Purchase Conversion Research Articles

Resource allocation and scheduling in the intelligent edge computing context

With the rapid development of the edge computing, Internet economy has become a new situation in economic development. What has brought development to the Internet economy is a variety of family-based platforms (i.e., edge computing). However, in the face of huge Internet users and different users’ shopping habits, reasonably personalized e-commerce platform allocations for users can improve the user’s website shopping experience. However, due to the huge user data and the diversity of SDN (software defined network) platforms, it is a huge challenge to reasonably allocate e-commerce resources/platforms to users. The quality of the SDN personalized resource allocations also affects the purchase conversion rate. Clustering algorithm is an algorithm involved in grouping data in machine learning. The same set of data has the same attributes and characteristics, and the attributes or features between different sets of data will be relatively large. In this paper, by using the mean shift clustering algorithm to characterize the behavior data. According to the characteristics of the grouping, we can allocate the e-commerce platform commonly used between the groups. However, using mean shift clustering for personalized allocation faces the problem of too high user data dimensions. Therefore, we first conduct computational efficiency analysis toward each user. We define user behavior sequences for user behavior data and classify user behavior. We transform the grouped user behavior into an embedded vector, and linearly transform the embedded vectors of different lengths into the same semantic space. We process the vectors in the semantic space through the self-attention layer and perform mean shift clustering. Experiments show that, in the edge computing context, our method can reduce the complexity of resource allocation toward complex data and improve the quality of the allocated data.

Intelligent Recommendation System Based on Mathematical Modeling in Personalized Data Mining

With the advent of the era of big data, data mining has become one of the key technologies in the field of research and business. In order to improve the efficiency of data mining, this paper studies data mining based on the intelligent recommendation system. Firstly, this paper makes mathematical modeling of the intelligent recommendation system based on association rules. After analyzing the requirements of the intelligent recommendation system, Java 2 Platform, Enterprise Edition, technology is used to divide the system architecture into the presentation layer, business logic layer, and data layer. Recommendation module is divided into three substages: data representation, model learning, and recommendation engine. Then, the fuzzy clustering algorithm is used to optimize the system. After the system is built, the performance of the system is evaluated, and the evaluation indexes include accuracy, coverage, and response time. Finally, the system is put into a trial operation of an e-commerce platform. The click-through rate and purchase conversion rate of recommended products before and after the operation are compared, and a questionnaire survey is randomly launched to the platform users to analyze the user satisfaction. The experimental data show that the MAE of this system is the lowest, maintained at about 0.73, and its accuracy is the highest; before the recommended threshold exceeds 0.5, the average coverage rate of this system is the highest: 0.75; in Q1–Q5 subsets, the shortest response time of the system is 0.2 s. Before and after the operation of the system, the average click-through rate increased by 11.04%, and the average purchase rate increased by 9.35%. Among the 1216 users, 43% of the users were satisfied with 4 and 9% with 1. This shows that the system algorithm convergence speed is fast; it can recommend products more in line with user needs and interests and promote higher click-through rate and purchase rate, but user satisfaction can be further improved.

Dynamic effects of store promotions on purchase conversion: Expanding technology applications with innovative analytics

Retail technology provides novel ways to investigate the effects of marketing actions (MAs) on consumers’ behaviors. To evaluate effects of MAs on conversion, we propose an approach that combines sensor-collected records of store entries with register-receipts within pre-determined time-bands. Studies of effects of MAs on conversion in brick-and-mortar settings, however, are surprisingly scarce. The approach has three stages: (a) build a model for the conversion probabilities with data outside the MA interval, (b) build a counterfactual baseline for the conversion probabilities during the MA, under the hypothesis that this action has not occurred, (c) compare the observed conversions to the corresponding counterfactual conversions to determine the effects of the MA. An analysis of a two-day promotion illustrates the method. It provides evidence that the promotion associates with increased counts of both visits and purchases, but the association with conversion remains unclear.

Prediction of Internet User’s Purchase Behavior Based on Mixed kernel SVM Model

In order to better predict the purchase behavior of online consumers and improve the purchase conversion rate of e-commerce companies, a model of consumer purchase intention prediction based on mixed kernel function support vector machine is proposed. In the research, the prediction model of consumer purchase intention is designed as a two-classification problem, and the feature selection algorithm mRMR is used to obtain the ranking of features, seeking to obtain better or similar classification results when choosing fewer features. On its basis, the particle swarm optimization algorithm (PSO) is used to optimize the parameters of the model. Experiments show that, to a certain extent, the constructed mixed kernel function can effectively improve the classification effect of SVM.

