graph search, facebook nearby & beyond: how social search impacts the future of local by greg...

Post on 10-May-2015

779 Views

Category:

Technology

4 Downloads

Preview:

Click to see full reader

DESCRIPTION

Social recommendations have been a primary driver of real-world commerce for generations. But the rise of online reviews and social networking have made them the rival of traditional word of mouth in local. Social connections and online comments now form the raw material of a new generation of discovery tools and search algorithms. After years of discussion and speculation, social search is finally real with Graph Search, Foursquare and Google+. What is the current state of social search and how will it impact the future of local?

TRANSCRIPT

Greg Sterling Opus Research

July 15, 2013

Social, Search and the Future of Local

•  Lawyereditorstartupsanalyst/blogger"•  Search + local + mobile + social media + SMB marketing"

•  Impact of digital media on real-world consumer behavior"•  Twitter: @gsterling!

"

About-Me Slide

The convergence of three major “online” trends: "

•  Social: word of mouth, user-generated content"

•  Local: online research offline buying "

•  Mobile: the internet “in context” (time/space)"

Last Year: SoLoMo

What Is ‘Social Search’?

Social search uses social data (likes, check-ins, social graph) to influence or determine ranking and relevance compared with conventional algorithmic search, which uses text and/or link analysis

Categories of ‘Social Data’

1.  Ratings/reviews (Online word of mouth) 2.  Social activities/actions (check-ins, likes,

comments)

3.  The “social graph” (connections, followers)

Local Word of Mouth (social search 1.0)

Word of Mouth

Sherman, my boy, traditional businesses have always relied on

word of mouth and personal recommendations for new leads

Gosh

Primary Source of Leads

Source: AMEX/Network Solutions 2/11 (n=400 US small businesses who did some form of online marketing)

Businesses historically viewed word of mouth as the primary driver of business

Source: Nielsen Q4 2011; n=28,000 Internet respondents in 56 countries.

Consumers Trust Each Other

Do you trust online customer reviews as much as personal recommendations?

Source: BrightLocal, 3/12 (n=2,862 respondents from the US, UK and Canada)

26%

21% 20%

33%

28%

24%

20%

28%

Yes, if there are multiple reviews

Yes, if the reviews are authentic

Yes, for some types of businesses no for

others

No

2010 2012

Reviews Trusted Like WoM

Source: Opus Research, 2012, n=1,001 US adults (multiple answers permitted)

Local: Reviews Most Important “When searching for a local businesses online, what types of information are most important?”

33.1%

30.3%

24.5%

20.3%

11.5%

Reviews of the business

Business name, address & phone

Pricing information

Maps & directions

Images of business

A local business needs at least 6 to 10 reviews to be credible and trusted

Credibility Threshold

Source: BrightLocal, 3/12 (n=2,862 respondents from the US, UK and Canada)

Source: BrightLocal Local Consumer Review Survey 2013

In the last 12 months have you recommended a local business to people you know by any of the following methods?

WoM Still Dominant

Rise of Social Directories (and the culture of participation)

Cityguides: WoM Online

•  In roughly 1994 – 1995 multiple cityguide sites launched

•  Restaurant, events & entertainment directories:

- Seen as profoundly threatening to newspapers; less so to YP (at the time)

•  In 1999 Citysearch (IAC) acquired Sidewalk from MSFT

Social Directories •  Yellow Pages 2.0 (directories + reviews)

-  YP publishers initially resisted reviews; perceived conflicts between advertiser and consumer interests

•  Sometimes called “social search” sites •  Angie’s List (est. 1995; online 1999) •  2003 – 2006:

-  Tribe.net -  Judy’s Book -  Insiderpages (acq’d by IAC ‘07) -  Yelp (2004; same year as FB) -  Kudzu -  Others

More Reviews, More Categories

•  “Social directories” had similar ambition as earlier generation of cityguides

•  But sought to bring more consumer reviews to more categories

•  Improve process of selecting a local business online

•  Yelp: “real people, real reviews”

Parallel Rise of Social Nets

•  The Well – 1985

•  Craigslist – 1995

•  Friendster – 2002

•  MySpace – 2003

•  LinkedIn – 2003

•  Facebook – 2004

•  Twitter – 2006

•  Google+ – 2011

•  Tumblr, Instagram, Snapchat, etc., etc.

Most social nets are not “utilitarian” initially

Social Evolution

Search Gets Social (and vice versa)

Search + Social Social + Search “You got peanut butter in my chocolate. You got chocolate in my peanut butter.”

