• IDE Projects 2015

    Please consider this as a sample of the research that we have currently underway. If you would like to have further information on any research at the IDE, please contact David Verrill (dverrill@mit.edu, 617-320-4923)

    Hint: Use ctrl+F to search by researchers name

    Financial Indicators from Big Data
    Professor Alex Pentland
    Theories of social physics tell us that human mobility and socialization patterns, along with diversity of behavior and communication patterns, are leading indicators of disposable income and social health.  We are using these patterns to mine telecommunications data and credit card data in order to provide a wide range of financial and social indicators, including personal and small business credit scores, development indices for neighborhoods and cities, and GDP and consumer attitude scores for entire economies.

    Sustainable Digital Ecology
    Professor Alex Pentland
    In the era of Big Data there is danger that the economic environment will be dominated by those with the biggest computers and most access to consumer personal data.  This is not just an economic worry; the scandal surrounding the US National Security Agency shows that similar dangers exist with governments.   Moreover, an ecology with such winner-take-all dynamics is likely to be unstable, with severe booms and busts.   This project investigates alternate architectures for a digital ecology that are more sustainable, such as moving from completely open market mechanisms to cooperative exchange network mechanisms.  Examples of such exchange networks include the SWIFT interbank exchange network or the Hawala networks in the Middle East.

    Digital Technologies and Their Impact on the Earnings Prospects of American Workers

    Professor Erik Brynjolfsson and Dr. Andrew McAfee

    This project focuses on an under-studied yet critical issue: the impact of digital technologies on the earnings prospects of American workers. The effects of the Great Recession, globalization and outsourcing, education, and many other factors have been extensively studied. Computerization, meanwhile, has received less attention. There is general agreement that technology often favors more highly skilled and educated workers. There is also a widely shared view among economists that the Luddites were wrong, and that more technology will not lead to fewer jobs. The research supporting these conclusions, however, was conducted before the Great Recession gave way to a dismayingly jobless recovery; before computers started displaying science-fiction abilities like driving cars, writing clean prose, and beating the best human contestants on Jeopardy!; and before the rise of crowdsourcing problems ranging from Innocentive to TaskRabbit to Amazon Mechanical Turk. The broad vision of this 3-year grant is to consider and utilize quantitative and qualitative studies that will inform us on digital technology trends and their impact on the earnings prospects of American workers.

    The Role of Advanced Technology and Automation

    Professor Erik Brynjolfsson and Dr. Yousef Alhammadi 

    This proposal will address the pros and cons of automation in a pedagogical method.  It is the objective to establish a set of metrics that based on well-established surveys and case studies.  This proposal will give specific consideration to key industry sectors including aerospace, semiconductor as well as construction.  The specific goal of proposed research is two-fold: (1) re-examining the job market dynamics in general and more specifically in Abu-Dhabi, for the coming decade, under highly automated job tasks. This can be analyzed within different industrial sectors including strategic industries (such as oil and gas), knowledge-intensive industries (e.g., aerospace), labor-intensive industries such as construction and civil infrastructure; governmental sector; or renewable energy sector (2) Quantifying the economic gains and the accompanying socioeconomic effects, under different designed scenarios.  The ultimate goal of this research is to consolidate the study findings into a balanced roadmap to manage the newly established job-market dynamics as applicable to Abu Dhabi. The research will be multifaceted and will explore legislation, technical solutions, and business models to encourage economic growth, job creation and technological advancement, in addition to attracting and incubating entrepreneurial ideas.

    The Attention Economy: Measuring the Value of Free Goods on the Internet

    Professor Erik Brynjolfsson and Dr. JooHee Oh

    Over the past decade, there has been an explosion of digital services on the Internet, from Google and Wikipedia to Facebook and YouTube. However, the value of these innovations is difficult to quantify, because consumers pay nothing to use them. Traditional approaches based on measuring prices and quantities do not work well for goods with zero price.  In this project, we develop a framework to estimate the value of consuming information goods on the Internet that is not measured in the traditional money-based measure of GDP. Our model of the “attention economy” yields an estimate of annual change in the value of online applications that have very low prices using the insight that even when people do not pay cash, they must still pay “attention,” or time. We estimate the increase in consumer surplus created by free internet services to be over $30 billion per year in the U.S. alone, about 0.23% of average annual GDP during the period of 2002-2011. We apply our methods to television and estimate that the level of consumer surplus from television is more than three times larger, but the annual increase is comparable to the increase in surplus from the Internet. Our analysis implies that most of welfare gain from digital services on the Internet would be overlooked by traditional approaches that rely only on the direct expenditures of money. Considering the time spent on consumption, as we do, makes it possible to assess the full value of these digital innovations.

