Taxonomy-based Contextual Ads Targeting

Jan 21, 2010Download as PPSX, PDF3 likes1,735 views

This presentation introduces taxonomy-based contextual ad targeting. It discusses defining taxonomies to semantically classify content and ads, then matching ads to content based on overlapping taxonomy terms. An ad targeting engine is described that uses an optimizer to select ads and a dispatcher to deliver them. The approach was tested against a rule-based system and showed click-through rate improvements for certain product categories.

Taxonomy-based
Contextual Ads Targeting
Patrick Nicolas
Dec 8, 2009
patricknicolas.blogspot.com
www.slideshare.net/pnicolas
https://github.com/prnicolas

Purpose
This presentation is a short introduction
to the different components of an ads
server that leverages semantic analysis
to segment and target audience.

Patrick Nicolas Copyright 2009-2011 - All rights reserved.

Semantics
A few definitions…
Contextual targeting is the process of
inserting the most appropriate advertising
into a published content (web pages, social
network, tweets, blogs..)
Taxonomy is the study or science of
classification of concept or concrete
items, in a logical and repeatable manner.
Patrick Nicolas Copyright 2009-2011 - All rights reserved.

Conundrum
Should targeting relies on
● audience preferences & behavior history
● content topic and style
● both?
Publisher
Content

User
Market
Advertiser
Promotion
Patrick Nicolas Copyright 2009-2011 - All rights reserved.

Taxonomy-based Targeting
It is assumed that the content
consumed by a visitor is reflective of
his/her
interests,
tastes
&
demographic characteristics.
Therefore, targeting (yield) consists of
analyzing content & extracting
context.
Patrick Nicolas Copyright 2009-2011 - All rights reserved.

Architecture
An ads targeting
components:

engine

two

key

● Optimizer
to
balance
objectives
(budget,
volume)
&
constraints
(placements, frequency, exclusivity,..)

● Dispatcher to select the ad with the
highest predicted yield according to
content
Patrick Nicolas Copyright 2009-2011 - All rights reserved.

Use Case
1. Campaign manager defines the objectives &
constraints
2. Optimizer computes the best promotion in the
inventory that satisfy constraints
3. Dispatcher formats & dispatches the promotion
Ads.
inventory
Campaign 1
1
Manager

Optimizer

1
2

1
Dispatcher 3
Content

Patrick Nicolas Copyright 2009-2011 - All rights reserved.

Semantic Analysis
1. The optimizer performs a semantic analysis of
both the content & promotional material.
1. The analysis generates taxonomy or semantic
classification graphs.
1. Finally, the promotion with the taxonomy
graph which is the closest to the content
graph is selected
Patrick Nicolas Copyright 2009-2011 - All rights reserved.

Taxonomy Match
Portability
Music

Device

Travel

Content taxonomy
Content

Matching
Listening
Device
Device
iPod

Promotion

Autonomy

Promotion taxonomy
Patrick Nicolas Copyright 2009-2011 - All rights reserved.

Test Results
The taxonomy match algorithm was
evaluated against a rule-based targeting
engine
for
consumer
discretionary
products.
The Click Through Rate (CTR) increased
from 0.8% to 0.193% (handbags) and
0.052% to 0.98% (upscale pen).
Patrick Nicolas Copyright 2009-2011 - All rights reserved.

References
● How much can Behavioral Targeting Help Online
Advertising? J. Yan, N. Lu, G. Wang, W. Zhang, Y Jiang
http://www2009.eprints.org/27/1/p261.pdf

● Introduction to Semantic Analysis
http://www.cs.tut.fi/sgn/arg/klap/introduction-semantics.pdf

Patrick Nicolas Copyright 2009-2011 - All rights reserved.

