Monetizing User Activity on Social Networks - Challenges

and Experiences

Meenakshi Nagarajan, Kamal Baid, Amit Sheth and Shaojun Wang

KNOESIS, Wright State University

M. Nagarajan, K. Baid, A. P. Sheth, and S. Wang, "Monetizing User Activity on Social Networks - Challenges and Experiences“, 2009 IEEE/WIC/ACM

International Conference on Web Intelligence, Milan, Italy

Targeted Content Delivery on SNSs

Content-based advertisements (CBAs) Well-known monetization model on the

Web but not translating well on SNSs

Monetizing content on Web 2.0 Where to monetize What to monetize

It’s the talk of the town!

State of the art – Content-Based Ads on SNSs

May 30,June 02 2009

June 01, 2009

State of the art – Content-Based Ads on SNSs

What is going on here..

Interests stated on user home/profile pages do not translate to purchase intents Interests are often outdated.. Intents are rarely stated on a profile..

Some highly demographic targeted cases work

Overall, click through stats are staggeringly low – show some

Intents in User Activity Elsewhere.. Missed Opportunities

June 01, 2009

Concert tickets

MP3 downloads

Services in and around location

June 01, 2009

Intents in User Activity Elsewhere.. Missed Opportunities

Challenges in Monetizing User Gener

Informal, casual nature of content▪ People are sharing experiences and events▪ Main message overloaded with off topic content

Non-policed content▪ Brand image, Unfavorable sentiments1

People are there to network▪ User attention to ads is not guaranteed

I NEED HELP WITH SONY VEGAS PRO 8!! Ugh and i have a video project due tomorrow for merrill lynch :(( all i need to do is simple: Extract several scenes from a clip, insert captions, transitions and thats it. really. omgg i cant figure out anything!! help!! and i got food poisoning from eggs. its not fun. Pleasssse, help? :(

1Learning from Multi-topic Web Documents for Contextual Advertisement, Zhang, Y., Surendran, A. C., Platt, J. C., and Narasimhan, M.  , KDD 2008

Talk Outline

System that generates ads based on activity (user generated content) elsewhere by

1. Identifying monetizable posts: intents behind user posts Pull content with monetization potential

2. Identifying keywords for advertizing from monetizable posts Dealing with off-topic chatter

System Overview and User Studies

User studies Hard to compare activity based ads to s.o.t.a

So we evaluate subgoals How well are we able to identify monetizable

posts (component 1) How targeted are ads generated using our

keywords vs. entire user generated content (component 2)

Intentions Behind ContentIdentification, Evaluation

Identifying Monetizable Intents Scribe Intent not same as Web Search Intent1

People write sentences, not keywords or phrases

Presence of a keyword does not imply navigational / transactional intents ‘am thinking of getting X’ (transactional) ‘i like my new X’ (information sharing) ‘what do you think about X’ (information

seeking)1B. J. Jansen, D. L. Booth, and A. Spink, “Determining the informational, navigational, and transactional intent of web

queries,” Inf. Process. Manage., vol. 44, no. 3, 2008.

From Entities to Action Patterns

Action patterns surrounding an entity (X)

How questions are asked and not topic words that indicate what the question is about

“where can I find a chotto psp cam” User post also has an entity

Bootstrapping to learn Information Seeking (IS) Patterns – offline step

MySpace User Posts (not annotated for intent)

Extract all 4-grams > freq 3

Using seed words (who, when, why, what, how)

Extract all 4-grams containing seed words

Candidate / Potential set of patterns (Sc)

‘does anyone know how’, ‘where do i find’, ‘someone tell me where’…

Bootstrapping to learn IS patterns

10 manually pickedInformation SeekingPatterns Sis‘how cool are we’ is not

Information Seeking

Remaining candidate patterns Sc = Sc - Sis

Candidate patternsSc

Goal: Evaluate candidate patterns and judge if it is Information Seeking or not

‘does anyone know how’, ‘where do i find’, ‘someone tell me where’…

Bootstrapping to learn IS patterns

‘.* anyone know how’

‘does .* know how’

‘does anyone .* how’

‘does anyone know .*’

For each fillerLook for patterns in candidate pool Sc-Functional compatibility of filler- words used in similar semantic contexts

- Empirical support for filler

‘does anyone know how’

For every known Information Seeking pattern in Sis generate set of filler patterns

Extracting and Scoring Patterns - Example

Known Information Seeking patterns Sis = {‘does anyone know how’, ‘where do I find’, ‘someone tell me where’}

pis from Sis = `does anyone know how’

Match ‘does * know how’ withpatterns in the Candidate Pool

‘does someone know how’ ▪ Functional Compatibility- Impersonal pronouns▪ Empirical Support – 1/3

‘does somebody know how’▪ Functional Compatibility - Impersonal pronouns▪ Empirical Support – 0▪ Pattern still retained – there might be support for somebody later on in the iterative

process

‘does john know how’▪ Pattern discarded

Functional Compatibility from a subset of LIWC1

-Cognitive mechanical (e.g., if, whether, wondering, find) ‘I am thinking about getting X’-Adverbs (e.g., how, somehow, where)-Impersonal pronouns (e.g., someone, anybody, whichever) ‘Someone tell me where can I find X’

1Linguistic Inquiry Word Count,LIWC, http://liwc.net

Other details in the paper..