Pricing Joint Sales and Rentals: When are Purchase Conversion Discounts Optimal?

Product sales and rentals often occur in parallel, allowing customers to choose their preferred option. In settings where consumers face significant value uncertainty, rentals also provide a mechanism for consumers to discover whether they like a product. For some products, such as movies and books, the rental option may yield a significant proportion of the utility associated with ownership. The convenience of renting products in digital forms (e.g., e‐books) makes the managerial challenge of jointly pricing rentals and sales increasingly relevant. Rental and sales prices are interdependent not only because consumers can choose between the two options, but also because a customer may rent before purchasing. Given these considerations, and allowing for consumer uncertainty, we analyze the optimal joint pricing of product rentals and sales. We show that when consumers have a relatively high probability of liking a product, and will experience a significant drop in their post‐rental residual utility, then the practice of allowing a portion of the rental price to apply to a future purchase is optimal.

Investigating the Role of Multiple Channels in Predicting Website Browsing Patterns and Purchase

In today’s online environment, consumers and sellers interact through multiple channels such as email, search engines, banner ads, affiliate websites and comparison-shopping websites. In this paper, we investigate whether knowing what channels a consumer used on their path until a point of time is predictive of their future browsing patterns and purchase conversion on the website. We propose a model in which the browsing patterns and conversion are stochastically dependent on the sequence of channels the consumer has used up to this point. Important features of our model are that we allow for clustered visits by consumers, which enables separation of their browsing patterns into intra- and inter-session components, and that channels can be evaluated using multiple metrics such as visit frequency and conversion. We estimate the model on a dataset obtained from an online retailer selling a durable good and find that information on both the identities and the frequencies of the channels through which consumers visited has significant predictive power for future conversion behavior. We find that channels may act as “closers” or “engagers” — consumers arriving through the former are more likely to make a purchase, while consumers arriving through the latter, even if they do not make a purchase, are more likely to visit again in the future or extend the current session. We discuss several managerial implications of the model including model based visualizations, landing page design, multi-dimensional propensity scoring, and the use of the model to evaluate a policy change regarding the encryption of channel information.

"I Think You'll Like It"

Understanding the roles of social factors in online purchase behavior has been a long standing research problem. The recently emerging social e-commerce platforms leverage the stimulated word-of-mouth effect to promote the sales of items, which offers a peek into the complex interplay between the social influence and online purchasing behavior. In this paper, we investigate this problem on a full-scale purchase behavior dataset from one of the leading social e-commerce platforms, Beidian. Specifically, we conduct a comparison study between the social e-commerce and conventional e-commerce that are both integrated in Beidian to examine how social factors affect user's purchase behavior. We reveal that social e-commerce leads to a 3.09~10.37 times higher purchase conversion rate compared with the conventional settings, which indicates users make purchase with significantly fewer item explorations. Then, we propose and validate four primary mechanisms that contribute to the efficient purchase conversion: better matching, social enrichment, social proof and price sensitivity. Moreover, we identify several behavioral indicators that are able to measure the effect of these mechanisms, based on which we design an accurate predictive model (AUC=0.7738) for user's purchase decision. These results combine to shed light on how to understand and model the purchase behavior in social e-commerce.

Coupons Along the Conversion Funnel: A Structural Approach

We study how couponing affects consumer behavior online along the conversion funnel. Toward this end, we explore three explanations: Couponing can affect consumer response by changing (1) the cost a consumer incurs to visit the website (i.e., visit cost), (2) the cost she incurs to search for a product (i.e., search cost) upon visit, and (3) the utility from purchase through price discount (i.e., price discount). Leveraging the exogeneous variation generated from a randomized field experiment of couponing, we develop a structural model to measure the causal effects of the coupon on website visit, product search, and purchase conversion in an e-commerce setting. We find couponing lifts consumer purchases primarily through an increase in consumer visits to the website, and most of the coupon effects on consumer visits arise from lowering the cost for consumers to visit the website. Using policy simulations, we demonstrate the presence of the coupon effects through visit cost, search cost, and price discount, and quantify the magnitude of these effects along the conversion funnel. We discuss the implications for customer targeting and promotion evaluation.

