Ritwik Sinha

Estimating the incremental effects of interactions for marketing attribution

Assigning credit to different marketing activities has long been an important but challenging goal for a marketer. With the advent of digital marketing, the marketer can now potentially record each interaction with a prospective customer. With this development it is possible to measure and assign credit for each marketing interaction. We propose an econometric model to estimate the true incremental number of purchases that can be attributed to a given marketing channel. We extend our model to attribute credit for revenue realization. We also propose an approach to automatically identify audience segments where attribution models differ. We build and test our model on a real world data set belonging to a travel and experience industry organizations web data. The results show that our approach improves upon industry standard rule-based approaches, by correcting for the biases inherent to these model.

A Non-parametric Approach to the Multi-channel Attribution Problem

Multi-channel marketing attribution modeling is a two-stage process. First, the value of exposure from different marketing channel needs to be estimated. Next, the total surplus achieved needs to be assigned to individual marketing channels by using the exposure effects from the first stage. There has been limited work in exploring possible choices and effects of determining the value of exposure to different marketing channels in the first stage. This paper proposes novel non-parametric and semi-parametric approaches to estimate the value function and compares it with other natural choices. We build a simulation engine that captures important behavioral phenomenon known to affect a customers purchase decision; and compare the performance of five attribution approaches in their ability to closely approximate the known ground truth. Our proposed method works well when marketing channels have high levels of synergy. We apply the proposed approaches on two real-world datasets and present the results.

Joint modeling tumor burden and time to event data in oncology trials.

The tumor burden (TB) process is postulated to be the primary mechanism through which most anticancer treatments provide benefit. In phase II oncology trials, the biologic effects of a therapeutic agent are often analyzed using conventional endpoints for best response, such as objective response rate and progression-free survival, both of which causes loss of information. On the other hand, graphical methods including spider plot and waterfall plot lack any statistical inference when there is more than one treatment arm. Therefore, longitudinal analysis of TB data is well recognized as a better approach for treatment evaluation. However, longitudinal TB process suffers from informative missingness because of progression or death. We propose to analyze the treatment effect on tumor growth kinetics using a joint modeling framework accounting for the informative missing mechanism. Our approach is illustrated by multisetting simulation studies and an application to a nonsmall-cell lung cancer data set. The proposed analyses can be performed in early-phase clinical trials to better characterize treatment effect and thereby inform decision-making.

Will your facebook post be engaging

Social media has become the ideal platform for promotional activities of organizations. However, due to the volatility of social media, the wrong message posted at the wrong time can result in significant damage to hard-built brand image. This calls for a mechanism to gauge the reactions a post will evoke from a given social community. The community can vary from customers of a particular brand to brand loyalists interacting through its social pages (for example, on Facebook). In this paper, we focus on learning the communitys reaction from past posts and providing a predictive model for gauging the reaction of the community before the post is published. This helps the marketer take better-informed decisions. Short-text posts in social media leads to a sparse feature space, we propose additional meta-features that improve reaction modeling. Given the feature representation, we discuss the possibility of casting the underlying problem under different paradigms - classification, regression and learning to rank. We study the performances of each of these paradigms on real data from Facebook. We will discuss the challenges involved, and ways to mitigate them, in addition to our observations, results and insights.

Probabilistic Deduplication of Anonymous Web Traffic

Cookies and log in-based authentication often provide incomplete data for stitching website visitors across multiple sources, necessitating probabilistic deduplication. We address this challenge by formulating the problem as a binary classification task for pairs of anonymous visitors. We compute visitor proximity vectors by converting categorical variables like IP addresses, product search keywords and URLs with very high cardinalities to continuous numeric variables using the Jaccard coefficient for each attribute. Our method achieves about 90% AUC and F-scores in identifying whether two cookies map to the same visitor, while providing insights on the relative importance of available features in Web analytics towards the deduplication process.

