Online Advertising€¦ · $ Online Ads: # Banner Ads, Sponsored Search Ads, Pay-per-Sale ads. $...
Transcript of Online Advertising€¦ · $ Online Ads: # Banner Ads, Sponsored Search Ads, Pay-per-Sale ads. $...
Web Science & Technologies University of Koblenz ▪ Landau, Germany
Online Advertising
Steffen Staab
Introduction to Web Science Slide 2 of 71 http://west.uni-koblenz.de
Topics
§ Introduction to online advertisement w Understanding the participants and their roles. w Targeted advertising.
§ Privacy Issues § Solutions
w User based solutions w Collaborative solutions
§ Conclusions
Introduction to Web Science Slide 3 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 4 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 5 of 71 http://west.uni-koblenz.de
Parties
§ Advertiser w Got money, wants publicity w e.g., Coca-Cola
§ Publisher w Got content, wants money w Cnn.com
§ Ad-network w Got advertising infrastructure, wants money w e.g., Google AdSense, Yahoo
§ Consumer w Wants free content
Introduction to Web Science Slide 6 of 71 http://west.uni-koblenz.de
Ad embedding
Introduction to Web Science Slide 7 of 71 http://west.uni-koblenz.de
Business Model
q CPM = Cost Per thousand impressions ¤ Impression: user just sees the ad. ¤ Rates vary from $0.25 to $100
q CPC = Cost Per Click ¤ This is the cost charged to an advertiser
every time their ad is "clicked" on ¤ Rates around 0.3$ per click
Introduction to Web Science Slide 8 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: w The publisher w Advertiser’s competitor w The publisher’s competitor
§ Ad-networks deal with it by trying to identify who clicks on the ads.
Introduction to Web Science Slide 9 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 10 of 71 http://west.uni-koblenz.de
n At the beginning: Traditional Ads q Posters, Magazines, Newspapers, Billboards.
n What is being Sold:
q Pay-per-Impression: Price depends on how many people your ad is shown to (whether or not they look at it)
n Pricing: q Complicated Negotiations (with high monthly premiums...)
q Form a barrier to entry for small advertisers
Traditional Advertising
Introduction to Web Science Slide 11 of 71 http://west.uni-koblenz.de
n Online Ads:
q Banner Ads, Sponsored Search Ads, Pay-per-Sale ads. n Targeting:
q Show to particular set of viewers. n Measurement:
q Accurate Metrics: Clicks, Tracked Purchases. n What is being Sold:
q Pay-per-Click, Pay-per-Action, Pay-per-Impression n Pricing:
q Auctions
Advertising on the Web
Introduction to Web Science Slide 12 of 71 http://west.uni-koblenz.de
1994: Banner ads, pay-per-impression
Banner ads for Zima and AT&T appear on hotwired.com.
1998: Sponsored search, pay-per-click 1st-price auction
GoTo.com develops keyword-based advertising with pay-per-click sales.
2002: Sponsored search, pay-per-click 2nd-price auction
Google introduces AdWords, a second-price keyword auction with a number of innovations.
1996: Affiliate marketing, pay-per-acquisition
Amazon/EPage/CDNow pay hosts for sales generated through ads on their sites.
History of Online Advertising
Introduction to Web Science Slide 13 of 71 http://west.uni-koblenz.de
Banner Ads
Introduction to Web Science Slide 14 of 71 http://west.uni-koblenz.de
n Pay-per-1000 impressions (PPM): advertiser pays each time ad is displayed q Models existing standards from magazine, radio, television q Main business model for banner ads to date q Corresponds to inventory host sells
n Exposes advertiser to risk of fluctuations in market q Banner blindness: effectiveness drops with user experience
n Barrier to entry for small advertisers q Contracts negotiated on a case-by-case basis with large minimums
(typically, a few thousand dollars per month)
Pay-Per-Impression
Introduction to Web Science Slide 15 of 71 http://west.uni-koblenz.de
n Pay-per-click (PPC): advertiser pays only when user clicks on ad q Common in search advertising q Middle ground between PPM and PPA
n Does not require host to trust advertiser n Provides incentives for host to improve ad displays
Pay-PerClick
Introduction to Web Science Slide 16 of 71 http://west.uni-koblenz.de
n Advertisements sold automatically through auctions: advertisers submit bids indicating value for clicks on particular keywords q Low barrier-to-entry q Increased transparency of mechanism
n Keyword bidding allowed increased targeting opportunities
Auction Mechanism
Introduction to Web Science Slide 17 of 71 http://west.uni-koblenz.de
n Initial GoTo model: first-price auction q Advertisers displayed in order of decreasing bids q Upon a click, advertiser is charged a price equal to his bid q Used first by Overture/Yahoo!