Over iterations, single-word substitutions, functional usage and empirical support conservatively expands Sis

Infusing new patterns and seed words

Stopping conditions

Sample Extracted Patterns

does anyone know how

anyone know how to

i dont know what

know where i can

tell me how to

i dont know how

anyone know where i

does anyone know where

does anyone know what

anybody know how to

anyone know how i

im not sure what

does anybody know how

does anyone know why

i was wondering how

does anyone know when

tell me what to

im not sure how

i was wondering what

no idea how to

someone tell me how

have no clue what

does anyone know if

i dont know if

know if i can

anyone know if i

im not sure if

i was wondering if

idea what you are

let me know how

and i dont know

now i dont know

but i dont really

was wondering if someone

would like to see

see what i can

anyone have any idea

wondering if someone could

was wondering how i

i do not want

Identifying the Monetization Potential of a new post

Information Seeking patterns generated offline

Monetization Potential of a post calculated by Finding its Information Seeking score : Extracting

and comparing patterns in posts with extracted patterns +

Finding its Transactional Intent Score: Using the LIWC ‘Money’ dictionary ▪ 173 words and word forms indicative of transactions,

e.g., trade, deal, buy, sell, worth, price etc.

Benchmarking with Facebook Marketplace

Using a training corpus of 8000 user posts MySpace Computers, Electronics, Gadgets forum Generated 309 unique new Information Seeking

patterns

Test Set: Using 3 sets of 150 posts each from Facebook ‘to buy’ Marketplace All these posts have Information Seeking and

Transactional intents 81 % of these posts were identified as monetizable in

nature using our algorithm

Validates usefulness of action patterns

Identifying KeywordsOff-topic Noise Elimination from posts with Monetization Potential

Identifying Keywords for Advertizing

Identifying keywords in monetizable posts Plethora of work in this space

Off-topic noise removal is our focusI NEED HELP WITH SONY VEGAS PRO 8!! Ugh and i have a video project due tomorrow for merrill lynch :(( all i need to do is simple: Extract several scenes from a clip, insert captions, transitions and thats it. really. omgg i cant figure out anything!! help!! and i got food poisoning from eggs. its not fun. Pleasssse, help? :(

Conceptual Overview – Details in Paper

Topical hints C1 - ['camcorder']

Keywords in post C2 - ['electronics forum', 'hd', 'camcorder', 'somethin',

'ive', 'canon', 'little camera', 'canon hv20', 'cameras', 'offtopic']

Move strongly related keywords from C2 to C1 Relatedness determined using concepts of information

gain Counts from Web as a corpus Makes for a domain independent solution

Off-topic Chatter - Example

C1 - ['camcorder'] C2 - ['electronics forum', 'hd',

'camcorder', 'somethin', 'ive', 'canon', 'little camera', 'canon hv20', 'cameras', 'offtopic']

Informative words ['camcorder', 'canon hv20', 'little camera', 'hd',

'cameras', 'canon']

EvaluationsOngoing Work

What are we evaluating..

Ideally, we would like to deploy on SNSs and observe click throughs

Approximating with subgoals

1. Effectiveness of using topical keywords instead of entire post content

2. Effectiveness of using user generated content on SNSs instead of profile (homepage) information

User Study – Set Up

Keywords from 60 picked monetizable user posts 45 MySpace Forums, 15 Facebook

Marketplace split into 10 sets of 6 posts each

30 graduate students, each set of 6 posts evaluated by 3 randomly selected users

1. Effectiveness of using topical keywords

Google AdSense ads for user post content vs. extracted topical keywords

Instructions – Example

Choose relevant Ad Impressions

VW 6 disc CD changer I need one thats compatible with a

2000 golf most are sold from years 1998-2004if anyone has one [or can get one] PLEASE let me know!

Result - 2X Relevant Impressions

Users picked ads relevant to the post At least 50% inter-evaluator agreement

For the 60 posts based on content Total of 144 ad impressions 17% of ads picked as relevant

For the topical keywords Total of 162 ad impressions 40% of ads picked as relevant

2. Profile Ads vs. Activity Ads

User’s profile information Interests, hobbies, tv shows.. Non-demographic information

Submit a post Looking to buy and why (induced noise)

Qsn asked: Select ads that generate interest, captured attention

Result - 8X Generated Interest Using profile ads

Total of 56 ad impressions 7% of ads generated interest

Using user submitted posts (entire content, already monetizable) Total of 56 ad impressions 43% of ads generated interest

Using topical keywords from submitted posts Total of 59 ad impressions 59% of ads generated interest

To note…

User studies small and results preliminary, but clearly suggest Monetization potential in user activity Improvement for Ad programs in terms of

relevant impressions

Evaluations based on forum, marketplace Verbose content May not work as well for micro-blog like

content, status updates etc.

To note…

A world between relevant impressions and clickthroughs Objectionable content, vocabulary

impedance, Ad placement, network behavior Our works fits in a pipeline of other

community efforts

No profile information taken into account Cannot custom send information to Google

AdSense

Thank you

Social Media Content Analysis @ Kno.e.sis Google/Bing: Meena Nagarajan

meena@knoesis.org http://knoesis.wright.edu/students/meena/

Google/Bing: Amit Sheth amit@knoesis.org http://knoesis.org/amit

Sponsors: NSF (Semantic Discovery - SemDis), IBM UIMA Innovation Award 2007: "UIMA-based Infrastructure for Summarizing Casual,

Unstructured Text”, Microsoft's Beyond Search - Semantic Computing and Internet Economics Award 2008: Chatter, Intent and Good Karma for Targeted Advertising in Social Networks