The Future of Search

WHAT IS SEARCH?

It’s more than a web destination and a few words plugged in a box

Search has become embedded in everything from apps like Yelp and Foursquare to Siri

Search is a program that organizes the world’s information/data and returns a list of results based on relevance to queries (Hopefully!)

But what is search really?

Technically speaking, Search is actually Database Keyword Technology

And the other half (and most important part) of Search…is being Found!

SEARCH PREHISTORY

The olden days: BEFORE GOOGLE

• CIRCA 1990-1995

WHAT’S CHANGED?

WE USED TO USE SEARCH ENGINES TO LOOK FOR INFORMATION

NOW WE USE SEARCH TO FIND US!

WHAT WAS ONCE TRANSACTIONAL NOW SEEMS AN EXTENSION OF OURSELVES

THEN VS. NOW

CONSIDER FINDING THE CURRENT TIME IN AUSTRALIA:ONCE IT WOULD HAVE REQUIRED THE FOLLOWING STEPS:

1) KNOWING CURRENT TIME WHERE YOU ARE2) SEARCHING FOR A REPUTABLE APPLICATION THAT

TRANSLATED TIME ZONES3) ENTERING YOUR LOCATION AND CHOOSING AUSTRALIA FROM

A SEPARATE MENU

NOW ON GOOGLE, YOU CAN SIMPLY TYPE OR SPEAK AND GOOGLE DOES THE WORK AND UNDERSTANDS WHAT YOU WANT

+ =

IT’S STILL NOT A PERFECT SCIENCE

AMBIGUITY IS THE ENEMY!“HOW BIG?” “ARE YOU LOOKING FOR LENGTH OR WIDTH?”“SHOW ME THE MONEY” VS. “SHOW ME AN IMAGE OF A DAHLIA”IT’S LIKE GOING TO THE DOCTOR AND SAYING, “GUESS WHERE IT HURTS?”THE PRESENT AND FUTURE:STRINGS ARE NOW “THINGS”GOOGLE IS INDEXING THINGS VS. WORDS

ENTER THE KNOWLEDGE GRAPH

CREATED IN 2005 AS DATABASE OF WORLD’S MOST USEFUL INFORMATION BY METAWEB500 MILLION THINGS LISTEDALSO REPRESENTS CONNECTIONS BETWEEN THINGS CALLED “EDGES”TAJ MAHAL=BUILDING=MUSICIAN=BEER=RESTAURANT=CASINO, ETC.THE GOAL IS ANSWERS NOT LINKS

HYPERNYMS!

WORDS THAT MEAN THE SAME THING AS A MORE SPECIFIC USAGE“JAGUAR” HAS 26 DIFFERENT MEANINGS:THE ANIMAL, THE MAC OS, A POPULAR SOUTH AMERICAN ARTIST, ETC.THE KNOWLEDGE GRAPH IS ATTEMPTING TO CHANGE OUR UNDERSTANDING OF HUMAN INTENT

HOW GOOGLE SEARCH WORKS:#1 SPIDER DISPATCH

• ROBOT PROGRAMS SCAN NEW AND UPDATED PAGES AND INDEX EVERY WORD (EXCEPT “A”, “AN”, “THE”)

• GOOGLEBOTS FOLLOW LINKS FROM PAGE TO PAGE MAKING ITS INDEX MORE COMPREHENSIVE

#2 INDEXING

KEYWORDS PLUS:• METADATA INFORMATION ON WHETHER

KEYWORDS CAPITALIZED

• FONT SIZE• WHERE FOUND ON PAGE (HEADERS, TITLE,

SUBTITLE) ALL ARE USED TO RANK IMPORTANCE OF A WEB PAGE

#3 RANKING

UNIQUE FEATURE OF GOOGLE’S ALGORITHM IS PAGERANKRATES PAGE’S IMPORTANCE BASED ON: • NUMBER AND REPUTATION OF LINKS THAT

POINTED TO THEM• TRAFFIC• HOW OFTEN KEYWORDS APPEAR• FRESHNESS OF PAGE AND SITES THAT LINK TO IT

#4 DEFINITION

GOOGLE’S ALGORITHM USES OVER 200 SIGNALS TO REFINE A SEARCH QUERY INCLUDING:

