population allows them to tweak. However, not telling folks they are a small expirmental group seems a bit cruel: witness madman's confusion.
And that resource led me to The Lycos 50 with Aaron Schatz, which provided insightful interpretation of why people search for the things they do.
"After September 11, the Lycos 50, like everything else in America, has been fundamentally changed.
Tuesday's terrorist attacks led to a massive search for information never before seen in the history of the Internet. Half of the subjects on this week's Lycos 50 are new this week, and of those subjects all except one are related to the attack on America.
Shocking, then, that the #1 subject is not the World Trade Center (#2) or terror suspect Osama Bin Laden (#3). No, the top subject is 16th century seer Nostradamus, thanks to an email hoax which attributed to him a prophesy foretelling this week's horrors."
What People Search For - Most Popular Keywords
NEWS & COMMENTARY
Google indexes blogs to stay current.
Forrester Research: Minitel to lose dominance in France
"Forrester predicts that the predominance of Minitel in France will
fade away over the next five years." Now there's a newflash....
Cyber Dialogue: Unsolicited email irks consumers
"Seventy-seven percent of US Internet users say their privacy is
invaded when they receive an unsolicited email from a company they do
not know, reports Cyber Dialogue." and another newsflash....
Nielsen NetRatings: Banks spending heavily on online ads
"Three of the top five spenders on online advertising in the US in
June were financial services companies, according to
AdSpectrum." maybe the standard should have gone to the banks?
The Industry Standard: Net population growth slows in US
"The growth of the US Internet population is slowing, reports The
Industry Standard."
lots more good news at http://www.nua.com/surveys