This document discusses supply chain analytics and how analyzing big data can help improve forecasting and decision making. It focuses on using social media data to analyze consumer sentiment regarding the beef supply chain. Key points include: 1. Supply chain analytics applies techniques like machine learning to patterns in data to improve forecasting and efficiency. 2. Analyzing social media, point-of-sale, and in-store data on topics like product choices and foot traffic can provide insights for forecasting sales. 3. A case study analyzed tweets about beef/steak, finding issues around food safety, quality, and coordination between suppliers affected consumer satisfaction. 4. Recommendations include developing consumer-centric supply chains and

Supply chain analytics convertednakshatraL

An improved approach to long tail advertising in sponsored searchAmar Budhiraja

This document proposes an approach for long tail advertising in sponsored search. It suggests that advertisers bid on high-level concepts through a taxonomy rather than individual keywords. Coverage patterns are extracted from search query logs and a taxonomy to allocate queries to advertisers based on their bids on concepts. Experiments on AOL search data show this concept-based approach improves ad space utilization and reach compared to keyword-based bidding. Future work plans to analyze the tradeoff between relevance and concept-based bidding, build taxonomies tailored for sponsored search, and consider truthful auctions for concept bidding.

Whatsapp analyticsDipesh Patel

This document discusses analyzing WhatsApp group data from two business groups - Frugetory and White Gold - to gain insights. Data was collected from WhatsApp web using web scraping and analyzed using sentiment analysis, keyword extraction, and topic modeling. The analysis found that most posts in Frugetory were positive while most in White Gold were neutral. It also identified frequent words and topics in each group. The document proposes continuing this analysis could help businesses better understand customers and improve strategies.

Industry updates on key mobile trends 8 09 13Performics

NativeX is unveiling new 'psychographic' mobile ad targeting using hundreds of attributes about users to target ads to the exact demographic an advertiser wants, like "a 20-year-old undergrad who likes to gamble". In tests, this led to a 500-600% increase in eCPMs. Google announced updates to help webmasters optimize mobile pages for faster loading. On average, mobile pages take over 7 seconds to load which is still too slow. A wedding gown company launched a mobile app letting brides share photos with friends of dresses they tried on using "magic mirrors" that sync to phones.

Hightech_marketing_1_Introduction.pdfyounghoon lee

The document discusses the history and concepts of algorithmic marketing. It begins by explaining how the rise of digital marketing created an environment requiring millions of micro-decisions that cannot be made efficiently without algorithms. It then discusses how algorithmic marketing automates marketing processes in areas like product, promotion, price, and placement. The document also provides historical examples that helped establish algorithmic marketing, such as how online advertising exchanges developed and how airlines pioneered revenue management systems. Finally, it proposes a framework for programmatic marketing services involving data collection, feature transformation, objective-driven modeling and decision-making, and feedback loops.

STATE OF THE ART CONTENT MINING USING SCAN TECHNOLOGYIRJET Journal

The document discusses the SCAN (specific context-aware network) technology for enhancing context understanding and extracting meaningful information from diverse content sources. SCAN leverages natural language processing, machine learning, and semantic analysis to improve the accuracy and relevance of content mining results. The document provides an overview of SCAN and how it addresses limitations of existing content mining techniques by incorporating context-awareness. It also discusses potential applications of SCAN such as sentiment analysis, recommendations, and knowledge discovery.

Behavioral Targeting and Dynamic Content creationSimon Hjorth

Behavioral targeting uses information about individuals' online behavior to select personalized advertising messages. It allows marketers to improve targeting and maximize returns. AdPeople helps clients set up behavioral targeting programs through services like data management, dynamic creative content, and media partnerships. Case studies show behavioral targeting can significantly increase conversion rates and returns on ad spending.

How to Use AI Ad Generator for Google Ads, LinkedIn and Instagram Ads.pdfNarratoSocial

Mar-Tech OversightDaniel McKean

This document provides an overview of various marketing technology platforms and how they could be integrated as part of an analytics framework. It reviews platforms for digital advertising, audience profiling, conversion tracking, social media management, mobile optimization, and business intelligence dashboards. The frameworks aims to optimize campaigns, audiences, and budgets by tracking user journeys across channels and leveraging machine learning. Dashboards would provide real-time data insights to stakeholders for continually improving performance.

demo_teralyticsKuhan Wang

This document summarizes the development of a pipeline to analyze textual features on content URLs and model their relationship to user engagement. It scrapes URLs, processes text, extracts keywords, and uses those keywords as features in a logistic regression classification model to predict whether ads will be clicked based on the content URL. The pipeline was delivered to the company for implementation to optimize ad placement.

