Top Banner
Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com http://searchenginewatch.com/
39

Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Dec 15, 2015

Download

Documents

Stacy Darcy
Welcome message from author
This document is posted to help you gain knowledge. Please leave a comment to let me know what you think about it! Share it to your friends and learn new things together.
Transcript
Page 1: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Evolving Internet Technologies:Web Search Engines

Danny SullivanEditor, SearchEngineWatch.com

http://searchenginewatch.com/

Page 2: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Overview

Key “technology” in 2001 was survival

Crawlers replacing humans New & old players to watch 11 September & Mindreading Other Things

Page 3: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

RIP 2001Go.com (Infoseek)You were one of the first web-wide spiders andlater added your own human directory of sites

NBCi (Snap)You provided your own human-compiledguide to the web

ExciteYou were another of the oldest web spiders to finally cease crawling

Page 4: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Search Economics

Economics is boring but important! Makes search engines viable; may impact results

Banner ads no longer sell

Listing services new way to make money Allow much needed “conversation” between

search engines and site owners…

But more interactive with results than banners,so searchers and site owners have new fears

What’s offered & should you worry?

Page 5: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Paid Placement

Buy your way to the top

All sell it, even Google

Overture (GoTo) sells for AOL, AltaVista, Ask, HotBot, Lycos, Yahoo and others

In Europe, Espotting sells for Yahoo, Lycos, others

Page 6: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Paid Placement Concerns

Users don’t really seem to mind -- yet

Similarity to “editorial” may cause distrust Main reason behind FTC complaint last July Ask Jeeves, Lycos recently improved labels

Why deny users top sites, if they don’t pay???

Heavy “ad break” might drive users away…

Page 7: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Meta Search or Meta Ads?

Meta Ads Dogpile, http://www.dogpile.com Search.com, http://www.search.com Mamma, Metacrawler like above

Meta Search Vivisimo, http://vivisimo.com IxQuick, http://www.ixquick.com qbSearch, http://www.qbsearch.com SurfWax, http://www.surfwax.com

Page 8: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Paid Submission

Pay to get your site reviewed quickly

No guaranteed ranking – no guarantee to even be included!

Yahoo and LookSmart both offer Mandatory for business categories

Annual charge at Yahoo: Yellow Pages

Page 9: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Paid Submission Concerns

Is it fair to miss some businesses?

How many florists do you want? 100, 1000?

What about non-profits, hobbyists?

Non-commercial categories exempt at Yahoo

LookSmart’s use of Zeal.com feeds itsnon-commercial listings, give good balance

Page 10: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Paid Inclusion

Get deeper representation in listingsand with crawlers, faster revisits

Usually doesn’t guarantee rankings, but…

Like having more tickets in the lottery – more chances to win

Every major crawler but Google sells this, as does LookSmart

Page 11: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Paid Inclusion: Example

Inktomi: $39 gets first URL listed in2 days, revisited each week Want more, $12-15 each, or CPC pricing No pay? Still might get included,

anyway Program has provisions for non-profits No rank boost

Page 12: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Paid Inclusion Concerns

Will we see important sites / pages dropped just because they don’t pay? That works against users and site owners

Fair those who pay better represented? The “real” world works this way

Northern Light worked this way for years

May depend on a case-by-case basis

Page 13: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Humans Were Supreme

From start of popular use of the web, human-powered Yahoo has been top search site

Why? It helped you refine. Search for “travel” gave 10 categories rather than 10 million results

Yahoo “seemed” to find things when it actually gave you less but forced you to be more specific

Others followed Yahoo’s lead…

Page 14: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Rise & Fall Of Humans

By 2000, 5 major human directories “powered” 6 of top 10 search engines

But now, 3 directories power 4 of top 10 Yahoo, MSN, Netscape, LookSmart

AOL, Lycos, Ask Jeeves abandoned humans in 2001/2002

Why the change?

Page 15: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Human Weaknesses

Editors cost money Go and NBCi ran out of this in 2001 Ask also scaled back on human answers

Machines can now do some of what humans originally did… “Related Searches” refine queries in the

way categories did “Autocategorization” also refines…

Page 16: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Auto-Categorization

Group pages into categories, on the fly One reason why Teoma, Wisenut and

Vivisimo get good reviews Not news to Northern Light!

Google says not necessary, but we’ll see They find human categorization better

(directory tab)

Page 17: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

But Humans Still Involved Crawlers better at being “human” because they

leverage human work more than in the past Human-made links used to determine importance Links used to determine context of pages Links used to autocategorize into “communities”

Crawlers also dependent on directories, giving them great weight in considering how to rank So what happens now that Yahoo & LookSmart are

more commercial? What happens if the Open Directory fails?