You Complete Me

•  Wants the social data to improve search (and compete with Facebook)

•  Wants to implement search to deliver

more utility and realize the financial opportunity

Crowdsourcing Search Crowdsourcing and social content have been at the heart of the search experience from beginning:

•  Yahoo Directory and DMOZ (‘98) used human editors to organize the web

•  Larry Page envisioned “back links” as “democratic voting” by the entire web re topic authority (1996) – superior to keyword density

•  Vertical sites (e.g., travel, shopping) enjoyed high rankings b/c of social content/reviews

Original Google Algo ‘Social’

Source: searcheverywhere.net (2012)

Social Evolution •  Social an “organic” development for search

- From html docsdocs, offline places, people

- Real-world input from people (WoM)

- More holistic treatment of query

•  Humans offer more direct and relevant “answers” vs. machine algorithm

- Every search query a question

Eurekster: 2004 Eurekster saw “social graph” as a filter/personalization tool (see Blekko)

Problem: not enough social/community

Q&A: ChaCha & Vark Real-time Q&A has always held promise but no one has made it work

•  Aardvark (2008) was effort at real-time Q&A/social search

•  Problem: not enough “critical mass” •  Acquired by Google for $50 million

•  Began as human-aided search (expert guides); now mostly machine generated

•  Problem: humans too expensive

Facebook Connect = Social Filtering •  Facebook Connect (2008) enabled users to see or filter content their

friends had “Liked” on Facebook

•  Integrated into Blekko and Bing in 2010, to differentiate from Google

Bing’s Social Sidebar

•  Bing introduced (2010) “social sidebar”

•  Relevant content from multiple networks

•  Ability to ask Facebook friends query

•  Sidebar changed, redesigned multiple times

Showing “asynchronous social recommendations” addresses the real-time critical mass problem

Makes Sense on Paper

You Made Us Do It •  Facebook wont allow Google to crawl site

•  Bing-Facebook alliance

•  Google wants social-graph data (hence privacy policy change for 360 view)

•  2009: Google introduces “social search” (small “s”)

- Public content from friends/contacts at bottom of search results; also a social filter at one point

•  2011: Google launches Google+ (also +1 buttons)

•  2012: “Search Plus Your World;” focus on personalization but social content instrumental

Search Social Feedback Loop

Search

Social

•  Beyond question that social activity improves ranking on Google

•  Specific variables open to debate "

Facebook Ranking Factors Here’s a non-exhaustive list of probable Graph Search ranking variables: •  Social graph/network

•  Completeness of business data on profile/Page •  Ratings •  Likes •  Check-ins •  Business location vis-à-vis user query

Pages that are more engaging/active (feature more content and interaction) are also going to rank higher

Social Search & Local (back where we started – sort of)

‘Local Is Social’ •  Social search + local a natural fit (see WoM)

•  Social content often local (e.g., Yelp, TripAdvisor, OpenTable)

•  Friend recommendations vs. “10 blue links”

-  Mobile factor: efficiency (“answers not links”)

•  “Local is social” (Marissa Mayer, June 2012)

Places in Graph Search

•  Graph Search (now wide) and Nearby Places Search (app) – use same underlying platform

•  Key feature of Graph Search is Places

•  Q: How committed is Facebook?

- Probably: statements + $$ oppty

- Experience is uneven (even crude) but shows promise

Limited but Promising Results

. . . Or Go Old School

MINUTES MINUTES 66

MINUTES 74

MINUTES 81

MINUTES 64

MINUTES 43

Sources: comScore Q1 2013

User Behavior Developing

Can’t Do That on G

Google’s Local Carousel

Google+ Local

New Google Maps w/Recs

Foursquare Social Rankings

Yelp Recommendations

Personal + contextual + social variables: •  Location •  Yelp check-ins and

reviews •  Yelp friends •  Time of day •  Weather

Pinterest: ‘Social Discovery’ Offline

Summary: Uses of Social Data •  To improve algorithm/results

-  Real-world feedback: comments, likes, follows, check-ins

-  Objective rankings (social actions = community voting)

•  Social graph: explicit filter (asynchronous WoM)

•  To enable “discovery” (implicit)

-  Way back: Amazon collaborative filtering

-  Personalization (along with history, etc.)

-  Search w/o searching (persistent/ambient)

•  Together w/mobile (“context”) social data enable next generation of services: PVAs

Future: Personal Assistant (the return of SoLoMo)

Siri & Google Now

Siri brought “assistant” concept into focus Google Now: “predictive search” (with multiple data inputs)

Other apps/entities use metaphor of virtual assistant (e.g., Nina, Tempo)

‘Conversational Search’

•  At Google I/O company demonstrated “conversational search”:

-  Phone understands context

-  Search can build on previous queries

•  Coming Motorola MotoX to have “always on” listening capability, ready to respond to voice commands

•  Wake-up phrase: “OK Google Now”

The Star Trek Computer

Google has repeatedly talked about building the “Star Trek computer”

Questions

Given the devastating completeness of the information

presented, Sherman, I should think not….

Do you think there will be any questions Mr. Peabody?

Greg  Sterling    greg.sterling@gmail.com    Twi4er.com/gsterling  

Questions?

top related