MSI’s annual industry-leading symposium will convene analytics leaders, data scientists, and top researchers to investigate issues at the front line of marketing analytics and practice. Drawing on domain expertise in economics, information systems, statistics, computer science, and marketing science, presenters will address the full range of questions, opportunities, and challenges posed by abundant and increasing customer data. These top researchers and scientists will offer the latest approaches and evidence-based findings, and will discuss implications for strategic decision making. Among the topics they will address:
- Optimizing data for predictive modelling and attribution
- Moving beyond transactional data
- Combining multiple data sets for a more complete customer view
- New approaches in machine learning, AI, data visualization, natural language processing, and predictive analytics
- Leveraging data to transform the “last mile” in the buying journey
- Visual analytics
- New insights on CRM and CLV
Please note that this is a rough agenda for planning purposes only. All times are subject to change.
|8:00 – 9:00 a.m.||Registration and Networking Breakfast|
|9:00 – 9:30||Welcome and Introductions|
|9:30 – 10:00||Field Experiments to Optimize Revenue Management
JP Dube, University of Chicago
For many companies, the management of revenues tends to be
based on ad hoc rules of thumb. Companies that are making
money often do not believe they have a revenue problem.
That is, they do not consider the opportunity costs of rules of
thumb (vs. true optimization). To document these costs and offer
an alternative, J. P. Dube will discuss several field experiments
involving randomizing prices for companies to optimize their
pricing against a specific set of objectives.
|10:00 – 10:30||The Role of Customer Data in Optimizing User Experience (UX): Opportunities for Competitive Advantage
Jim Maurer, Data Scientist, Volkswagen of America, Inc.
Today, with relentless increases in informational efficiencies, customers expect high quality at competitive prices. Price and quality are seen as table stakes, and the overall user experience becomes a vital point of differentiation. For competitive advantage in UX design, brands should expand their use of customer data from targeting, prediction, and recommendation to informing an optimal UX. Jim Maurer will discuss how customer data and machine learning methods can be leveraged to engineer enhanced user experiences, and will offer examples in case studies from automotive, consumer products and digital media verticals.
|10:30 – 11:00||Break|
|11:00 – 11:30||Experimentation and Causality in Modern Digital Advertising
Harikesh Nair, Stanford University
Worldwide digital ad spending in 2018 is around $300B, accounting for about 50% of total media advertising. A significant driver of this growth is the better measurability of digital ads compared to traditional formats. Digital ads leverage sophisticated tracking to understand outcomes and exposure, and utilize experimentation to determine causality and attribution. In this talk, Harikesh Nair describes how modern ad-systems do this at scale, using examples from the ad-experimentation platform of JD.com in China. He will describe how experimentation is now delivered “as a service” via self-serve products deployed on platforms; how advertisers and brands can leverage those products to improve their marketing activities; and how modern marketplaces are moving towards continuous experimentation that seamlessly merges experimentation with optimization of marketing instruments.
|11:30 – 12:00 p.m.||A Dynamic Model of Customer Efficiency
Gary Class, Senior Vice President, Wells Fargo & Company
Gary Class will discuss how Wells Fargo has enabled a deep understanding of the key drivers of customer engagement through fine-grained observation and interpretation of routine banking activity over time. Analyzing customers’ detailed interactions with the bank’s multiple service channels has resulted in a robust behavioral segmentation scheme, “Customer Efficiency,” which explicitly identifies the customer’s role in the service delivery process. By understanding the task the customer seeks to accomplish as well as the preferred channel, the bank can identify the marketing strategy to move the client to a higher-order “Customer Efficiency” segment.
|12:00 – 12:30||Presentation to be Announced
Justine Hastings, Brown University
|12:30 – 2:00||Lunch|
|2:00 – 2:30||Consumer Friendly Advertising
Phillip Leslie, Chief Digital Economist & Vice President, Amazon
In the last 5 years Amazon has developed a significant advertising business, consisting of both brand and performance advertising products that are available for advertisers that sell products at Amazon as well as elsewhere. These ad products are designed to create value for advertisers, as well as enhancing the shopping experience of Amazon’s customers. Phil Leslie will discuss how Amazon uses numerous inputs to accomplish this, including the signals from advertisers (campaign inputs) as well as contextual and behavioral signals from customers. These signals and advanced modeling techniques have allowed Amazon to develop consumer friendly advertising that delivers performance to its advertisers.
|2:30 – 3:00||From Cookies to Customers
Sarah Moshary, University of Chicago
Many companies collect data on visitors to their website, tracking detailed behavior about cookies: time on the site, pages viewed, and traffic source, among others. These data are ideal for offering insight about session-level outcomes, but firms are often more interested in optimizing for a longer horizon for users. A single user might touch a website on multiple browsers, computers, or mobile devices, generating multiple cookies. This fragmentation poses a problem for simple A/B tests run at the cookie-level. How can firms map cookies back to users? How can we design experiments to avoid the problems of fragmentation? In this talk, Sarah Moshary will discuss recent advances in the marketing literature that seek to address these problems.
