Hyper Analytics for Social Media Blogworld 2010 by PeopleBrowsr and Klout

Hyper Analytics and the Collective Stream of Consciousness

Jodee Rich CEO @PeopleBrowsr@WingDude

Joe Fernandez CEO@JoeFernandez @Klout

2

HUMAN SOCIALISATION

Hyper Analytics for Social Media – October 2010

Swinging through the trees..

3

HUMAN SOCIALISATION

Emerging from the jungle with Language


4

HUMAN SOCIALISATION

Thousands of years later we wrote it down


5

HUMAN SOCIALISATION

PCs, the internet, cell phones have come together to enable a vast distributed network of human intelligence


6

HUMAN SOCIALISATION

A Persistent Stream of Consciousness..


7

HUMAN SOCIALISATION

Established Industry, the Stream and the Cloud


8

HUMAN SOCIALISATION

Persistent Social Framework displaces Industry


9

YEARS TO REACH 50 MILLION SUBSCRIBERS

Radio 38 Years

TV 13 Years

Facebook 100 Million in less than 9 months

Source: Socialnomics09

9Hyper Analytics for Social Media – October 2010

“That’s what social networking is all about: worshipping at the Altar of Self.”

nothing more than an exercise in unashamed public vanity for the Look at Me generation”

Sandra Lee in The Sunday Telegraph - May 25, 2008

PERCEPTION TWO YEARS AGO…


11

1 YEAR OF TWITTER TRAFFIC:

View chart and stats on analytic.ly

Now at 50 Million Tweets/day


http://www.analytic.ly/

12

STREAM SIZE 2010: IT IS DAVID AND GOLIATHNext to Facebook, the Twitter Stream is still small


2010 OPENNESSLittle Twitter is dragging the others out of the cave and into the open


OPENNESSBecause it is open, the Twitter Stream will become the core transport layer for rich MetaData and Cross Network Links



EVERY BROADCST MEDIUM IS MEASURED

http://www.flickr.com/photos/moriza/126238642/


THE INDIVIDUAL IS NOW THE BROADCASTER


NETWORK VALUE IS A COMPONENT OF LIFETIME VALUE


SIMPLE COUNTING IS NOT THE FUTURE OF ANALYTICS


ANALYTICS CAN BE USED FOR MORE THAN JUST GRAPHS

20

1 MILLION BRAND MENTIONS PER DAY


21

CASE STUDIES IN 2010

BlogWorld 2009

Starbucks

eBay

Toyota Recall

Super Bowl Ads


22

Goals:

• Build Audience

• Sell tickets to the Conference

• Create buzz around the event

• Monitor the social stream

Industry: Events and Marketing

Background: Producer of the largest Social Media Conference that annually hosts 10,000 people

Length of Engagement: 15 months

Metrics for Success

• Identify Key Influencers

• Increase # of Conversations in Public timeline

• Increase # of Followers

• Maximize sales of tickets

CASE OVERVIEW: CONFERENCE


23

Solution: Campaign Target Audience

• Users mentioning BlogWorld keywords

• Bloggers

• Social Media Enthusiasts


PeopleBrowsr extracted all the users who mentioned BlogWorld and relevant keywords for the last 360 days.

The list has also been populated with the participants to the conference in 2008

PeopleBrowsr selected the TOP influential users to engage with.

These Tweeters received a Private Message with a special discount code to the conference


24

Performance and Results

30% Percentage Increase in Relevant Followers

11,000# of Mentions of the hashtag during the conference

39%# of Clicks Through


180,000# of Tweets extracted for 360 days

171,423Additional Followers reached by ReTweets

962# of Users who ReTweeted the Message


STARBUCK BUZZ CREATED Campaign Period


26

TOYOTA RECALL SENTIMENT ANALYSIS

Overwhelming negative mainstream media coverage and very little negative social media comment


We analyze 18 months of historical data and live streams in real timeWe use a special algorithm and then have real people improvingthe result up to 95% accuracy.

95%

accuracyVs 70-80% automation alone

27

US AIRLINE INDUSTRY STUDY JUNE 2009


Analytics:

• Effect of Traditional Media on Social Media

• Mechanical Turk to measure accurate Sentiment

•Metrics to measure Success:

• Total Mentions

• Positive Mentions

By Volume

Mullen and Radian6

28

SUPER BOWL


29

SUPER BOWL

Results:

• 103,158 Total Mentions

• Sampled 1000 Tweets from Every Brand and used Mechanical Turk Human Sentiment to analyze

• Polarized:

• 50% Positive

• 28% Negative

• 18% Neutral


30

SUPER BOWL Correlation of Tweets and Ads


31

SUPER BOWL

By Volume

Mullen and Radian6

O’Reilly and PeopleBrowsrO’Reilly and PeopleBrowsr

By VolumeBy Positive


32

Seek an effective way to measure brand sentiment

accurately. The goal is to find a list of influencers

speaking in both positive and negative terms and engage.