    User Investment and Firm Value: Case of Internet Firms

    Professor Erik Brynjolfsson, Dr. Seon Tae Kim, and Dr. JooHee Oh

    This project quantifies business implications of the recent rapid growth of the amount of users’ online activities. Users collectively spend billions of person-hours creating user generated contents which is then becomes an asset that produces a stream of value for fellow users, and of course, the host site itself, labeled as user generated capital. This fact has important implications for business strategies and values of Internet companies because growth opportunities of many Internet firms are determined largely by the user activities, in particular, the users’ investment of their valuable contents and time. The relationship between the amount of users’ online activities and an Internet firm’s value has been, even though a popular topic discussed in press or in policy debates, under-studied in the academic literature and is studied both empirically and theoretically in this project. More specifically, we focus on addressing three questions: (i) what is the mechanism of user generated capital itself, (ii) what is the consequence of user generated capital for an Internet firm’s value, and (iii) what are the Internet firm's optimal strategies to manage user generated capital. For such a purpose, we document facts on the relationship between an Internet firm's market value and its tangible and intangible capital and the amount of users' online activities. Then we study them in a dynamic general equilibrium model of the user's time-allocation between online and offline activities. This project aims to enhance understanding more broadly intangible capital by studying a particular class of intangibles, user generated capital.

    The Co-Evolution of Income and Product Sales Distributions on eBay

    Professor Erik Brynjolfsson and Sagit Bar-Gill

    Online markets for a large variety of goods have been shown to exhibit long tail properties, with significant aggregate sales of niche products. Mathematically, this translates into sales distributions that follow power laws. Power laws have also been documented in firm-size and personal income distributions, but a study of both income and product sales distributions and their dynamics in a single marketplace has yet to be carried out.  EBay is the ultimate online marketplace, with annual net revenue of over $15 billion, a population of nearly 150 million active buyers, and over one million sellers who make their living on eBay. In the proposed project we plan to explore eBay’s long tail, in both product sales and seller income distributions, and their co-evolution over time. We will then explore possible drivers of the trends observed, and of individual mobility within the income distribution. The study of eBay’s income distribution will thus provide meaningful insights on economy-wide trends.


    Measuring the Data Equity of Top American Companies

    Professors Erik Brynjolfsson and Kristina McElheran

    The objective of this project is to determine the “data equity” that top American firms enjoy as a function of their use of data and information technology.  It will address the question of whether firms that use data more intensively or more effectively tend to enjoy higher valuations and better performance outcomes.  The primary deliverable will provide insight into the data use-performance relationship for top public firms in the U.S., but it will also be informative of IT strategy more generally. 

    Digital Technologies and Their Impact on the Earnings Prospects of American Workers

    Professor Erik Brynjolfsson and Dr. Andrew McAfee

    Sponsor:  Markle Foundation

    This project investigates an under-studied yet critical issue: the impact of digital technologies on the earnings prospects of American workers.  Some of the news is clearly bad. The median American household earns less than it did fifteen years ago, labor's share of national income peaked in the early 1980s has been falling pretty steadily since then, long-term unemployment has emerged as a vexing problem, and startups are creating many fewer jobs than they used to. Technological progress has been linked to these and other trends, including by us in our 2011 short book Race Against the Machine.  However, there is also some good news -- reason to believe that technology is helping at least some workers and households earn money in ways that were not previously possible eBay and newer marketplaces like Etsy allow people to reach large markets for the goods they make and trade, and startups like AirBnB and RelayRides let people earn incremental income by renting their houses and cars in lightweight and flexible ways.  Some other recent technology developments, however, are ambiguous at this point; it's not clear if they're good news or bad news for workers overall. For example, there is now a large variety of online labor sourcing platforms including (but certainly not limited to) Innocentive, Liveops, oDesk, 99Designs, TaskRabbit, Samasource, Crowdflower, and Amazon Mechanical Turk. Are they providing much-needed supplemental income, putting people out of work, or both? Are they tools of economic empowerment for individuals, or digital sweatshops? It's also not clear if the new generation of robotics coming very soon to factories, warehouses, and distribution centers will assist human workers or replace them.