• WEB SITE’S PAGERANK• SEARCHER’S GEOGRAPHICAL LOCATION• WHICH LINKS A USER USUALLY CLICKS• HOW THEY MODIFY SEARCH QUERIES

WHEN THEY ARE UNSATISFIED• SEARCH HISTORY

3 FURTHER SEARCHES

BING • RELATED SEARCHES APPEAR IN THE RIGHT- HAND COLUMN• VERY REFINED IMAGE SEARCH• INTEGRATES FARECAST GIVES BEST FARES• CANT SEARCH FOR SPECIFIC DATES• CANT FILL IN GAPS IF YOU GET SEARCH TERM

PARTY WRONG

YAHOO!

• INTEGRATES RESULTS WITH ORIGINAL CONTENT, AUDIO, VIDEO, IMAGES

• CAN PERSONALIZE PAGE TO MAKE THE MOST RELEVANT RESULT TO YOU APPEAR FIRST

• USER INTENTION ISSUE—OFTEN GIVES LINKS TO IRRELEVANT SITES WHILE GOOGLE GIVES SHOW-TIMES IN YOUR CITY

YANDEX

YANDEX IS MARKET LEADER AND #1 WEB SITE IN RUSSIA• LOCALIZATION IS STRENGTH FOR RUSSIAN

RESULTS WHEREAS GOOGLE FOCUSES BROADLY ON COUNTRIES

• FAR FEWER SERVERS/DATA CENTERS THAN GOOGLE & CANNOT INDEX ENGLISH LANGUAGE PAGES AS FAST AS GOOGLEBOT

BAIDU

• 78% MARKET SHARE IN CHINA VS. GOOGLE’S 18%

• RESULTS CENSORED BY CHINESE GOVERNMENT

• STILL BEST OF THE CHINESE DOTCOMS AND #5 RANKED WORLD WEB SITE

THE FUTURE: SEARCH EVERYWHERE

•What is TIVO but TV Search?•What is iTunes but Music

Search• Local and Personal

THE FUTURE NOW: KEY TRENDS TO WATCH

• LONGTAIL KEYWORDS RULE• GEOLOCATION IS HERE• AUTHCODES ARE STEALTH• CLICKSTREAMS MAKE SEARCH THE NEW

INTERFACE• THE SEMANTIC WEB

THE FUTURE NOW:SEARCH IS SOCIAL

• SOCIAL MEDIA SEARCH— TRUE FREQUENCY MARKETING VS. TRADITIONAL PAGE RANKING• REAL-TIME POSTING• COMMUNITY-BASED SEARCH (MOG,

SPOTIFY)• SOCIAL INTELLIGENCE AND REPORTING (RELISHMIX.COM)

THE FUTURE NOW

• HUMAN CURATED SEARCH FROM BOTTOM UP NOT TOP DOWN (BLEKKO)• OWNED VS. EARNED VIEWS• MOBILE SEARCH (BOOPSIE)

THE FUTURE NOW

• SENTIMENT ANALYSIS• SEGMENTED SEARCH ENGINES

(PRODUCT SEARCH—DECIDE.COM)• CONTEXTUAL SEARCH

(ADWORDS)

“THE FUTURE OF SEARCH IS VERBS”

WHAT DOES BILL GATES MEAN BY THIS?

MAYBE IT’S THAT WE WANT TO SEARCH TO DO THINGS, AND NOT JUST SUGGEST THINGS TO US

THE FAR FRONTIER IS “UNIVERSAL SEARCH”

WE THOUGHT THAT WHEN THE WEB UNIVERSE WAS COMPLETE THE ENTIRE WEB WOULD BE SEARCHABLE:

WITH DIGITAL LIVES FRACTURED—IN APPS, WALLED GARDENS LIKE FACEBOOK, CLUNKY INTERFACES—REUNITING IN ONE UMBRELLA PLATFORM THAT IS SEARCHABLE IS LARGEST PROBLEM TODAY

(ACCORDING TO JOHN BATELLE, AUTHOR, THE SEARCH)