Page 18: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Who’s New: Crawlers

Teoma.com Potential there, but will Ask have the funds

and know how (this time) to make it happen

Coverage is set to grow; of course, paid inclusion was first “improvement” shown

WiseNut.com LookSmart set to buy it; will this solve the

freshness issue?

Page 19: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Who’s Still Hot

Google (and Google Toolbar) Everything they do is magic Good: finally, a tool you can learn and depend

on

AllTheWeb.com Big improvements recently; take another look

Don’t forget to visit Yahoo’s categories or surprise yourself with a search at MSN

Page 20: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Lessons Of 11 September

People hit Google & others for news

Terms included: cnn, news, world trade center, bbc, reuters, msnbc, sky news, new york times, pentagon, bin laden, american airlines, united airlines

What did they get?…

Page 21: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Google results, 4 hours after attack

Page 22: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

FAST/AllTheWeb.com results, 4 hours after attack

Page 23: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Ask Jeeves results, 4 hours after attack

Page 24: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

But at AltaVista, less than 2 hours after the attack…

Page 25: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

“Blended” results mixed in news content,even if news option on home page

had been ignored

Why Did AltaVista Succeed?

We know historically that home page optionsDO get ignored, but we learn this again onSept. 11, by watching Google and others

Page 26: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Read My Mind

Sept. 11 dramatically illustrates the main search challenge – the need to somehow automatically hit the correct dataset Images for search on “pictures of spain” MP3 files for search on “madonna” Movie info for search on Harry Potter

this month How NOT to do it, then good examples…

Page 27: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com
Page 28: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com
Page 29: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Examples Of Mind Reading

Smart query analysis, then suggestionsor insertions of non-web material Products at AltaVista Sidebar results at AllTheWeb News, dictionary, stock & more at Google Encyclopedia at MSN

Careful not to take away all control Power search for few who want to drive

Page 30: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Specialty Search / Vortals

To mind read, you need specialty datasets Among the majors, Google & AllTheWeb

pushing here, & I think they’ll keep going Also think (and hope) we’ll still see more

“vortals” or “vertical portals” Moreover, http://www.moreover.com MessageKing, http://www.messageking.com xrefer, http://www.xrefer.com LawCrawler, http://www.lawcrawler.com

Page 31: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Other Issues & Trends

Freshness Size “Off the page” ranking criteria

Page 32: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Feeling For Freshness

AllTheWeb pushed end of last year to be 9-12 days old at most – now more likely to be a month, like others

Google & others aiming to be less than a month old or fresher for key documents

Just show dates when pages were visited!

Page 33: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Still Growing, But Still Missing

The leaders? Google, 1.1 to 1.6 billion documents AllTheWeb, 625 million

Large index probably more comprehensive

We do want more index growth! However, don’t judge a search engine

only based on its index size…

Page 34: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Does Size Matter?

To professionals, yes. Coverage helps them find unusual or obscure material What good is half a haystack?

To average users, not really. They desperately need better relevancy. How about I dump a haystack on your head? 100 million extra pages makes no difference to

best matches for “horoscopes” or “britney spears”

Page 35: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

“Off The Page” Ranking

Looking beyond content of the page,since webmasters can’t easily control this

Link analysis still going strong But can produce oddities, like infamous Bush

result Under new pressure from link spammers

Clickthrough measurements not as hot Personalization might get revived with Google

Past fears would limit results, rather than help

Page 36: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Some Closing Thoughts

Yet we’ve had them less than 10 years!

Answers to everything weren’t on web before, aren’t now and never will be, so…

Search engines are the top resource for Americans seeking answers, used 32% of the time

--Consumer Daily Question Study, Fall 2000

Page 37: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Just One Of Many Tools

Don’t expect miracles from search engines

They’re great “Swiss Army Knives,”but you’ll still want an entire toolbox

My hot search tools? Telephone & Email! Use them to avoid “search rage” Stop searching after 10 minutes

and try other means. Also…

Page 38: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

Be Non-Traditional

Forget Boolean, please Don’t cast your net wide

You don’t need every synonym in your query…

Instead, explore what’s in the first catch! Unlike traditional tools, web documents LINK A few good pages usually lead you to more

good pages – your answer may be a few clicks away

You’ll also find links bring you to documents that contain the synonyms you would have tried

Page 39: Evolving Internet Technologies: Web Search Engines Danny Sullivan Editor, SearchEngineWatch.com

http://searchenginewatch.com

This Presentation - http://calafia.com/presentations

Search Engine Watch - http://searchenginewatch.com

Web Searching Tips – Search Engine Listings

Free Search Engine Newsletters(SearchDay – Search Engine Report)

Become A Member – You Support Me & Chris Sherman!(and get some extra benefits)