MBA Content Product 19.12.14.compressedThomas Roache

This document provides an overview of MediaCom Beyond Advertising's content strategy and services. It discusses conducting an understanding of audiences and brands to identify opportunities. It then outlines developing content solutions such as identifying territories, platforms, and an 'Involve stretch' to extend campaigns. Content execution services include pre-testing, distribution strategies, and measurement. An ideal client working relationship is also described where MediaCom provides integrated campaigns and validates content strategy opportunities.

Business Analytics Case on GYF AdsFitrie Ratnasari

Boosting Personalization In SaaS Using Machine Learning.pdfReza Rahimi

The document discusses personalization in software as a service (SaaS) using machine learning. It defines personalization and outlines a general personalization system architecture. Examples of personalization use cases at Dropbox are also provided, such as personalized search, lifetime value estimates, GUI rendering, Dropbox plan recommendations, churn scores, messaging, content ranking, promotions, and marketing. The key takeaways are that keeping customers happy is critical for SaaS success, personalization can help automate connecting and serving customers for better experiences, and machine learning enables personalization at scale.

Getting it Right: Building Quality into your Content (July 2014)Joe Gollner

The document discusses key steps for achieving and sustaining content quality, including defining content, understanding the content lifecycle, establishing a content strategy, using appropriate tools, leveraging automation, engaging authors, and understanding that information quality outcomes depend on content quality. It emphasizes balancing the various components, getting feedback from information use, and customizing tools and processes to support authors.

Best digital marketing institute in indiapomypaul20

The Robot MarketeerBart De Waele

The document discusses how artificial intelligence (AI) and machine learning will impact marketers. It argues that AI will not replace marketers but will change their roles. Marketers will need to develop new skills like systems thinking that are strengths of AI. They will also need breadth of knowledge across marketing disciplines as well as depth of expertise in specific areas. The organization of marketing teams will change and include roles focused on data, content, and digital experiences. Marketers should assess technologies' maturity levels and install a digital culture to help their organizations adapt to these changes.

Boosting Personalization In SaaS Using Machine LearningGlorium Tech

The document discusses personalization in software as a service (SaaS) using machine learning. It defines personalization and provides examples of use cases like personalized search, content ranking, marketing, and SaaS plans. The document also outlines a general two-phase personalization system architecture consisting of a feature store, personalization engine, and business logic/ranking components. It provides Dropbox as an example that uses various personalization techniques like personalized search, lifetime value, and churn scores to improve customer experience.

Module3 offsite webanalyticsGayathri Choda

This document provides an overview of off-site web analytics. It discusses various display advertising tools and metrics used to analyze online advertising data. It also covers search engine optimization, search engine marketing, and social media reporting. The document includes sections on display media, SEO, SEM, and social media metrics. It describes tools, metrics, and how to analyze data from various online marketing channels.

My Charity Connects: SEM/ SEO 101Tamera Kremer

This document provides an overview of search engine optimization (SEO) and search engine marketing (SEM), also known as pay-per-click (PPC) advertising. It discusses the importance of search engine prominence, and covers topics like keyword research, competitive analysis, ad copywriting, landing pages, analytics, budgeting, campaign implementation and optimization for both SEO and PPC strategies. The document also addresses basics of SEO like content development, internal linking, metadata, and social media optimization.

Tamera Kremer - When SEO Met SEM: All You Need To Know About Search Engine Op...CanadaHelps / MyCharityConnects

If a charity’s website launches on the internet and no one visits – does it make an impact? Don’t spend all your time and resources building a website that no one sees! Search Engine Optimization (SEO) and Search Engine Marketing (SEM) are just fancy ways of saying “drive more traffic to your website!” Learn easy, low-cost (and even free) techniques for jumping to the top of search engine results and getting more people to see your site. Make your website the place to be!