    The Economic Impact of Advanced Technology and Automation on the Oil and Gas Sector

    Professor Erik Brynjolfsson (MIT IDE) and Dr. Mohammad Omar (Masdar Institute, Abu Dhabi)

    Technology is having profound effects on the nature of jobs in economies around the world.  While past innovation has typically created more jobs than those displaced, the pattern has been changing in the recent years.  The new dynamic challenges the efforts of governments around the world to sustain positive job creation trends, and represents perhaps the greatest economic challenge and opportunity of our era.  An additional challenge for the Oil and Gas sector in Abu Dhabi is managing these changes in a workforce that contains a significant number of expatriate workers.  Leadership in automation may help to the Emirate to maintain positive economic growth in the Oil and Gas sector while sustaining a balanced socioeconomic and political environment.  What skills and occupations are most likely to be displaced by technology?  Which ones will likely continue and even grow?  How can governments foster policies that will lead to sustainable job trends for the foreseeable future?  This project analyzes these questions by using rigorous econometric analyses and case studies to advance the frontier of economic research while informing a framework for policymaking in the Abu Dhabi 2030 vision. 

    Digital Organization: How fast-moving digital innovations are changing the way we work and manage

    Deborah Soule and George Westerman  

    As the digital age unfolds, new opportunities arise continually to transform customer experience, internal business processes, and business models. But far less attention is being paid to the design of the organization that must fulfill the new digital strategies. This project draws on interview and survey data to: 1) understand how digital technologies are influencing organization design across a variety of industries, 2) identify the characteristics that are particularly effective in enabling digital dexterity, and 3) provide insight into leader and workforce experiences of “becoming a digital organization.”

    Digital Leadership: How leader behaviors influence digital practices, attitudes and outcomes

    Deborah Soule and George Westerman    

    Sponsor: Haier Corporation

    Earlier IDE research identified practices that are useful in leading and profiting from digital transformation. This research extends our understanding of how specific digital mastery practices relate to key design characteristics of the digital organization. Through deep interview and survey investigation of a single complex company, we will examine how new technologies enable new organization designs that provide new competitive advantage.  We will also investigate the ways in which leadership practices influence the formation of attitudes, skills, practices, and resources fit for the digital future.

    The Economic Impact of Skill-Biased Technical Change

    Erik Brynjolfsson, Frank MacCrory, and George Westerman

    Rapid advances in digital technologies have profound implications for work. Many middle and low skill jobs have disappeared, contributing to increasing inequality, falling labor force participation and stagnating median incomes. Yet this stagnating median masks a surprising amount of variation among different groups. A large literature shows that skill-biased technical change (SBTC) has led, on average, to real wage declines for those with less than a college degree and real increases for those with more education.  However, college education is not a simple on/off trait, and skill is not a single dimensional construct. How do education and skills interact with technical change to affect the labor market?  Using rigorous econometric analysis, we identify distinct dimensions of skill.  Then, we assess their interactions with education and technology to affect wages, occupational demand, and other economically important outcomes.

    Measuring Organizational Strain as a Leading Indicator for Operational Execution Risk

    Erina Ytsma and George Westerman

    Sponsor: MassMutual

    Organizational and IT changes, combined with unforeseen changes in external demands, can create strain on employees and organizational units. Organizational strain can reduce performance in the short term while also creating long term risks in areas such as compliance, capacity, and agility. This project aims to construct a statistical model of early indicators of organizational strain as well as policy guidelines to prevent or alleviate strain.  We will construct potential indicators of strain using detailed microdata, and link the indicators to organizational unit performance and risk.  By selecting units appropriately, we will identify natural experiments which we can exploit to further test our models of organizational strain and outcomes.