Do cognitive and affective expressions matter in purchase conversion? A live chat perspective

Live chat, which is embedded in some shopping websites, allows a retailer to communicate with its shoppers directly and respond to their inquiries promptly. We extended the Speech Act Theory (SAT) to this context by developing the Dialog Conversion Model (DCM) that elucidates the characteristics of online dialog between shoppers and retailers. We validated the DCM using quantitative and qualitative data. The quantitative analyses of 1,185 communication sessions and related purchase records reflect that shoppers' cognitive, rather than affective, expressions tend to indicate their purchase intentions, and retailers should emphasize cognitive expressions to the extent possible while being strategic in their choice of expressions. Furthermore, the affective expressions (in positive and negative forms) articulated by retailers have a considerable influence on purchase conversion. However, the same is not true for shoppers. Post‐hoc interviews were conducted with shoppers to gain additional insights.

Customer Purchase Intent Prediction Under Online Multi-Channel Promotion: A Feature-Combined Deep Learning Framework

The micro-level customer purchase intent in promotions is crucial for the overall purchase conversion rate of promotions. In the context of joint promotions on multiple online channels, customers can access and compare prices and services by navigating between channels to aid their purchase decision. The interactions between customer and promotion channels offer another angle to predict their purchase intent during promotions. In this paper, we propose a feature-combined deep learning framework, in which a full-connected long short-term networks (FC-LSTM) is used for modeling the interactions between customers and promotion channels, as well as the nonlinear sequence correlations and cumulative effects between customer's browsing behavior. To improve the performance of the prediction, the framework incorporates other features of customer profile including purchase history and demographics, integrating them into an end-to-end framework. We apply our method in a real prediction task for online multichannel promotion for concert tickets. Extensive experiments show that the proposed approach exhibits overall good performance compared with state-of-the-art methods on standard metrics such as precision, recall, f-measure, area under curve (AUC), and lift.

Modeling the effects of dynamic group influence on shopper zone choice, purchase conversion, and spending

In many retail contexts, social interaction plays an important role in the shopping process. We propose a three-stage dynamic linear model that captures the influence of group discussion on shopper behavior within a hierarchical Bayes framework. The model is tested using a video tracking and transaction dataset from a specialty apparel store. The research reveals that group conversations have a significant impact on the shopper’s department or “zone” choice, purchase likelihood, and spending over time. This group influence is magnified by the size of the group (particularly for zone penetration and purchase conversion), and is also moderated by group composition and cohesiveness. The conversations of mixed-age groups and groups who stay together while shopping have a significant influence on shopper behavior across all three stages, while discussions by adult groups exhibit a marginal carryover effect for purchase conversion. When shoppers have repeated discussions in a specific department, they are more likely to return to and buy from this department, while the cumulative number of discussions in the store drives higher spending levels. We also observe that group shoppers visit more departments than their solo counterparts; and mixed-age groups and solo shoppers are more likely to buy than adults-only or teen groups. This study has important implications for how retailers manage shopper engagement and group interaction in their stores.

Real-time prediction of online shoppers’ purchasing intention using multilayer perceptron and LSTM recurrent neural networks

In this paper, we propose a real-time online shopper behavior analysis system consisting of two modules which simultaneously predicts the visitor’s shopping intent and Web site abandonment likelihood. In the first module, we predict the purchasing intention of the visitor using aggregated pageview data kept track during the visit along with some session and user information. The extracted features are fed to random forest (RF), support vector machines (SVMs), and multilayer perceptron (MLP) classifiers as input. We use oversampling and feature selection preprocessing steps to improve the performance and scalability of the classifiers. The results show that MLP that is calculated using resilient backpropagation algorithm with weight backtracking produces significantly higher accuracy and F1 Score than RF and SVM. Another finding is that although clickstream data obtained from the navigation path followed during the online visit convey important information about the purchasing intention of the visitor, combining them with session information-based features that possess unique information about the purchasing interest improves the success rate of the system. In the second module, using only sequential clickstream data, we train a long short-term memory-based recurrent neural network that generates a sigmoid output showing the probability estimate of visitor’s intention to leave the site without finalizing the transaction in a prediction horizon. The modules are used together to determine the visitors which have purchasing intention but are likely to leave the site in the prediction horizon and take actions accordingly to improve the Web site abandonment and purchase conversion rates. Our findings support the feasibility of accurate and scalable purchasing intention prediction for virtual shopping environment using clickstream and session information data.

