Transcript PPT - Institute for Web Science and Technologies

Web Science & Technologies
University of Koblenz ▪ Landau, Germany
Online Advertising
Steffen Staab
Topics
 Introduction to online advertisement
 Understanding the participants and their roles.
 Targeted advertising.
 Privacy Issues
 Solutions
 User based solutions
 Collaborative solutions
 Conclusions
Introduction to Web Science
Slide 2 of 71
http://west.uni-koblenz.de
Introduction
 Online Advertising plays a critically important role in
the Internet world.
 advertising is the main way of profiting from the
Internet, the history of Internet advertising developed
alongside the growth of the medium itself
Introduction to Web Science
Slide 3 of 71
http://west.uni-koblenz.de
Facts and short history
 First internet banner, 1994, AT&T.
 Also in 1994, the first commercial spam, a "Green Card
Lottery".
 The first ad server was developed by FocaLink Media
Services and introduced on 1995.
 In March 2008, Google acquired DoubleClick for US$3.1
billion in cash.
Introduction to Web Science
Slide 4 of 71
http://west.uni-koblenz.de
Parties
 Advertiser
 Got money, wants publicity
 e.g., Coca-Cola
 Publisher
 Got content, wants money
 Cnn.com
 Ad-network
 Got advertising infrastructure, wants money
 e.g., Google AdSense, Yahoo
 Consumer
 Wants free content
Introduction to Web Science
Slide 5 of 71
http://west.uni-koblenz.de
Ad embedding
Introduction to Web Science
Slide 6 of 71
http://west.uni-koblenz.de
Business Model
 CPM = Cost Per thousand impressions
Impression: user just sees the ad.
Rates vary from $0.25 to $100
 CPC = Cost Per Click
This is the cost charged to an advertiser
every time their ad is "clicked" on
Rates around 0.3$ per click
Introduction to Web Science
Slide 7 of 71
http://west.uni-koblenz.de
Click fraud
 clicking on an ad for the purpose of generating a charge
per click without having actual interest.
 Might be:
 The publisher
 Advertiser’s competitor
 The publisher’s competitor
 Ad-networks deal with it by trying to identify who clicks on
the ads.
Introduction to Web Science
Slide 8 of 71
http://west.uni-koblenz.de
Online Advertising and Ad
Auctions at Google
Vahab Mirrokni
Google Research, New York
Introduction to Web Science
Slide 9 of 71
http://west.uni-koblenz.de
Traditional Advertising

At the beginning: Traditional Ads
 Posters, Magazines, Newspapers, Billboards.

What is being Sold:
 Pay-per-Impression: Price depends on how many people
your ad is shown to (whether or not they look at it)

Pricing:

Complicated Negotiations (with high monthly premiums...)

Form a barrier to entry for small advertisers
Introduction to Web Science
Slide 10 of 71
http://west.uni-koblenz.de
Advertising on the Web

Online Ads:

Banner Ads, Sponsored Search Ads, Pay-per-Sale ads.
Targeting:


Show to particular set of viewers.
Measurement:
 Accurate Metrics: Clicks, Tracked Purchases.
What is being Sold:

Pay-per-Click, Pay-per-Action, Pay-per-Impression
Pricing:




Auctions
Introduction to Web Science
Slide 11 of 71
http://west.uni-koblenz.de
History of Online Advertising
1994: Banner ads,
pay-per-impression
1998: Sponsored search,
pay-per-click 1st-price auction
Banner ads for Zima
and AT&T appear on
hotwired.com.
GoTo.com develops keywordbased advertising with pay-perclick sales.
1996: Affiliate marketing,
pay-per-acquisition
2002: Sponsored search,
pay-per-click 2nd-price auction
Amazon/EPage/CDNow
pay hosts for sales generated
through ads on their sites.
Google introduces AdWords, a
second-price keyword auction with
a number of innovations.
Introduction to Web Science
Slide 12 of 71
http://west.uni-koblenz.de
Banner Ads
Introduction to Web Science
Slide 13 of 71
http://west.uni-koblenz.de
Pay-Per-Impression