n Google model: stylized second-price auction q Advertisers ranked according to bid and click-through-
rate (CTR), or probability user clicks on ad q Upon a click, advertiser is charged minimum amount
required to maintain position in ranking
Auction Mechanism
Introduction to Web Science Slide 18 of 71 http://west.uni-koblenz.de
Graph from [Zhang 2006]
n Bidding history in Yahoo! First-Price Auction:
Bidding Patterns
Introduction to Web Science Slide 19 of 71 http://west.uni-koblenz.de
Graph from [Zhang 2006]
Bidding Patterns
Introduction to Web Science Slide 20 of 71 http://west.uni-koblenz.de
4
Targeting Populations
Advert Creation
Keyword Selection
Bids and Budget
3 2 1
“You don’t get it, Daddy, because they’re not targeting you.”
Bidding Process
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
Introduction to Web Science Slide 25 of 71 http://west.uni-koblenz.de
4
Targeting Populations
Advert Creation
Keyword Selection
Bids and Budget
“Here it is – the plain unvarnished truth. Varnish it.”
3 2 1
Bidding Process
Introduction to Web Science Slide 26 of 71 http://west.uni-koblenz.de
Ad title Ad text
Display url
Introduction to Web Science Slide 27 of 71 http://west.uni-koblenz.de
4
Targeting Populations
Advert Creation
Keyword Selection
Bids and Budget
“Now, that’s product placement!”
3 2 1
Bidding Process
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
Introduction to Web Science Slide 31 of 71 http://west.uni-koblenz.de
4
Targeting Populations
Advert Creation
Keyword Selection
Bids and Budget
3 2 1
Bidding Process
Introduction to Web Science Slide 32 of 71 http://west.uni-koblenz.de
Daily Budget
Introduction to Web Science Slide 33 of 71 http://west.uni-koblenz.de
Introduction to Web Science Slide 34 of 71 http://west.uni-koblenz.de
n A repeated mechanism! n Upon each search,
q Interested advertisers are selected from database using keyword matching algorithm
q Budget allocation algorithm retains interested advertisers with sufficient budget
q Advertisers compete for ad slots in allocation mechanism q Upon click, advertiser charged with pricing scheme
n CTR updated according to CTR learning algorithm for future auctions
Auction Mechanism
Introduction to Web Science Slide 35 of 71 http://west.uni-koblenz.de
n Click-through rate (CTR): a parameter estimating the probability that a user clicks on an ad
n A separate parameter for each ad/keyword pair n 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
n CTR learning algorithm uses a weighted averaging of past performance of ad to estimate CTR
Click-Through Rates
Introduction to Web Science Slide 36 of 71 http://west.uni-koblenz.de
Advertiser
A
B
C
Bid Allocation Price
$10 2 $5
$5
$50
X
1
$0
$10
per click!
Ad slot 1
Ad slot 2
Keyword
Algorithmic search results
(Old) Yahoo! 2nd-Price Auction
Introduction to Web Science Slide 37 of 71 http://west.uni-koblenz.de
Advertiser
A
B
C
Bid CTR Bid x CTR Allocation Price
$10 0.10 1.0 2 $5
$5
$50
0.50
0.01
2.5
0.5
1
X
$2
$0
(expected bid per impression)
per click!