Call center to respond to negative sentiment everyday

Velocity

10,000Mentions/day

filtered to

180Meaningful

comments

Velocity

10,000Mentions/day

filtered to

180Meaningful

comments


33

TWEETS PER CAPITA

Source: PeopleBrowsr DataMine

A new SocioEconomic KPI


http://peoplebrowsr.com/

34

THE NEXT TWO YEARS

The Conversation Stream becomes the Conversation Cloud

A real time historical record


35

THE NEXT TWO YEARS

The Conversation Cloud becomes the Rich Metadata Cloud

Social Media Stream will become the core backbone for people data


36

CUSTOM ANALYTICS


37

SENTIMENT ANALYSIS


38

SENTIMENT ANALYSIS BY DAY


39

SAMPLE POSITIVE TWEETS


40

SAMPLE NEGATIVE TWEETS


41

VOLUME OF MENTIONS


42

REACH


43

WORD CLOUD


44

TIME TARGETING


45

GEO TARGETING


46

GEO TARGETING


47

AUDIENCE PROFILING


48

AUDIENCE PROFILING


49

BLOGWORLD ANALYTICS PLATFORM

http://Analytic.ly/Blogworld


50

EXPERIMENTAL APPS RESEARCH.LYWhat can we build?

RESEARCH.LY

Contextual Search and PostInline Content

RESEARCH.LY

HyperLocalLinked to other netwoks



SEARCH THE WORLD OR JUST THE COUNTRY OR CITY YOU WANT

The first thing you will notice are the Country and City drop down menus

Use these as you search, so the results you get are more relevant

t2.ly uses the geo metadata in Tweets


SEARCH EVERY TOPIC OR JUST THE ONE YOU LIKE

Refine the search for Your Community

Choose any Topic that is more interesting and relevant to you


CUSTOM URL FOR DIRECT SEARCH

Define your search directly from the URL

Select Location, Category and Keyword and start browsing!


TWO WAYS TO START SEARCHING

Click on any of the words in the ‘Most Retweeted’ box to start searching

Start typing a Tweet in the ‘Compose a Tweet’ box. Click any word you type to start searching on that word to show related results

OR


RECURSIVE SEARCH

Now t2.ly comes to life; a Twitter stream with related, relevant Tweets appears below

Click on any word in the Search Stream to continue drilling down for refined results


FIND A RELATIONSHIP PATH…

Degrees of separation (1º, 2º, 3º and 4º) show how many ‘steps’ you are from the original Tweeter

The power of demonstrating this connectedness is of enormous benefit in creating trust and value within the search results


WHO KNOWS WHO KNOWS YOU…

1º - you follow them

2º - you follow someone, who follows them

3º - you follow someone, who follows someone, who follows them

4º - you follow someone, who follows someone, who follows someone, who follows them


SEARCH IN FRIENDS STREAM

Find all your friends who Tweeted about any search keyword

Choose to filter the stream by gender or sentiment


RICH MEDIA

These are currently related and trending pictures you can use in your Tweet

These are currently related and trending @Names, Hashtags, URLs you can use in your Tweet

These are the analytics of your chosen keyword


RELATEDNESS & CONNECTEDNESS

Every search also shows related and connected data in the form of images, videos, @names, #tags, retweets and URLs

Each result can be used to search deeper just by clicking it or used as content for your own tweet in the Compose box

VAST DYNAMIC DATA STORES POWERCOLLECTIVE CONSCIOUSNESS


62

REFERENCES

This Deckhttp://slidesha.re/HyperAnalytics

www.Analytic.ly

Socialnomics09

Social Media in Plain English

PeopleBrowsr Super Bowl Study

PeopleBrowsr Top 20 Brands Study

www.ReSearch.ly

http://www.slideshare.net/peoplebrowsr/the-twitter-metadata-revolution-and-collective-consciousness

http://www.nytimes.com/external/readwriteweb/2010/05/17/17readwriteweb-twitter-forefather-leaves-aims-to-disrupt-b-89770.html

http://blogs.hbr.org/research/2010/05/why-gallup-when-you-can-tweet.html

http://www.briansolis.com/2010/05/report-top-20-brands-on-twitter-april-2010/


Jodee Rich CEO @PeopleBrowsr@WingDude

Joe Fernandez CEO@JoeFernandez @Klout

http://www.youtube.com/user/Socialnomics09

http://www.youtube.com/user/Socialnomics09

http://www.youtube.com/watch?v=MpIOClX1jPE

http://www.slideshare.net/peoplebrowsr/people-browsr-super-bowl-deck

Hyper Analytics for Social Media Blogworld 2010 by PeopleBrowsr and Klout

Technology