Pay-per-1000 impressions (PPM): advertiser pays each time
ad is displayed




Exposes advertiser to risk of fluctuations in market


Models existing standards from magazine, radio, television
Main business model for banner ads to date
Corresponds to inventory host sells
Banner blindness: effectiveness drops with user experience
Barrier to entry for small advertisers

Contracts negotiated on a case-by-case basis with large minimums
(typically, a few thousand dollars per month)
Introduction to Web Science
Slide 14 of 71
http://west.uni-koblenz.de
Pay-PerClick

Pay-per-click (PPC): advertiser pays only when user
clicks on ad




Common in search advertising
Middle ground between PPM and PPA
Does not require host to trust advertiser
Provides incentives for host to improve ad displays
Introduction to Web Science
Slide 15 of 71
http://west.uni-koblenz.de
Auction Mechanism

Advertisements sold automatically through
auctions: advertisers submit bids indicating value
for clicks on particular keywords



Low barrier-to-entry
Increased transparency of mechanism
Keyword bidding allowed increased targeting
opportunities
Introduction to Web Science
Slide 16 of 71
http://west.uni-koblenz.de
Auction Mechanism

Initial GoTo model: first-price auction




Advertisers displayed in order of decreasing bids
Upon a click, advertiser is charged a price equal to his bid
Used first by Overture/Yahoo!
Google model: stylized second-price auction


Advertisers ranked according to bid and click-throughrate (CTR), or probability user clicks on ad
Upon a click, advertiser is charged minimum amount
required to maintain position in ranking
Introduction to Web Science
Slide 17 of 71
http://west.uni-koblenz.de
Bidding Patterns

Bidding history in Yahoo! First-Price Auction:
Graph from [Zhang 2006]
Introduction to Web Science
Slide 18 of 71
http://west.uni-koblenz.de
Bidding Patterns
Graph from [Zhang 2006]
Introduction to Web Science
Slide 19 of 71
http://west.uni-koblenz.de
Bidding Process
Targeting
Populations
Advert
Creation
Keyword
Selection
Bids and
Budget
1
2
3
4
“You don’t get it, Daddy, because they’re not targeting you.”
Introduction to Web Science
Slide 20 of 71
http://west.uni-koblenz.de
Introduction to Web Science
Slide 21 of 71
http://west.uni-koblenz.de
Introduction to Web Science
Slide 22 of 71
http://west.uni-koblenz.de
Introduction to Web Science
Slide 23 of 71
http://west.uni-koblenz.de
Introduction to Web Science
Slide 24 of 71
http://west.uni-koblenz.de
Bidding Process
Targeting
Populations
Advert
Creation
Keyword
Selection
Bids and
Budget
1
2
3
4
“Here it is – the plain unvarnished truth. Varnish it.”
Introduction to Web Science
Slide 25 of 71
http://west.uni-koblenz.de
Display url
Ad title
Ad text
Introduction to Web Science
Slide 26 of 71
http://west.uni-koblenz.de
Bidding Process
Targeting
Populations
Advert
Creation
Keyword
Selection
Bids and
Budget
“Now, that’s product placement!”
Introduction to Web Science
Slide 27 of 71
http://west.uni-koblenz.de
1
2
3
4
Introduction to Web Science
Slide 28 of 71
http://west.uni-koblenz.de
Introduction to Web Science
Slide 29 of 71
http://west.uni-koblenz.de
Introduction to Web Science
Slide 30 of 71
http://west.uni-koblenz.de
Bidding Process
Targeting
Populations
Advert
Creation
Keyword
Selection
Bids and
Budget
Introduction to Web Science
Slide 31 of 71
http://west.uni-koblenz.de
1
2
3
4
Daily Budget
Introduction to Web Science
Slide 32 of 71
http://west.uni-koblenz.de
Introduction to Web Science
Slide 33 of 71
http://west.uni-koblenz.de
Auction Mechanism


A repeated mechanism!
Upon each search,





Interested advertisers are selected from database using
keyword matching algorithm
Budget allocation algorithm retains interested advertisers
with sufficient budget
Advertisers compete for ad slots in allocation mechanism
Upon click, advertiser charged with pricing scheme
CTR updated according to CTR learning algorithm
for future auctions
Introduction to Web Science
Slide 34 of 71
http://west.uni-koblenz.de
Click-Through Rates