Ad slot 1
Ad slot 2
Keyword
Algorithmic search results
Google Single-Shot Auction
Introduction to Web Science Slide 38 of 71 http://west.uni-koblenz.de
n Exact match: keyword phrase equals search phrase n Phrase match: keyword phrase appears in search
(“red roses” matches to “red roses for valentines”) n Broad match: each word of keyword phrase appears
in search (“red roses” matches to “red and white roses”)
n Issues: q Tradeoff between relevance and competition q How to handle spelling mistakes
Keyword Matching
Introduction to Web Science Slide 39 of 71 http://west.uni-koblenz.de
n Basic algorithm q Spread monthly budget evenly over each day q If budget leftover at end of day, allocate to next day q When advertiser runs out of budget, eliminate from
auctions
n Issues: q Need to smooth allocation through-out day q Allocation of budget across keywords
Budget Allocation
Introduction to Web Science Slide 40 of 71 http://west.uni-koblenz.de
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
n PPC of most popular searches in Google, 4/06
Typical Parameters
Introduction to Web Science Slide 41 of 71 http://west.uni-koblenz.de
Keyword Top Bid 2nd Bid mesothelioma $100 $100
structured settlement $100 $52
vioxx attorney $38 $38
student loan consolidation $29 $9
n Bids on some valuable keywords n CTRs are typically around 1%
Typical Parameters
Introduction to Web Science Slide 42 of 71 http://west.uni-koblenz.de
n Avoiding click fraud n Bidding with budget constraints n Externalities between advertisers n User search models
Typical Parameters
Introduction to Web Science Slide 43 of 71 http://west.uni-koblenz.de
n Adwords FrontEnd: Bid Simulations q Clicks and Cost for other bids.
n Google Analytics q Traffic Patterns, Site visitors.
n Search insights: q Search Patterns
n Interest-Based Advertising q Indicate your interests so that you get more relevant ads
Measurement: Information
Introduction to Web Science Slide 44 of 71 http://west.uni-koblenz.de
AdWords FrontEnd
Introduction to Web Science Slide 45 of 71 http://west.uni-koblenz.de
Introduction to Web Science Slide 46 of 71 http://west.uni-koblenz.de
Web Analytics
Introduction to Web Science Slide 47 of 71 http://west.uni-koblenz.de
47
n Distinguish Causality and Correlation. n Experimentation:
q Ad Rotation: 3 different creatives q Website Optimizer q E.g. 6000 search quality experiments, 500 of
which were launched.
n Repeated experimentation: q Continuous Improvement (Multi-armed bandit)
Re-acting to Metrics
Introduction to Web Science Slide 48 of 71 http://west.uni-koblenz.de
48
n Google Ad Systems: q Sponsored Search: AdWord Auctions. q Contextual Ads (AdSense) & Display Ads (DoubleClick) q Ad Exchange q Social Ads, YouTube, TV ads.
n Bid Management & Campaign Optimization for Advertisers q Short-term vs. Long-term effect of ads.
n Planning: Ad Auctions & Ad Reservations. q Stochastic/Dynamic Inventory Planning q Pricing: Auctions vs Contracts
n Ad Serving q Online Stochastic Assignment Problems
Other Online Advertising Aspects
Introduction to Web Science Slide 49 of 71 http://west.uni-koblenz.de
49
n Efficiency, Fairness, Smoothness. n Sponsored Search: Repeated Auctions, Budget
Constraints, Throttling, Dynamics(?) n Display Ads: Online Stochastic Allocation
q Impressions arrive online, and should be assigned to Advertisers (with established contracts) n Online Primal-Dual Algorithms. n Offline Optimization for Online Stochastic Optimization: Power
of Two Choices.
q Learning+Optimization: Exploration vs Exploitation??
n Ad Exchange Ad Serving: Bandwidth Constraints. n Social Ads: Ad Serving over Social Networks
Ad Serving
Introduction to Web Science Slide 50 of 71 http://west.uni-koblenz.de
Itay Gonshorovitz Foundation of privacy
TARGETED ONLINE ADVERTISING
Introduction to Web Science Slide 51 of 71 http://west.uni-koblenz.de
Online behavioral advertising
§ Online behavioral advertising refers to the practice of ad-networks tracking users across web sites in order to learn user interests and preferences.