    Personality Traits and Job Performance: A Novel Approach Using Short Games

    Erik Brynjolfsson, Hossein Ghasemkhani, and George Westerman

    Many companies use paper assessments and structured interviews to screen new job applicants or to help employees improve their performance. These instruments, while useful, have limited validity.  In particular, they can be subject to employee manipulation and unconscious rater biases, and do not measure actual behaviors.  A new class of assessments attempts to reduce potential biases by measuring a large number of difficult-to-measure traits, such as persistence and creativity, through actual behavior in online games.  Our research project aims to investigate the relation between employee personality traits and performance in different occupations using these new digital games.  Research questions include: How do personality traits vary across job categories?  What personality traits are associated with performance in each job category?  How do digital behavior assessments compare with more traditional assessments such as Myers Briggs?

    Evaluation Schemes for Innovation Contests

    Professor Sam Ransbotham (Boston College) and Dr. George Westerman (MIT IDE)

    The innovation consulting industry has created hundreds of potential prescriptions ranging from open innovation tools to 20% time to broad culture training.   Most methods are promoted as silver bullets that will improve an organization’s innovativeness with one simple shot.  Yet evidence about the effectiveness of each method, or even how to do each method well, is often much less powerful than the sales pitch.  We examine one particular innovation method, the innovation contest.  We compare two different methods of evaluating innovation ideas using detailed data on more than ten thousand submissions over four years of a large firm’s successful process.  This contest, which requires very little direct investment, has generated more than fifty winning ideas, many of which are now moving into production.  It has also been credited with instilling a culture of innovation and collaboration among many employees who would not have seen innovation as part of their roles in the past.  The selection mechanisms exhibit different cost profiles and different drivers of success.  A combined approach can improve selection while optimizing use of expensive resources.    The findings can also energize employees, provide advice to potential innovators, and help the firm identify hidden innovators across the organization. 


    Improving Automotive Demand Predictions Using Online Activity Data

    Dr. Sagit Bar-Gill and Dr. Shachar Reichman   

    Brands’ website traffic, along with non-proprietary online resources, reflect consumers’ interests and intensions as part of their decision making process. We thus measure the relationship between traffic on the BMW website, and both Google and Wikipedia searches for the brand, to offline car sales. We then develop models for automobile market-level sales predictions, and test these models’ performance on recent sales numbers. We then proceed to construct prediction models for customer level purchase and churn probabilities, based on fine-grained CRM data, along with individual browsing patterns. These models assign purchase and churn scores for the brand’s existing and prospective customers, allowing for personalized marketing activities.



    Income Mobility and Distribution in E-Commerce: The Case of eBay Sellers

    Professor Erik Brynjolfsson (MIT IDE), Dr. Sagit Bar-Gill (MIT IDE), Dr. Peter Coles (eBay), and Nir Hak (Harvard)

    We explore the dynamics of sellers in the eBay marketplace, studying trends in sellers’ income distribution, and individual level mobility. We find that in the past decade income inequality has been rising on eBay, as the Gini coefficient for professional sellers’ revenue has been monotonically increasing. We further find that seller revenue exhibits a Power Law distribution, and that the tail of high earners has become heavier over the past decade. The documented increase in inequality is thus driven by top sellers that are commanding a growing share of total income on eBay. This reflects a pattern that has received much attention in the context of individual income dynamics. We then move from distribution level analysis to study success and mobility drivers at the individual level. We find that sellers with very diverse offerings or else very specialized offerings are most likely to experience growth, and that upward mobility has increased over time. This ongoing project will further examine the role of IT tools provided by the eBay platform on mobility levels, for sellers with different volumes of operations. The study will include several experiments conducted in collaboration with eBay. Since eBay serves a spectrum of small and large sellers, this research will shed light on overall trends in income distribution and mobility in e-commerce.