Click-through rate (CTR): a parameter estimating
the probability that a user clicks on an ad
A separate parameter for each ad/keyword pair
Assumption: CTR of an ad in a slot is equal to the
CTR of the ad in slot 1 times a scaling parameter
which depends only on the slot and not the ad
CTR learning algorithm uses a weighted averaging
of past performance of ad to estimate CTR
Introduction to Web Science
Slide 35 of 71
http://west.uni-koblenz.de
(Old) Yahoo! 2nd-Price Auction
Keywor
d
Ad slot
1
Ad slot
2
Algorithmic
search
results
Advertiser
Bid
Allocation
Price
per click!
A
B
$10
$5
2
X
$5
$0
C
$50
1
$10
Introduction to Web Science
Slide 36 of 71
http://west.uni-koblenz.de
Google Single-Shot Auction
Keywor
d
Ad slot
1
Ad slot
2
Algorithmic
search
results
Bid
CTR
A
B
$10
$5
0.10
0.50
1.0
2.5
2
1
$5
$2
C
$50
0.01
0.5
X
$0
Introduction to Web Science
Bid x CTR
(expected bid per
Allocation
Price
impression)
per click!
Advertiser
Slide 37 of 71
http://west.uni-koblenz.de
Keyword Matching




Exact match: keyword phrase equals search phrase
Phrase match: keyword phrase appears in search
(“red roses” matches to “red roses for valentines”)
Broad match: each word of keyword phrase appears
in search (“red roses” matches to “red and white
roses”)
Issues:


Tradeoff between relevance and competition
How to handle spelling mistakes
Introduction to Web Science
Slide 38 of 71
http://west.uni-koblenz.de
Budget Allocation

Basic algorithm




Spread monthly budget evenly over each day
If budget leftover at end of day, allocate to next day
When advertiser runs out of budget, eliminate from
auctions
Issues:


Need to smooth allocation through-out day
Allocation of budget across keywords
Introduction to Web Science
Slide 39 of 71
http://west.uni-koblenz.de
Typical Parameters

PPC of most popular searches in Google,
4/06
Keyword Price in 3rd slot
# of Keywords
$20-$50
2
$10.00 - $19.99
22
$5.00 - $9.99
206
$3.00 - $4.99
635
$1.00 - $2.99
3,566
$0.50 - $0.99
4,946
$0.25 - $0.49
5,501
$0.11 - $0.24
5,269
Introduction to Web Science
Slide 40 of 71
http://west.uni-koblenz.de
Typical Parameters


Bids on some valuable keywords
Keyword
Top Bid
2nd Bid
mesothelioma
$100
$100
structured settlement
$100
$52
vioxx attorney
$38
$38
student loan consolidation
$29
$9
CTRs are typically around 1%
Introduction to Web Science
Slide 41 of 71
http://west.uni-koblenz.de
Typical Parameters




Avoiding click fraud
Bidding with budget constraints
Externalities between advertisers
User search models
Introduction to Web Science
Slide 42 of 71
http://west.uni-koblenz.de
Measurement: Information

Adwords FrontEnd: Bid Simulations


Google Analytics


Traffic Patterns, Site visitors.
Search insights:


Clicks and Cost for other bids.
Search Patterns
Interest-Based Advertising

Indicate your interests so that you get more relevant ads
Introduction to Web Science
Slide 43 of 71
http://west.uni-koblenz.de
AdWords FrontEnd
Introduction to Web Science
Slide 44 of 71
http://west.uni-koblenz.de
Introduction to Web Science
Slide 45 of 71
http://west.uni-koblenz.de
Web Analytics
Introduction to Web Science
Slide 46 of 71
http://west.uni-koblenz.de
Re-acting to Metrics

Distinguish Causality and Correlation.

Experimentation:
 Ad Rotation: 3 different creatives
 Website Optimizer
 E.g. 6000 search quality experiments, 500 of
which were launched.