§ Benefits
w Advertisers targets a more focused audience which increases the effectively.
w Consumer is “bothered” by more relevant and interesting ads.
Introduction to Web Science Slide 52 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 w Search history – analyzing search keywords w Browse history - analyzing content of visited pages w Purchase history w Social networks w Geography
Introduction to Web Science Slide 53 of 71 http://west.uni-koblenz.de
How Ad-Networks track users
§ Cookies w 3rd Party cookies w Flash cookies
§ Web bug § IP address § User-agent Headers
w Browser + OS w More than 24,000 signatures
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
Levis.com case study
Introduction to Web Science Slide 56 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 57 of 71 http://west.uni-koblenz.de
So… What we have so far?
§ User - Preserve his privacy § Ad-Network & Publisher –
w Maintain targeting and preserve their effectiveness and income
w Still want to be able to fight click fraud § Questions:
w Do the two goals necessarily conflict? w Or can they be both achieved?
Introduction to Web Science Slide 58 of 71 http://west.uni-koblenz.de
Naive (paranoid) solution
§ Surf only across anonymizing proxies. w TOR
§ Surf in private mode § Advantages
w Effective from the user’s perspective.
§ Disadvantages w Are proxies really anonymizing? w Very awkward w Slower w Damages targeted advertising
Introduction to Web Science Slide 59 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:
w Query timing w Click through behavior
Introduction to Web Science Slide 60 of 71 http://west.uni-koblenz.de
TrackMeNot
§ Advantages w Simple
§ Disadvantages w Still the real queries can be connected to real identity. w Might have problems with offensive contents. w Again, damages targeted advertising
Introduction to Web Science Slide 61 of 71 http://west.uni-koblenz.de
Privad (Guha, Reznichenko, Tang , et al., 2009)
§ Require client software:
w saves locally database of ads (served by the ad-network) w Learn user interests in order to match ads. w Match add from the local database according
to the User interests.
Introduction to Web Science Slide 62 of 71 http://west.uni-koblenz.de
Privad
§ Introduce new party – Dealer:
w Proxies anonymously all communication between the user and the ad-network.
w might be government regulatory agency. w 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 63 of 71 http://west.uni-koblenz.de
Introduction to Web Science Slide 64 of 71 http://west.uni-koblenz.de
Privad
§ Advantages w Ad-Networks can still target ads without violates user
privacy.
§ Disadvantages w Complicated to add the new party. w Ad-Network has to trust the dealer in order to fight click-
fraud which might unmotivated them to cooperate.
Introduction to Web Science Slide 65 of 71 http://west.uni-koblenz.de
Adnostic (Toubina, Narayanan, Boneh, et al., 2009)
§ Two party solution: w Client side: Implemented as a Firefox plugin. w 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 66 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.
w given the public key and ciphertexts , anyone can calculate
w given the public key and ciphertexts , and scalar c, can be calculated.
Introduction to Web Science Slide 67 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 68 of 71 http://west.uni-koblenz.de
Adnostic - charge per impression protocol
q Validates the proof. If the proof is valid then using homomorphic encryption calculates
when c is the price of viewing the ad. q 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 ad-network.
Introduction to Web Science Slide 69 of 71 http://west.uni-koblenz.de
Adnostic
§ Advantages w Ad-networks can still target ads without violates user
privacy. w Ad-networks can still detect click fraud though it will be
difficult without gathering information on IP even for a short time.
§ Disadvantages w Ad-networks become weaker. w Ad-networks can still track user if they are willing to, and
the protocol is built on trust.
Introduction to Web Science Slide 70 of 71 http://west.uni-koblenz.de
n Measurements n Pricing n Experimentation n Other form of Advertising:
q TV Ads q Ad Exchanges q Social Ads
Future of Online Advertising
Introduction to Web Science Slide 71 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.