Published on February 27, 2014
Privacy-Friendly Personalization Charlie Reverte VP Engineering @numbakrrunch
Personalized Don't judge me...
Why Personalize? ● Funnel optimization ○ Shorten the path between a person and something they want ● Increased conversion through higher relevance ● Super-serve individual audience segments ● Streamline your site to focus users on your content, products and ads
The Next Phase of the Web ● In the beginning there were directories.. ○ Navigate a hierarchy to find content ● Then came search ○ Skip the hierarchy, just tell us what you want ● Now the web is personalizing ○ We already know what you want, here’s a great experience
Good Data for Personalization ● Infer intent from behavior ○ Demographics are a poor proxy ● Observed vs. Declarative ○ What I do vs. what I say or "like" ● Context ○ Show me something related ○ Match my mood
User-Defined Segmentation Declarative Inputs Data Output Behavioral Inputs
Personalization Pitfall: Boxing ● Boxing - is where a consumer’s vision and choices are limited by .. analytics that make judgments based on their digital history ○ Martin Abrams - Boxing and Concepts of Harm ● Already a self-selecting behavior ● Don't reinforce, however users don't want to be contradicted either
Think Discovery ● Don't spend all of your inventory on reinforcement ● Diversify user interests to broaden conversion
Cold Start: Use 3rd-Party Data ● Cold-start problem ○ Bootstrap with 3rd-party data ● Maximize yield on your earned and owned user acquisition ○ The widest part of your funnel ● Treat first time visitors as well as your loyal customers
3rd-Party Data Landscape ● Cookie-based behavioral data ○ Will be fragmented but not dead if 3rd-party cookies are limited ● Contextual extraction ○ Scrape, classify, referer, geo, time of day ● Hybrid: Contextualized based on observed behavior ○ Classify sites based on behavior + context ○ Apply to cold start
Context and Surprise What's the difference? 1st- vs. 3rd-party context? Creepy? Useful?
Minimizing the Creep Factor Privacy isn't just about PII vs. non-PII and 3rd-party cookies ● Take a user-centric approach ● Base your application of data on user sensitivities and expectations of privacy ● Avoid surprising users ● Clarify whether they're anonymous or not ● Leverage real-world paradigms ○ Users know how their "street identity" works
Transparency and Control Empower your users..
Summary ● Privacy-friendly personalization is a valueadd to users ● Take a user-centric approach to privacy ● Build trust through transparency and control ● Apply 3rd-party data to unlock your site's full potential
Questions? Charlie Reverte VP Engineering @numbakrrunch email@example.com
Presentación que realice en el Evento Nacional de Gobierno Abierto, realizado los ...
In this presentation we will describe our experience developing with a highly dyna...
Presentation to the LITA Forum 7th November 2014 Albuquerque, NM
Un recorrido por los cambios que nos generará el wearabletech en el futuro
Um paralelo entre as novidades & mercado em Wearable Computing e Tecnologias Assis...
Special Section: Personalization vs. Privacy. ... Personalization vs. Privacy. ... "The best personalization is done in a privacy-friendly manner where all ...
Personalization Privacy - Download as PDF File (.pdf), Text File (.txt) or read online. Scribd is the world's largest social reading and publishing site.
Journal of Management and Marketing Research E-personalization and online, Page 1 E-personalization and online privacy features: the case with
Let’s Do It at My Place Instead? Attitudinal and Behavioral ... as yet whether users also perceive CSP as more privacy-friendly than remote personalization.
What are the guidelines when using data to personalize digital experiences, and how can we help people feel comfortable with personalization services?
Personalization vs. Privacy: Drawing the Line on Customer Information. Privacy advocates struggle to protect consumers and data collection ...
PRIVACY-ENHANCED PERSONALIZATION (PEP2005) Proceedings Edited by: ... when it comes to requirements specification for privacy in UC, user-friendly
Accepted for publication in the Journal of the Association for Information Science and Technology (JASIST) !!! The!Effect!of!Personalization!Provider ...
User Modeling and Personalization 12: Personalization and Privacy Eelco Herder L3S Research Center / Leibniz University of Hanover Hannover, Germany