    When Online Engagement Gets in the Way of Offline Sales: A Natural Experiment

    Dr. Sagit Bar-Gill (MIT IDE) and Dr. Shachar Reichman (MIT IDE and TAU)

    E-commerce platforms and brand websites have become an integral part of almost any consumption process, and both online and offline retailers have been focusing their efforts on improving their online presence and enhancing firm websites. One of the challenges still faced by researchers and practitioners alike is to understand and measure the end effect of web presence on offline sales, and how it is mediated by the effect on online consumer behavior. Online engagement and information gathering serves as a substitute to offline information acquisition, while also complementing offline consumption processes by increasing the probability that a consumer arrives at a store or dealership, and by increasing conversion rates offline. In this research, we exploit a natural experiment setting, to study the effect of the launch of a new interactive website by a leading automaker, on the brand’s offline car sales. We find that website launch lead to a decrease in sales, as online engagement lead to a decrease in sales leads and subsequent dealership contact. Our setting provides a unique opportunity to identify a causal effect of online engagement on offline sales. Furthermore, the research offers a practical contribution – the negative effect we identify suggests that managers should be cautious in setting high online engagement goals for offline products and services.

    Identifying the Role of Social Media in Brand Loyalty and Desertion

    Dr. Sagit Bar-Gill (MIT IDE) and Dr. Shachar Reichman (MIT IDE and TAU)

    Online word-of-mouth for a non-digital product (such as a car) is likely to be triggered by an offline experience.  A trigger experience may be the purchase of a new car, a positive or negative service experience, or even advertising and promotional activities (where these may occur offline or online). Online social networks therefore have the potential of increasing both the magnitude and reach of brand-related word-of-mouth. In this research we will construct a comprehensive database of BMW’s CRM and service data, activity on the BMW website, and talk of brand-related experiences in online forums and social networks. We will then analyze the effects of social media traffic on offline repurchasing and desertion behavior by both originators and followers of brand-related online word-of-mouth, further measuring the spillovers of one customer’s offline experiences on fellow customers’ decisions, created by social media.

    Are Occupational Licenses Needed in a Digitized World?

    Professor Brynjolfsson (MIT IDE), Professor Chiara Farronato (Harvard Business School), Dr. Andrey Fradkin (MIT IDE), and Professor Bradley Larsen (Stanford)

    The inability to verify the quality of a service provider represents a major challenge to the functioning of markets. One traditional solution to this problem is for the government to regulate services through certification or mandatory licensing. These regulations are meant to ensure that consumers receive quality service by excluding unqualified sellers. However, occupational licenses also have the potential drawbacks of raising the prices consumers face and rationing workers out of the market. Furthermore, although there are over 800 licensed occupations, little is known about the efficacy of licenses in raising the level of service. At the same time, digital reputation systems provide an alternative way for consumers to match the service providers. These reputation systems aggregate high-frequency information about the quality of a service and use it to help consumers match with appropriate service providers. In this paper, we use data from Thumbtack.com, a large online marketplace for professional services, to study how licensing affects market outcomes, how licensing compares to online reviews, and whether consumers value licensing information. Thumbtack is the perfect environment for studying these questions because consumers are able to see the reviews and licensing information of bidding service providers. Therefore, we can use this data to measure whether more stringent licensing leads to higher quality services and how consumers value licenses and online reviews.

    Search, Matching, and Discovery on the Internet

    Dr. Andrey Fradkin (MIT IDE)

    Online marketplaces have transformed how consumers search and match in the labor, housing, and dating markets through technologies such as search engines, reputation systems, and mobile apps. Not only have these platforms reduced the cost of search but, for transactions such as short-term apartment rentals (Airbnb), cross-border professional services (Odesk and Freelancer), and rides (Uber and Lyft), they have greatly expanded the size of their respective markets. However, little is known about how matches form on these platforms, the role of technology in creating those matches, and the potential for making matching more efficient through better marketplace policies. Furthermore, although search and matching plays a pivotal role in theories of market failure, data on this process has historically been scarce and often incomplete. The comprehensive data regarding browsing, communication, and transactions collected by these marketplaces allows for novel tests of search and matching theories. This project uses this type of data from Airbnb, as well as new models of search, to study the role of technology in making the search and matching process in this marketplace more efficient.