The Impact of Coupons on the Search-to-Purchase Funnel: Theory and Empirical Evidence

Firms use coupons to stimulate demand. Although couponing is popular in practice, limited research has examined the causal effects of coupons on visit, search, and purchase behaviors among heterogeneous customers. In this paper, we explore coupon effects using data from a randomized field experiment with an online retailer in which customers were divided into two heterogeneous customer segments (low value and high value) with two types of coupon discounts (base value and better value). We find couponing is effective in increasing revenue, primarily by attracting customers who purchase without coupon redemption, and the lift in revenue per customer is larger for the high-value segment. Using clickstream data of customer visit and search behavior, we find most of the revenue lift arises from a corresponding lift in the likelihood of visiting the website under couponing. Though the lift in visit likelihood is relatively homogeneous across customer segments under the base coupon, the high-value segment has a higher purchase conversion rate than the low-value segment, leading to an amplified revenue lift. We also find a deeper discount leads to higher redemption and purchase conversion for the high-value segment but does not change visit likelihood. Finally, most of the search behaviors are unchanged under couponing, suggesting the mix of customers brought in under couponing are similar to those who visit without receiving the coupon promotion.

Attribution Modeling In Digital Advertising

ABSTRACT Digital advertising campaigns often are launched across multiple channels, a selection of which may include search, display advertisements, social media, mobile, video, and e-mail. By exposing consumers to advertisement impressions, these channels help consumers make purchase decisions or sign up to an advertised service (Fulgoni, 2016; Yang and Ghose, 2010). To gauge the effectiveness of such advertising campaigns, one must know which media channels or advertising formats have contributed to a purchase conversion. This process is known as attribution. A better understanding of attribution, or assigning conversion credit to the various relevant channels, can serve a number of research and industry purposes. Marketing managers may use such attribution models to interpret the influence of advertisements on consumer behavior and to optimize their advertising campaigns.

M Commerce: Experiencing the Phygital Retail

Emergence and evolution of information technologies paved way for the online transactions, termed as e-commerce. E-commerce was soon rechristened as wired e-commerce due to the rapid rise of mobile phone and other handheld devices facilitating access to internet while on the go and this led to the coining of the term m-commerce, the wireless avatar of e-commerce. Term Mobile commerce or simply m-commerce has been attributed to the collection of location based commercial services that are delivered by various internet enabled handheld devices such as mobile phones, tablets and palmtop devices. Though, essentially, both-e-commerce and m-commerce- are similar as the transactions are electronic in nature and are facilitated by internet but the element of mobility is one big differentiator between the two. Now, it has become difficult for the people to imagine e-commerce without the mobility dimension appended to it. In today’s online business environment, m-commerce is growing as next stage of e-commerce. Further, M-commerce has opened up new vistas for marketing the products, for targeting the customers at multiple points, for offering customized services using location based features and for enabling enhanced shopping and service experiences. It also offers opportunities for services that are more efficient and more user-friendly. The marketers may also offer new approaches, apps, and in-store solutions for enhanced customer engagement after researching the needs, roles and use contexts of the customers. Marketers have already started talking about mobile conversion funnel besides desktop conversion funnel. It is quite interesting to note that in mature markets, shape of the funnel is different and has higher purchase conversion rate.

Investigating Purchase Conversion by Uncovering Online Visit Patterns

This research aims to understand and predict online customers’ store visit and purchase behaviors. To this end, we develop a model that accounts for different patterns of online store visits at the individual level. Given the latency of visit patterns, we employ a changepoint modeling framework and statistically infer them using a Bayesian approach. The inferences obtained are then used to examine the effects of visit patterns on purchase dynamics across store visits. Using Internet clickstream data at an online retailer, we find that online store visit patterns tend to be clustered with significant variation across customers in terms of the number and size of visit clusters as well as the visit frequencies, both within and between clusters. Furthermore, the conversion rates vary significantly, depending on store visit patterns, such that they tend to be higher at later visits within a visit cluster, compared with earlier visits. The proposed model thereby offers superior fit and predictive performance than benchmark models that ignore clustered visit patterns and their impact on purchase behavior. We demonstrate the model’s ability to better identify prospective customers by utilizing their visit patterns, which can assist marketers in scoring customers and making targeting decisions across individuals for marketing activity.Data, as supplemental material, are available at https://doi.org/10.1287/mksc.2016.0990 .

The effectiveness of different forms of online advertising for purchase conversion in a multiple-channel attribution framework

The Internet has given rise to many new forms of advertising. Scientific studies have focused on individual reactions to specific advertising forms in isolation and have offered little guidance for aggregate-level budget allocation decisions, which are typically based on simple rules. This article compares the long-term effectiveness of nine forms of advertising—seven online and two offline—by means of a structural vector autoregressive model and restricted impulse responses. For five product categories, we investigate how these forms of advertising generate traffic, affect conversion, and contribute to revenue. We find that content-integrated advertising is the most effective form, followed by content-separated advertising and firm-initiated advertising. Although online advertising forms have similar power to drive traffic, content integration dominates content separation in the area of progression toward purchase. Last-click attribution underestimates content-integrated activities and suggests online advertising budget allocations that yield 10%–12% less revenue than the status quo, whereas the model's proposed online advertising budget allocation yields a 21% revenue increase over the status quo. These results highlight the payoffs for companies that integrate content into online media.