Repeated experimentation:
 Continuous Improvement (Multi-armed bandit)
Introduction to Web Science
Slide 47 of 71
http://west.uni-koblenz.de
47
Other Online Advertising Aspects

Google Ad Systems:





Bid Management & Campaign Optimization for Advertisers


Short-term vs. Long-term effect of ads.
Planning: Ad Auctions & Ad Reservations.
 Stochastic/Dynamic Inventory Planning


Sponsored Search: AdWord Auctions.
Contextual Ads (AdSense) & Display Ads (DoubleClick)
Ad Exchange
Social Ads, YouTube, TV ads.
Pricing: Auctions vs Contracts
Ad Serving

Online Stochastic Assignment Problems
Introduction to Web Science
Slide 48 of 71
http://west.uni-koblenz.de
48
Ad Serving



Efficiency, Fairness, Smoothness.
Sponsored Search: Repeated Auctions, Budget
Constraints, Throttling, Dynamics(?)
Display Ads: Online Stochastic Allocation

Impressions arrive online, and should be assigned to
Advertisers (with established contracts)





Online Primal-Dual Algorithms.
Offline Optimization for Online Stochastic Optimization: Power
of Two Choices.
Learning+Optimization: Exploration vs Exploitation??
Ad Exchange Ad Serving: Bandwidth Constraints.
Social Ads: Ad Serving over Social Networks
Introduction to Web Science
Slide 49 of 71
http://west.uni-koblenz.de
49
TARGETED ONLINE ADVERTISING
Itay Gonshorovitz
Foundation of privacy
Introduction to Web Science
Slide 50 of 71
http://west.uni-koblenz.de
Online behavioral advertising
 Online behavioral advertising refers to the practice of adnetworks tracking users across web sites in order to learn
user interests and preferences.
 Benefits
 Advertisers targets a more focused audience
which increases the effectively.
 Consumer is “bothered” by more relevant and
interesting ads.
Introduction to Web Science
Slide 51 of 71
http://west.uni-koblenz.de
How ad-networks match ads
 Most behavioral targeting systems work by categorizing
users into one or more audience segments.
 Profiling users based on collected data
 Search history – analyzing search keywords
 Browse history - analyzing content of visited pages
 Purchase history
 Social networks
 Geography
Introduction to Web Science
Slide 52 of 71
http://west.uni-koblenz.de
How Ad-Networks track users
 Cookies
 3rd Party cookies
 Flash cookies
 Web bug
 IP address
 User-agent Headers
 Browser + OS
 More than 24,000 signatures
Introduction to Web Science
Slide 53 of 71
http://west.uni-koblenz.de
Levis.com case study
Introduction to Web Science
Slide 54 of 71
http://west.uni-koblenz.de
Levis.com case study
Introduction to Web Science
Slide 55 of 71
http://west.uni-koblenz.de
Privacy
 Tracking and categorizing users by the ad-networks tend to
violate user’s privacy.
 The gathered information, linked with the users real identity,
form a violation of privacy in its most basic form.
 For example, if a person is searching the web for
information on a serious genetic disease, that information
can be collected and stored along with that consumer's other
information - including information that can uniquely identify
the consumer.
Introduction to Web Science
Slide 56 of 71
http://west.uni-koblenz.de
So… What we have so far?
 User - Preserve his privacy
 Ad-Network & Publisher –
 Maintain targeting and preserve their effectiveness and
income
 Still want to be able to fight click fraud
 Questions:
 Do the two goals necessarily conflict?
 Or can they be both achieved?
Introduction to Web Science
Slide 57 of 71
http://west.uni-koblenz.de
Naive (paranoid) solution
 Surf only across anonymizing proxies.
 TOR
 Surf in private mode
 Advantages
 Effective from the user’s perspective.
 Disadvantages
 Are proxies really anonymizing?
 Very awkward
 Slower
 Damages targeted advertising
Introduction to Web Science
Slide 58 of 71
http://west.uni-koblenz.de
TrackMeNot (Howe, Nissenbaum, 2005)