|3:00 – 3:30||Break|
|3:30 – 4:00||Measuring Consumer Sensitivity to Audio Advertising Load: A Field Experiment on Pandora Internet Radio
David Reiley, Principal Scientist, Pandora
In an era of ad-supported content provision, field experiments offer a crucial way to measure consumer sensitivity to advertising. David Reiley will present key insights and findings from a randomized experiment with almost 35 million Pandora listeners over 21 months. This duration allowed Pandora to determine the long-term effects of different levels of ad exposure. Among these: (1) the number of hours listened decreased linearly in number of ads per hour, resulting in a strikingly linear demand curve, and (2) increased ad load caused a significant increase in paid ad-free subscriptions to Pandora.
|4:00 – 4:30||Optimizing Messages
Sanjog Misra, University of Chicago
Constructing and implementing messages is fundamental to the task of marketing. New developments in technology and machine learning as well as the emergence of scalable experimentation have facilitated the emergence of algorithmic solutions to what has been traditionally been thought of as a creative exercise. Can algorithms create realistic and relevant marketing messages? Are such messages better from an objective point of view? Can the messages be personalized? In this talk Sanjog Misra will present ongoing research that discusses the answers to these and other related questions.
|4:30 – 4:45||Reflections and Closing Remarks|
|4:45 – 6:15||Networking Reception|
|8:00 – 9:00 a.m.||Registration and Networking Breakfast|
|9:00 – 9:15||Opening Remarks|
|9:15 – 9:45||Precisely Wrong: The Limits of Targeting
Karin Kricorian, Director, Management Science and Integration, The Walt Disney Company
Despite our decades of effort, marketers’ understanding of consumers remains remarkably poor. The explosion of new data and the popularization of new analytic techniques seems like a godsend: we will finally know which of our marketing is working and be able to spend our media dollars with laser precision. However, quantity doesn’t always equal quality and this abundance of data is sometimes an illusion. In this presentation, Karin Kricorian will discuss the challenge of navigating between the dangers of over- and under-targeting and methods for doing so effectively.
|9:45 –10:15||Psychological Targeting and Tailoring in Advertising
Michal Kosinski, Stanford University
A growing proportion of human activities such as social interactions, entertainment, and gathering information, are now mediated by online social networks. Such activities can be easily recorded, offering an unprecedented opportunity to study and assess psychological traits using actual―rather than self-reported―behavior. In this presentation, Michal Kosinski shows that digital records of behavior, such as Tweets or Facebook Likes can be used to accurately measure a wide range of psychological traits. Such Big Data assessment has a number of advantages: it does not require participants’ active involvement; it can be easily and inexpensively applied to large populations; and it is relatively immune to cheating or misrepresentation. Furthermore, he will show that resulting psychological profiles can be used to optimize the targeting and tailoring of marketing communication. If the ethical challenges can be overcome, psychological assessment based on Big Data has the potential to revolutionize marketing.
|10:15 – 10:45||Break|
|10:45 – 11:15||Connecting the Dots: Data-Informed Approach to Improving Marketing Performance in B2B
Anne-Lise Kopp, EU Marketing Manager, Performance & Insights, Shell
Ivan Vlahinic, Senior Manager, Data Science, Ekimetrics
It’s time to stop being “data-driven” – and start being “customer-driven but data-informed”. We are caught in a data deluge today. Most organizations, whether they know it or not, are already sitting on tremendous data assets, with pace of digitization only intensifying. What is lacking instead is a way to connect key business questions to these data points, in the service of one’s customers. In this talk, Anne-Lise and Ivan will discuss their partnership and how they’ve used the pragmatic approach above to generate a step improvement in the way they measure impact from B2B marketing investments – and with the feedback loop in place, how they will be driving toward more optimal marketing activities going forward. The outcome of the collaboration is an integrated set of capabilities covering, market share tracking, customer lifetime value, and business performance lenses.
|11:15 – 11:45||State-of-the-Art Methods for Targeting Interventions
Dean Eckles, Massachusetts Institute of Technology
All marketers face the question of who to target with costly marketing interventions (e.g., advertising, discounts, detailing). Dean Eckles will illustrate state-of-the-art in methods for evaluating and optimizing targeting in two settings: first, using a series of experiments to iteratively improve targeting of discounts as part of pro-active churn management at the Boston Globe, and second, using existing data from field experiments in networks to evaluate strategies for seeding a product or behavior in networks. His presentation draws on joint work with Sinan Aral, Alex Chin, Paramveer Dhillon, Johan Ugander, and Jeremy Yang.
|11:45 – 12:00 p.m.||Closing Remarks|
MSI has secured a block of rooms at the Conrad Chicago at a discounted rate of $275 per night for a single or double room. In order to secure a room at the group rate please book online here, or contact the Conrad at (844) 676-2522 and reference the group code “MSI”.
Availability and the discount for rooms can only be guaranteed until Monday, October 14th, so please make sure to book your accommodations early.
101 E. Erie St.
Chicago, IL 60611