广告与推荐系统.pdf

Advertising Recommendation Dou Shen Baidu @沈抖 (weibo) Outline  Overview  Evolution of Display Ads  Metrics  User Data  Targeting Recommendation  Future Overview Definition  Advertising • A form of communication used to encourage or persuade an audience to continue or take some new actions.  Computational Advertising • A principled way to find the best match between a user in a context and a suitable advertisement  Recommendation • A particular form of information filtering, that exploits past behaviors and user similarities to generate a list of information items that is personally tailored to an end- user’s preferences Ad examples How about these? Major players  Players Traffic • Inventory • Users / Searcher • Supply • Se ers / Publishers / Search Engines Engine • Demand • Buyers / Advertisers / Marketers User Ad  Campaign Types • Branding (awareness and message reinforcement) • Direct response (ROI/acquisition-focused) Pricing models  CPM = cost per thousand impressions  CPC = cost per click  CPT/CPA = cost per transaction/action Advertising revenue market share by media Historical annual revenue trends (U.S.) Advertising format share Internet ad revenues by industry category Internet ad revenues by pricing model Evolution Changes Ad format Interactive Ads Display Ads Text Ads Traffic Search Web page Game Video App Emai Socia PC Mobile The evolution of display advertising - IAB  Video from IAB Direc