Marketing^AI

This academic paper proposes a novel method called the Experimental Selection Correction (ESC) estimator for combining observational and experimental data to estimate the causal effects of a treatment on a primary long-term outcome . Researchers often have access to large observational datasets with many outcomes but non-random treatment assignment , and smaller experimental datasets with random a...

This article introduces the use of double machine learning (DML) in survey statistics to understand nonresponse in a high-dimensional setting. It explains how traditional machine learning methods are good for prediction but can produce biased estimates of causal relationships, while DML can provide approximately unbiased estimators by effectively handling numerous potential confounding variables....

This source presents a trajectory for generative AI (GenAI) in consumer research, outlining three stages: democratization , where GenAI expands accessibility for consumers and researchers; the average trap , where GenAI's predictive nature leads to generic outcomes; and model collapse , where GenAI increasingly learns from its own output, losing connection to real-world human behavior. The aut...

We discuss how augmenting structured data with features from unstructured sources, like text or images, impacts statistical analysis. They propose that while this augmentation can introduce bias through the use of machine learning models for feature extraction, it also leads to a significant reduction in variance due to the richer information available. The text explores whether this variance redu...

This academic paper presents Missing At Random Structured data (MARS) , a framework for valid statistical inference using predictions from black box AI models , particularly neural networks. It addresses the issue of bias propagation when these models are used to impute missing structured data from unstructured sources. MARS achieves robust and efficient estimation by requiring access to a ground...

We explain Particle Filtering , also known as Sequential Monte Carlo , as a method for tracking hidden states in systems with noisy measurements where traditional filters struggle due to non-linear or non-Gaussian characteristics. The core idea involves using a "swarm" of particles representing possible states, which are weighted based on how well they match observations and then resampl...

 the ongoing  struggle for data ownership and transparency  within the digital marketing ecosystem. It explains that  "walled gardens" , controlled by major platforms, limit brands' access to crucial performance data, making analysis and attribution difficult. Furthermore, the text highlights how some  marketing agencies  can inadvertently or deliberately create a secondary barrier b...

We analyze how the rise of Generative AI (GenAI) has fundamentally changed marketing, arguing that rather than making the creative brief obsolete, it makes it even more crucial. The document emphasizes that AI requires a detailed and strategic brief as quality input to produce relevant and on-brand content, preventing the "garbage in, garbage out" problem. It highlights that the brief no...

This survey paper focuses on inference-time techniques for controlled generation with diffusion models , contrasting them with post-training methods. It introduces the concept of optimizing downstream reward functions during the sampling process , often referred to as alignment, which can involve conditioning on target properties or maximizing regressor outputs. The document outlines various deriv...

We examine a significant  challenges faced by digital marketers  today, such as the  opacity of large ad platforms  that rely on "black box" AI, a lack of control over campaign details, and difficulties with  personalization and creative scaling . They highlight how generic AI tools can  dilute brand identity  and how managing diverse data across numerous channels creates fragmentation....

These sources collectively explore the evolving landscape of digital advertising , specifically focusing on the integration of advertising within AI chatbot and search interfaces . The first article details Google's expansion of its AdSense network into AI conversations , driven by the need to maintain advertising dominance amidst the rise of generative AI competitors and ongoing regulatory ch...

This academic paper focuses on differentially private submodular maximization , a technique crucial for data summarization in scenarios involving sensitive information. It explores algorithms that can maximize submodular functions —which capture diminishing returns, useful for tasks like feature selection and data summarization—while also adhering to differential privacy , ensuring individual data...

We provide a  comprehensive analysis  of Perplexity AI's advertising system, detailing its  novel, contextual approach  that diverges from traditional search advertising by focusing on user queries and avoiding deep personalization or tracking. It explains the  ad formats , primarily sponsored follow-up questions and sidebar placements, and highlights the  high CPM pricing  and the  challenges...

This paper presents a novel  automated mechanism design framework  that utilizes  large language models (LLMs)  to reformulate the process as a  code generation task . The framework generates heuristic mechanisms in code and evolves them to  optimize for performance metrics  while ensuring crucial design criteria through a  problem-specific fixing process . This approach addresses limitations in t...

This academic paper examines the  value of offsite tracking data for online advertisers , specifically on platforms like Meta (Facebook and Instagram). Through a large-scale experiment involving over 70,000 advertisers, the study  quantifies the impact of losing the ability to optimize ad delivery using data about user actions outside the platform , such as website purchases. The findings indicate...

We analyze the strategic placement of advertisements within conversational AI platforms, specifically focusing on how the dynamic nature of user interaction influences optimal ad auction design. The core tension explored is the trade-off between  acquiring more information about ad quality over time  and the risk of  decreasing market thickness , which negatively impacts revenue. The research find...

This conceptual analysis explores  Meta's proposed AI-driven advertising strategy , where algorithms handle everything from creative generation to targeting and optimization, drawing parallels to  established quantitative marketing concepts  like ad response modeling, consumer heterogeneity, and dynamic optimization. The text discusses how this automated approach  aims to maximize advertising...

This analysis examines Meta Platforms' vision for its advertising business, which is undergoing a fundamental shift driven by artificial intelligence. Mark Zuckerberg, Meta's CEO, envisions an automated ecosystem where AI handles everything from creative development to targeting and measurement, aiming for greater efficiency and simplified processes, especially for small businesses. However, the t...

We discuss the  MM* condition  in game theory, where players lack a single secure pure strategy to guarantee their best possible worst-case outcome. The text explains how  repeated games satisfying this condition  can lead to  pathological learning behaviors  like non-convergence and chaotic adaptation between players, because agents are forced into complex or randomized strategies. However, the s...

This paper focuses on  learning efficient equilibria in repeated games.  The author proposes a  stochastic learning rule  that selects a subgame-perfect equilibrium where players receive efficient payoffs. The paper details how this rule works, involving  bounded-memory models, approximate best responses, and periodic testing  of these models against observed actions. A key element is the probabil...

This academic article from  Econometrica , published by The Econometric Society and available on JSTOR, explores the concept of  subjective equilibrium  in the context of repeated games. It contrasts this idea with the more established notion of  Nash equilibrium , which assumes players have accurate knowledge of their opponents' strategies. The authors, Ehud Kalai and Ehud Lehrer, propose tha...

This academic paper, published in the journal  Games and Economic Behavior , discusses the  possibility of Bayesian learning in infinitely repeated games . The author examines prior research suggesting limitations on rational learning in such games, particularly the difficulty of achieving both belief consistency and strategy learnability. The paper proposes a modified concept called  "optimi...

This paper published in the September 1993 issue of  Econometrica: Journal of the Econometric Society  and available through  JSTOR , presents a theoretical study on how players in infinitely repeated games with rational learning will eventually converge to playing a  Nash equilibrium . The authors, Ehud Kalai and Ehud Lehrer, explore how players with  subjective beliefs  about their opponents&#39...

These sources discuss QALIGN, a new technique for aligning Large Language Model (LLM) outputs with desired preferences at the time of generating the response, rather than requiring expensive prior training or fine-tuning of the model weights. The core idea is to use Markov chain Monte Carlo (MCMC) sampling, guided by an external reward model (RM), to iteratively refine outputs and converge towards...

This paper discusses a high-dimensional choice model  for online retailers to address the challenge of understanding  substitution patterns  among a large number of products, which is difficult with traditional models. By leveraging  consumer clickstream data  and combining  econometric and machine learning methods , specifically the graphical lasso technique, the authors aim to  learn flexible su...