Improving Marketing Interactions by Mining Sequences

The advent of digital marketing has revolutionized how a marketer reaches the organizations customers. Since each interaction with the customer is recorded today, the marketer can do a better job of measuring the effectiveness of marketing efforts. With multi-channel marketing data, comes a new set of challenges; those of measuring the effect of individual channels, understanding synergistic effects, and finally leveraging the information stored in the sequence of marketing activities. While there is some work addressing the first two challenges, we aim to shed light on the last question. The combinatorial explosion in the number of possible marketing sequences requires a systematic approach to address this problem. We propose an approach based on sequence mining to identify marketing touch sequences that are most likely to lead to a stated marketing goal. Our approach provides a rapid way of creating marketing campaigns with the highest chance of success. We test our proposed approach on a real world dataset of a retail chain with web visits, digital marketing channels, email data, instore and online purchase data. We compare against baseline approaches, and observe interesting insights in the real data.

Community reaction: from blogs to Facebook

Online social media is all pervasive in this digitally connected world. It provides a great platform to share information and news, and have public discussions on these topics. These interactions happen on owned-sites as well as on earned social media. But it is reasonable to hypothesize that the communities on the various platforms will engage differently. We explore this hypothesis using data collected from a blog as well as their Facebook page. We observe significant differences between the blog and Facebook as two mediums. First, we see how the topic of the post leads to differences in how posts are received on the two mediums. We next characterize the distribution of the time to comments, displaying how blog and Facebook differ. Finally, we describe how the sentiment of posts and popular entities differ across mediums. In conclusion, reactions across the two mediums are so diverse that it calls for different strategies across different social mediums, we provide our recommendations towards this goal.

Linking Clicks to Bricks: Spillover Benefits of Online Advertising

Businesses have widely used email ads to directly send promotional information to consumers. While email ads serve as a convenient channel that allows firms to target consumers online, are they effective in increasing offline revenues for firms that predominantly sell in brick-and-mortar stores? Is the effect of email ads, if any, heterogeneous across different consumer segments? If so, on which consumers is the effect highest? In this research, we address these questions using a unique high-dimensional observational dataset from one of the largest retailers in the US, which links each consumer’s online behaviors to the item-level purchase records in physical stores. We use a doubly robust estimator (DRE) that incorporates nonparametric machine learning methods and allows us to perform causal estimation on observational data. Using the DRE we find that receiving email ads can increase a consumer’s spending in physical stores by approximately

Audience Prism: Segmentation and Early Classification of Visitors Based on Reading Interests

11.82. Additionally, we find that the increased offline sales result from increased purchase probability and a wider variety of products being purchased. Further, we use a data-driven approach to demonstrate that the effect of email ads is heterogeneous across different consumer segments. Interestingly, the effect is highest among consumers who have fewer interactions with the focal retailer recently (i.e., lower email opening frequency). Overall, our results suggest a reminder effect of email ads. Receiving email ads from the retailer can generate awareness and remind the consumer of the retailer’s offerings of various products and services, which gradually increase the consumer’s purchase probability in the retailer’s physical stores. These findings have direct implications for marketers to improve their digital marketing strategy design and for policy makers who are interested in evaluating the economic impact of prevalent email advertising.

Predicting Reactions to Short-Text Posts

The largest Media and Entertainment (M&E) web portals today cater to more than 100 Million unique visitors every month. In Customer Relationship Management, customer segmentation plays an important role, with the goal of targeting different products for different segments. Marketers segment their customers based on customer attributes. In the non-subscription based media business, the customer is analogous to the visitor, the product to the content, and a purchase to consumption. Knowing which segment an audience member belongs to, enables better engagement. In this work, we address the problems: 1) How can we segment audience members of an M&E web property based on their media consumption interests? 2) When a new visitor arrives, how can we classify them into one of the above defined segments (without having to wait for consumption history)? We apply our proposed solution to a real world data-set and show that we can achieve coherent clusters and can predict cluster membership with a high level of accuracy. We also build a tool that the editors can find valuable towards understanding their audience.