Implemented as a Firefox plugin.
Achieves privacy through obfuscation.
Generates noisy queries.
Starts with fixed a seed query list and evolve queries base
on previous results.
 Mimics user behavior so fake queries be indistinguishable:
 Query timing
 Click through behavior
Introduction to Web Science
Slide 59 of 71
http://west.uni-koblenz.de
TrackMeNot
 Advantages
 Simple
 Disadvantages
 Still the real queries can be connected to real identity.
 Might have problems with offensive contents.
 Again, damages targeted advertising
Introduction to Web Science
Slide 60 of 71
http://west.uni-koblenz.de
Privad (Guha, Reznichenko, Tang , et al., 2009)
 Require client software:
 saves locally database of ads (served by the
ad-network)
 Learn user interests in order to match ads.
 Match add from the local database according
to the User interests.
Introduction to Web Science
Slide 61 of 71
http://west.uni-koblenz.de
Privad
 Introduce new party – Dealer:
 Proxies anonymously all communication between
the user and the ad-network.
 might be government regulatory agency.
 hides user’s identity from the ad-network, but
itself does not learn any profile information
about the user since all messages between the
user and ad-network are encrypted.
Introduction to Web Science
Slide 62 of 71
http://west.uni-koblenz.de
Introduction to Web Science
Slide 63 of 71
http://west.uni-koblenz.de
Privad
 Advantages
 Ad-Networks can still target ads without violates user
privacy.
 Disadvantages
 Complicated to add the new party.
 Ad-Network has to trust the dealer in order to fight clickfraud which might unmotivated them to cooperate.
Introduction to Web Science
Slide 64 of 71
http://west.uni-koblenz.de
Adnostic (Toubina, Narayanan, Boneh, et al., 2009)
 Two party solution:
 Client side: Implemented as a Firefox plugin.
 Server side: requires Ad-Network support
 User’s preferences and interests are stored locally by the
plugin, instead of at the Ad-network.
 The targeted ad is selected by the plugin locally at the
users computer, instead of at the Ad-Network servers.
Introduction to Web Science
Slide 65 of 71
http://west.uni-koblenz.de
Adnostic - Accounting
 “charge per click” model remains unchanged.
 “charge per impression” is harder.
 It uses homomorphic encryption scheme.
 given the public key
and ciphertexts
, anyone can calculate
 given the public key
and ciphertexts
, and scalar
c,
can be calculated.
Introduction to Web Science
Slide 66 of 71
http://west.uni-koblenz.de
Adnostic - charge per impression protocol
Client: Track user activity and maintains the
data locally.
Visits an Ad supported website.
Server: Sends a list of n ads ids along with
public key
The browser chooses an ad to display to the
user.
Then creates
that matches the
selected ad, then send
, Along with zero-knowledge proof that
and each is 0 or 1.
Introduction to Web Science
Slide 67 of 71
http://west.uni-koblenz.de
Adnostic - charge per impression protocol
Validates the proof. If the proof is valid then using
homomorphic encryption calculates
when c is the price of viewing the ad.
The server save encrypted counter for each ad and
add to it the previous values. Only one counter’s real
value change.
 At the end of the billing period, say a month, each
counter is decrypted (should be done by trusted
authority) and the advertisers pays for the adnetwork.
Introduction to Web Science
Slide 68 of 71
http://west.uni-koblenz.de
Adnostic
 Advantages
 Ad-networks can still target ads without violates user
privacy.
 Ad-networks can still detect click fraud though it will be
difficult without gathering information on IP even for a short
time.
 Disadvantages
 Ad-networks become weaker.
 Ad-networks can still track user if they are willing to, and
the protocol is built on trust.
Introduction to Web Science
Slide 69 of 71
http://west.uni-koblenz.de
Future of Online Advertising




Measurements
Pricing
Experimentation
Other form of Advertising:
 TV Ads
 Ad Exchanges
 Social Ads
Introduction to Web Science
Slide 70 of 71
http://west.uni-koblenz.de
Conclusions
 In my opinion, It is hard to believe that ad-networks will
give up the power of tracking users without legislation.
 Nevertheless, There are reasonable solutions that still
support targeted advertising without violating users privacy.
Introduction to Web Science
Slide 71 of 71
http://west.uni-koblenz.de