A Web Usage Mining for Modeling Buying Behavior at a Web Store using Network Analysis

Understanding visitors’ invisible behaviors and responding with appropriate answers are important issues in continually increasing online market. To promote online transactions, customers’ behavior should be predicted correctly to keep low purchase conversion rate. In this study, we suggest an approach based on the idea that customers’ sessions in a web store can be transformed into the structure of a graph, which are represented as density of a session based on a graph theory. Online users visit lots of sites and their activities include information acquisition and browsing. The history of these activities can be used to construct a relationship network among web sites. This study analyzes this visit history made by website visitors with graph theory. The density of a network refers to the differentiated degrees of relationship among objects. In this study, we dichotomize into “purchase” and “no purchase group” since predicting whether a customer will buy or not buy our products is an important issue in web stores. We collect data on sessions which are a sequence of page views or a period of sustained web browsing. We model the sessions on the basis of density of a graph, which resulted in DOS (Density of a Session). The performance of other predictors including DOS is compared to that of suggested method in this study. Predictors are TVT (Total Visit Time during a period of a visit), AVT (The Average Time per Page Viewed), TNC (Total Number of Clicks), TPP (Total Number of Product-Related Pages Viewed), and DOS (Density of a Session Based on Graph Analysis). The study found that all predictors except total visit time are useful to differentiate between “purchase” and “no purchase” group. And we conducted Logit Analysis to examine the performance of each purchase prediction method. The results from Logit Analysis show that DOS predicts purchase behavior better in comparison with other predictors. It means understanding customers’ sessions with respect to a graph structure is useful to predict whether a customer will buy or not buy products in a web store.

Learning User Real-Time Intent for Optimal Dynamic Web Page Transformation

Many e-commerce websites struggle to turn visitors into real buyers. Understanding online users’ real-time intent and dynamic shopping cart choices may have important implications in this realm. This study presents an individual-level, dynamic model with concurrent optimal page adaptation that learns users’ real-time, unobserved intent from their online cart choices, then immediately performs optimal Web page adaptation to enhance the conversion of users into buyers. To suggest optimal strategies for concurrent page adaptation, the model analyzes each individual user’s browsing behavior, tests the effectiveness of different marketing and Web stimuli, as well as comparison shopping activities at other sites, and performs optimal Web page transformation. Data from an online retailer and a laboratory experiment reveal that concurrent learning of the user’s unobserved purchase intent and real-time, intent-based optimal interventions greatly reduce shopping cart abandonment and increase purchase conversions. If the concurrent, intent-based optimal page transformation for the focal site starts after the first page view, shopping cart abandonment declines by 32.4% and purchase conversion improves by 6.9%. The optimal timing for the site to intervene is after three page views, to achieve efficient learning of users’ intent and early intervention simultaneously.

Path to Purchase: A Mutually Exciting Point Process Model for Online Advertising and Conversion

This paper studies the effects of various types of online advertisements on purchase conversion by capturing the dynamic interactions among advertisement clicks themselves. It is motivated by the observation that certain advertisement clicks may not result in immediate purchases, but they stimulate subsequent clicks on other advertisements, which then lead to purchases. We develop a novel model based on mutually exciting point processes, which consider advertisement clicks and purchases as dependent random events in continuous time. We incorporate individual random effects to account for consumer heterogeneity and cast the model in the Bayesian hierarchical framework. We construct conversion probability to properly evaluate the conversion effects of online advertisements. We develop simulation algorithms for mutually exciting point processes to compute the conversion probability and for out-of-sample prediction. Model comparison results show the proposed model outperforms the benchmark models that ignore exciting effects among advertisement clicks. Using a proprietary data set, we find that display advertisements have relatively low direct effect on purchase conversion, but they are more likely to stimulate subsequent visits through other advertisement formats. We show that the commonly used measure of conversion rate is biased in favor of search advertisements and underestimates the conversion effect of display advertisements the most. Our model also furnishes a useful tool to predict future purchases and advertisement clicks for the purpose of targeted marketing and customer relationship management.This paper was accepted by Eric Bradlow, special issue on business analytics.