When SEO Met SEM: All You Need to Know About Search Engine Optimization and ...Social Innovation Generation

Presenter: Tamera Kremer If a charity’s website launches on the internet and no one visits – does it make an impact? Don’t spend all your time and resources building a website that no one sees! Search Engine Optimization (SEO) and Search Engine Marketing (SEM) are just fancy ways of saying “drive more traffic to your website!” Learn easy, low-cost (and even free) techniques for jumping to the top of search engine results and getting more people to see your site. Make your website the place to be!

IRJET- Hybrid Recommendation System for MoviesIRJET Journal

This document describes a hybrid recommendation system for movies that combines collaborative and content-based filtering. It uses the MovieLens rating dataset and supplements it with additional data from IMDB, such as movie details. Algorithms like nearest neighbors collaborative filtering and content-based filtering are used to provide personalized movie recommendations to users. The system architecture and design are outlined, including user profiles, movie searching, and success prediction for upcoming movies. An evaluation of the system demonstrates how additional content features can improve recommendation accuracy over collaborative filtering alone.

AUTOMATION & CROSS-CHANNELTinuiti

Search is getting so complicated: multiple search engines, bidding algorithms and AI, forecasting, and budget management. In this world of machine automation how do I maximize my productivity without losing control? In this session we’ll have a look at the tools available to search marketers in Google Ads as well as some time-saving technologies that have been implemented by 3rd party platforms to help you save time, extend your reach, and be more competitive in your SEM program.

Accenture presentation sydneyMichael Buckley

1. The document discusses how digital advertising is evolving with real-time bidding and the ability to target ads at a granular level based on audience characteristics. 2. Click-through rates for banner ads have been steadily declining over the past two decades according to data from various sources, following what is called the "law of shitty clickthroughs". 3. New approaches are needed to improve relevance and extend the useful life of advertising, such as using transient messages tailored to intent to maintain relevance over time as audiences move across devices and platforms.

Infopresse Plateformes 2017 - Maximising the Value of Your Digital Content to...iProspect Canada

IWT PARIS project 16 Nov.DarrenJin

This document summarizes the design and evaluation of two demonstrators for personalized online advertisements. The first demonstrator focuses on offline adaptation of ads based on movie context, user profile, social media posts, and preferences. An evaluation with 200 participants found that transparency and user control improved perceived quality, intention, understanding, and attitude toward ads. The second demonstrator aims to perform online adaptation by recognizing objects in videos and retrieving personalized ads to be integrated in real-time. It outlines the required technical capabilities and plans to leverage external APIs and partner technologies from the PARIS project.

Complete Guide to Digital Marketing:AdvantagesRama Devi K

Autonomous medical coding with discriminative transformersPatrick Nicolas

Open Source Lambda Architecture for deep learningPatrick Nicolas

More Related Content

Similar to Taxonomy-based Contextual Ads Targeting (20)

How to Use AI Ad Generator for Google Ads, LinkedIn and Instagram Ads.pdfNarratoSocial

Mar-Tech OversightDaniel McKean

demo_teralyticsKuhan Wang

MBA Content Product 19.12.14.compressedThomas Roache

Business Analytics Case on GYF AdsFitrie Ratnasari

Boosting Personalization In SaaS Using Machine Learning.pdfReza Rahimi

Getting it Right: Building Quality into your Content (July 2014)Joe Gollner

Best digital marketing institute in indiapomypaul20

The Robot MarketeerBart De Waele

Boosting Personalization In SaaS Using Machine LearningGlorium Tech

Module3 offsite webanalyticsGayathri Choda

My Charity Connects: SEM/ SEO 101Tamera Kremer

Tamera Kremer - When SEO Met SEM: All You Need To Know About Search Engine Op...CanadaHelps / MyCharityConnects