    The Economic Effects of Peer-to-Peer Online Marketplaces

    Professor Chiara Farronato (Harvard Business School) and Dr. Andrey Fradkin (MIT IDE)

    Online marketplaces have reduced entry costs across a variety of industries. These marketplaces allow small and part-time service providers (peers) to participate in economic exchange. For example, Airbnb allows anyone to become a hotelier and Uber allows anyone to become a cab driver. The entry of peer-to-peer competitors has two effects: market expansion and business stealing. The first effect occurs when the peer-to-peer sector supplies price sensitive or niche consumers who were previously underserved. The second effect occurs when the peer-to-peer sector attracts consumers away from conventional suppliers. We study these two effects using data from the hotel industry and Airbnb. We show that the market expansion and business stealing effects differ by location, and attribute this heterogeneity to supply constraints - legal and geographic - relative to the level of demand. We then derive a simple model of competition between a peer-to-peer marketplace and hotels to explain these findings. In the model, hotels and peer-to-peer suppliers differ in their fixed (higher for hotels) and marginal costs (higher for peer-to-peer suppliers). The model allows us to study how the efficient market structure depends on the level and variability of demand, and to quantify the welfare gains from peer-to-peer entry in the accommodation industry.

    The Impact of Unemployment Insurance on Job Search: Evidence from Google Search Data

    Professor Scott R. Baker (Northwestern University) and Dr. Andrey Fradkin (MIT IDE)

    Job search is a key choice variable in theories of labor markets but has proven difficult to measure directly. We develop and validate the Google Job Search Index (GJSI), a measure of job search based on Google search data. The GJSI is publicly available, high-frequency, and location-specific. We use the GJSI to study the effects of the unemployment insurance (UI) expansions between 2008 and 2011. We show that states with higher potential durations of UI experience decreased search. A calibration exercise suggests that the decrease in job search due to UI expansions had minor effects on unemployment rates.

    The Design of Online Reputation Systems

    Dr. Andrey Fradkin (MIT IDE), Elena Grewal (Airbnb), David Holtz (MIT and Airbnb), and Matthew Pearson (Airbnb)

    Reviews and other evaluations are used by consumers to decide what goods to buy and by firms to choose whom to trade with, hire, or promote. However, because potential reviewers are not compensated for submitting reviews and may have reasons to omit relevant information in their reviews, reviews may be biased. We use the setting of Airbnb to study the determinants of reviewing behavior, the extent to which reviews are biased, and whether changes in the design of reputation systems can reduce that bias. This first stage of the project uses large scale field experiments on Airbnb to measure the size of bias in reviews. The second stage uses these experiments to measure the effects of reviews on seller outcomes and overall market efficiency.


    Storytelling in Social Media – The Impact on Brand Relationships

    Professor Glen Urban

    Social media offer many opportunities to influence brand relationships.  Most firms have Facebook pages and post stories and brand items to pages of their fans.  But social media has evolved to become much more than a tool for gathering information. Today, social media has gotten more complex with real-time micro-blogging updates and combined interactions between Facebook, Twitter, Instagram, YouTube, and Pinterest.  Consumers are increasingly taking control of social media with, for example, Twitter comments that may be negative as well as positive.  Brands can broadcast brand-to-customer (B2C) measures, but cannot control the consumer-to-consumer (C2C) information very tightly.  Consumers are telling real and imagined stories about their experiences and expectations about brands.  The purpose of this project is to better understand the impact of social media and, specifically, storytelling upon consumer-brand relationships with BMW.  In 2013 BMW funded a project to examine the effects of social media exposure on brand image and consideration.  This scope of work is for continuing and expanded work in this area of making social media more effective in brand building and sales generation.

    The Digital Organization of the Future

    Drs. George Westerman and Andrew McAfee
      Capgemini Consulting

    The Center for Digital Business at MIT engages in fundamental research on the effects of digital technology on business and the economy.  Previous work by the research team investigated how large firms are using, and benefiting from, new digital technologies.  It identified technology and management drivers of digital mastery, and their implications for the performance of firms.  The CDB has also done pioneering research on the effects of recent technology advances on workers and the economy.  More remains to be learned, however, about the digital organization of the future.  Exponential growth in the price/performance of technologies will have immense impacts on the nature of work, the design of organizations, and the shape of industries.  What will ubiquitous information do to organization structures?  As computers move from replacing manual labor to replacing knowledge work, what will be the implications for workers?  How will large incumbent firms compete against networks of new digitally-powered competitors?    Answers to these questions are obviously important, yet have been lacking.