When SEO Met SEM: All You Need to Know About Search Engine Optimization and ...Social Innovation Generation

IRJET- Hybrid Recommendation System for MoviesIRJET Journal

AUTOMATION & CROSS-CHANNELTinuiti

Accenture presentation sydneyMichael Buckley

Infopresse Plateformes 2017 - Maximising the Value of Your Digital Content to...iProspect Canada

IWT PARIS project 16 Nov.DarrenJin

Complete Guide to Digital Marketing:AdvantagesRama Devi K

How to Use AI Ad Generator for Google Ads, LinkedIn and Instagram Ads.pdfNarratoSocial

Mar-Tech OversightDaniel McKean

demo_teralyticsKuhan Wang

MBA Content Product 19.12.14.compressedThomas Roache

Business Analytics Case on GYF AdsFitrie Ratnasari

Boosting Personalization In SaaS Using Machine Learning.pdfReza Rahimi

Getting it Right: Building Quality into your Content (July 2014)Joe Gollner

Best digital marketing institute in indiapomypaul20

The Robot MarketeerBart De Waele

Boosting Personalization In SaaS Using Machine LearningGlorium Tech

Module3 offsite webanalyticsGayathri Choda

My Charity Connects: SEM/ SEO 101Tamera Kremer

Tamera Kremer - When SEO Met SEM: All You Need To Know About Search Engine Op...CanadaHelps / MyCharityConnects

When SEO Met SEM: All You Need to Know About Search Engine Optimization and ...Social Innovation Generation

IRJET- Hybrid Recommendation System for MoviesIRJET Journal

AUTOMATION & CROSS-CHANNELTinuiti

Accenture presentation sydneyMichael Buckley

Infopresse Plateformes 2017 - Maximising the Value of Your Digital Content to...iProspect Canada

IWT PARIS project 16 Nov.DarrenJin

Complete Guide to Digital Marketing:AdvantagesRama Devi K

More from Patrick Nicolas (12)

Autonomous medical coding with discriminative transformersPatrick Nicolas

Open Source Lambda Architecture for deep learningPatrick Nicolas

AI for electronic health recordsPatrick Nicolas

This presentation discusses two AI strategies for mining medical records: ontologies and machine learning. Ontologies rely on domain experts and rules to extract information and provide high accuracy but do not scale well. Machine learning can process large amounts of data but provides less interpretable results and requires significant data. A balanced approach is needed depending on the use case and requirements for accuracy, scalability, and data availability.

Monadic genetic kernels in ScalaPatrick Nicolas

Scala for Machine LearningPatrick Nicolas

This presentation describes some key features of Scala uses in the creation of machine learning algorithms: 1 Functorial definition of tensors for learning non-linear models (manifolds) 2. Monads to compose of explicit kernel functions in Euclidean space 3. Implicit class to extends Scala standard library 4. Stackable traits and dependency injection to build formal models and dynamic workflows 5. Tail recursion to implementation dynamic programming techniques 6. Streaming to reduce memory consumption for big data 7. Control of back pressure in data flows http://patricknicolas.blogspot.com http://bit.ly/12GjRu9

Stock Market Prediction using Hidden Markov Models and Investor sentimentPatrick Nicolas

This presentation describes hidden Markov Models to predict financial markets indices using the weekly sentiment survey from the American Association of Individual Investors. The first section describes the hidden Markov model (HMM), followed by selection of features (investors' sentiment) and labeled data (S&P 500 index). The second section dives into HMMs for continuous observations and detection of regime shifts/structural breaks using an auto-regressive Markov chain The last section is devoted to alternative models to HMM.

Advanced Functional Programming in ScalaPatrick Nicolas

Adaptive Intrusion Detection Using Learning ClassifiersPatrick Nicolas

The document discusses using learning classifiers and genetic algorithms to implement an adaptive intrusion detection system. Traditional data mining techniques are limited in their ability to adapt to changing environments, but learning classifiers systems combine genetic algorithms and reinforcement learning to discover and evolve security policies and rules from real-time data. The rules are represented as genes and evolved over time through processes of crossover, mutation, and selection to accurately identify threats.

Data Modeling using Symbolic RegressionPatrick Nicolas

This is an introduction to the concept of symbolic regression for managing effectively data stream. Symbolic regression combines genetic algorithm, reinforcement learning and flexible policies to extract meaning or knowledge from data, in an ever changing environment. As the knowledge extracted from real-time data is human readable and consumable, decision makers can validate the findings of the algorithm and act appropriately. Symbolic regression is used in signal processing, process monitoring and adaptive caching in data centers.

Semantic Analysis using Wikipedia TaxonomyPatrick Nicolas

Hadoop EcosystemPatrick Nicolas

Multi-tenancy in Private CloudsPatrick Nicolas

Autonomous medical coding with discriminative transformersPatrick Nicolas

Open Source Lambda Architecture for deep learningPatrick Nicolas

AI for electronic health recordsPatrick Nicolas

Monadic genetic kernels in ScalaPatrick Nicolas

Scala for Machine LearningPatrick Nicolas

Stock Market Prediction using Hidden Markov Models and Investor sentimentPatrick Nicolas

Advanced Functional Programming in ScalaPatrick Nicolas

Adaptive Intrusion Detection Using Learning ClassifiersPatrick Nicolas

Data Modeling using Symbolic RegressionPatrick Nicolas

Semantic Analysis using Wikipedia TaxonomyPatrick Nicolas

Hadoop EcosystemPatrick Nicolas

Multi-tenancy in Private CloudsPatrick Nicolas

Recently uploaded (11)

Matplot Lib Practicals artificial intelligence.pptxPianoPianist

POSSIBLE WAYS TO CATCH A CHEATING PARTNER WITHOUT TRACEadriasalome95

The Digital Wave: Social Media Marketing in Kolkata’s Businessestechibo

Step-by-Step Strategy for Google Ads with a $500 Monthly Budget.pptxSubhash Jain

How Social Media Can Transform Your Businessdigitalhappiness4

berlin_wall_presentation.pptx12345678978nischayagarwal354

In social media, half-life is the time it takes for a post to receive half of...Scott M. Graffius

In social media, half-life is the time it takes for a post to receive half of its total engagement—such as likes, shares, and comments “Lifespan (Half-Life) of Social Media Posts: Update for 2025” by Scott M. Graffius is available at ResearchGate 👉 https://www.researchgate.net/publication/387700277_Lifespan_Half-Life_of_Social_Media_Posts_Update_for_2025 #SocialMedia #SozialeMedien #MediosSociales #MédiasSociaux #MediaSociali #RedesSociais

The Smart Way to Grow Your Quora ReputationSocioCosmos

Inspired Trendy Women Clothing and FashionTony Alonso

Smart TikTok Comments Strategies Workingsociomarketingcom

Tired of posting TikToks and getting zero comments? Yeah, same. That empty comment section just makes your content feel... flat. This short guide breaks down how to buy TikTok comments that look real and actually fit your vibe. Whether you're looking to buy custom TikTok comments with your own text or just want a quick boost to make your post feel alive, this shows exactly how it works. No logins, no weird steps — just pick a package, drop your link, and the comments start rolling in. I’ve been using SocioTraffic for a while to get TikTok comments that don’t scream “fake.” You can buy comments on TikTok that match your tone, promote your brand, or spark engagement. The best part? They actually help your post get seen more. If you're stuck or just tired of that awkward silence after you post, check this out. It’s simple, and it works. 👉 https://sociotraffic.com/buy-custom-tiktok-comments/

Best google ads course in jalandharlakshchaudhry404

Matplot Lib Practicals artificial intelligence.pptxPianoPianist

POSSIBLE WAYS TO CATCH A CHEATING PARTNER WITHOUT TRACEadriasalome95

The Digital Wave: Social Media Marketing in Kolkata’s Businessestechibo

Step-by-Step Strategy for Google Ads with a $500 Monthly Budget.pptxSubhash Jain

How Social Media Can Transform Your Businessdigitalhappiness4

berlin_wall_presentation.pptx12345678978nischayagarwal354

In social media, half-life is the time it takes for a post to receive half of...Scott M. Graffius

The Smart Way to Grow Your Quora ReputationSocioCosmos

Inspired Trendy Women Clothing and FashionTony Alonso

Smart TikTok Comments Strategies Workingsociomarketingcom

Best google ads course in jalandharlakshchaudhry404

Taxonomy-based Contextual Ads Targeting

1. Taxonomy-based Contextual Ads Targeting Patrick Nicolas Dec 8, 2009 patricknicolas.blogspot.com www.slideshare.net/pnicolas https://github.com/prnicolas

2. Purpose This presentation is a short introduction to the different components of an ads server that leverages semantic analysis to segment and target audience. Patrick Nicolas Copyright 2009-2011 - All rights reserved.

3. Semantics A few definitions… Contextual targeting is the process of inserting the most appropriate advertising into a published content (web pages, social network, tweets, blogs..) Taxonomy is the study or science of classification of concept or concrete items, in a logical and repeatable manner. Patrick Nicolas Copyright 2009-2011 - All rights reserved.

4. Conundrum Should targeting relies on ● audience preferences & behavior history ● content topic and style ● both? Publisher Content User Market Advertiser Promotion Patrick Nicolas Copyright 2009-2011 - All rights reserved.

5. Taxonomy-based Targeting It is assumed that the content consumed by a visitor is reflective of his/her interests, tastes & demographic characteristics. Therefore, targeting (yield) consists of analyzing content & extracting context. Patrick Nicolas Copyright 2009-2011 - All rights reserved.

6. Architecture An ads targeting components: engine two key ● Optimizer to balance objectives (budget, volume) & constraints (placements, frequency, exclusivity,..) ● Dispatcher to select the ad with the highest predicted yield according to content Patrick Nicolas Copyright 2009-2011 - All rights reserved.

7. Use Case 1. Campaign manager defines the objectives & constraints 2. Optimizer computes the best promotion in the inventory that satisfy constraints 3. Dispatcher formats & dispatches the promotion Ads. inventory Campaign 1 1 Manager Optimizer 1 2 1 Dispatcher 3 Content Patrick Nicolas Copyright 2009-2011 - All rights reserved.

8. Semantic Analysis 1. The optimizer performs a semantic analysis of both the content & promotional material. 1. The analysis generates taxonomy or semantic classification graphs. 1. Finally, the promotion with the taxonomy graph which is the closest to the content graph is selected Patrick Nicolas Copyright 2009-2011 - All rights reserved.

9. Taxonomy Match Portability Music Device Travel Content taxonomy Content Matching Listening Device Device iPod Promotion Autonomy Promotion taxonomy Patrick Nicolas Copyright 2009-2011 - All rights reserved.

10. Test Results The taxonomy match algorithm was evaluated against a rule-based targeting engine for consumer discretionary products. The Click Through Rate (CTR) increased from 0.8% to 0.193% (handbags) and 0.052% to 0.98% (upscale pen). Patrick Nicolas Copyright 2009-2011 - All rights reserved.

11. References ● How much can Behavioral Targeting Help Online Advertising? J. Yan, N. Lu, G. Wang, W. Zhang, Y Jiang http://www2009.eprints.org/27/1/p261.pdf ● Introduction to Semantic Analysis http://www.cs.tut.fi/sgn/arg/klap/introduction-semantics.pdf Patrick Nicolas Copyright 2009-2011 - All rights reserved.

Taxonomy-based Contextual Ads Targeting

Recommended

More Related Content

Similar to Taxonomy-based Contextual Ads Targeting (20)

More from Patrick Nicolas (12)

Recently uploaded (11)

Taxonomy-based Contextual Ads Targeting