    Prospective Meta-Analysis to Improve Advertising and Productivity

    Professor Glen Urban

    Sponsor:  General Motors

    Prospective Meta-analysis is used in the Medical Field to establish causative links between treatment and results (e.g., a recent meta-analysis of 14 clinical studies and 90,000 patients established that statins do reduce cholesterol and death due to heart disease).  With this project we will apply a clinical experimental methodology to advertising to develop causative links between media and steps in the purchase process model.  With testing the deep funnel links of advertising exposure to consideration, brand perception and preference, MIT can then model the sales impact of changes in advertising spending and the use of the most recent and established media.  As an example, how effective are apps, social network communications, blogs, video banners, advanced search links versus, TV, brochures and print?  MIT’s aim is to measure the ROI of advertising and to improve the productivity of advertising.

    Digital Leadership: Assessment and Analysis

    Drs. Andrew McAfee and George Westerman

    Sponsor:  Haier

    The initial focus of digital leadership research is to focus on assessing the state of digital maturity across Haier’s major business units, and recommending appropriate digital leadership mechanisms.   We will focus on social media and big data analytics, but also gather some data on mobility and potentially embedded devices.  Additional data gathering on organizational design characteristics will be useful for research purposes and for further planning by Haier’s digital leadership team.

    From Products to Platforms

    Professors Erik Brynjolfsson, Marshall Van Alstyne and Geoffrey Parker

    Sponsor:  Haier

    The first objective will be to undertake a “platform maturity” review.  This will address the following questions – how are (a) current technological, (b) economic (incentives), and (c) organizational strategies aligned with a platform strategy?  An important output from the review will be to provide guidance on linking platforms across white goods, electronics, and the connected home ecosystems.  The second objective, based on the maturity review, will be to suggest developer strategies to assist with innovation.   How can Haier open a platform?  How can Haier entice developers and 3rd parties to contribute to their platform?  How can Haier innovate at a faster pace using a platform strategy? 


    Platform Foundations

    Professors Erik Brynjolfsson, Marshall Van Alstyne and Geoff Parker

    Sponsor:  MassMutual

    Over MassMutual is in the early stages of creating a new social community (comprised of both physical and digital experiences) for prospective insurance customers.  The community is an example of a platform-mediated market, which is subject to specific and often non-intuitive strategic rules.  The new platform may be a way to inform current and potential customers, provide customer support, serve as a test bed for new market ideas, identify leads, and allow for long-term interactions with customers in an environment of trust and transparency.  It can also be a foundation for big-data analytics into customer behavior and preferences.  In a multi-year collaboration, The MIT Center for Digital Business will use its deep experience in platform strategy and big-data analytics to conduct research and advise MassMutual in these areas.  Common products such as personal computers, cell phones, streaming media, telecommunications infrastructure, travel services, and financial services can be described in terms of systems where developers build “applications” on top of a “platform.”  Such systems have become an important way in which firms go to market and organize innovation.  The goal year one is to investigate foundations.  We will describe platform market foundations and business models that firms use to participate in platform markets, especially in financial services.  As part of the effort, we will also lay the foundation for future analytics efforts on existing datasets and, eventually, data from the new platform itself.

    The Future of Prediction: Reducing Uncertainty in Enterprise Risk Management

    Professor Erik Brynjolfsson, Shachar Reichman and George Westerman

    Sponsor:  MassMutual

    We have demonstrated how data from search engines like Google provide a highly accurate but simple way to predict future business activities. Applying our methodology to predict housing market trends, we found that a housing search index is strongly predictive of the future housing market sales and prices in U.S. states. Other studies have shown the power of using twitter and other data streams to predict economic indicators such as inflation or unemployment, or trends such as flu outbreaks.  This kind of prediction has not been applied to the area of risk, where prediction is traditionally difficult and the cost of failing to anticipate an incident is very high.  We will investigate the use Google searches along with other indicators to improve upon existing methods of estimating risk likelihood and predicting events.  In this project, we plan to extend our approach to move from economic indicators and product sales to a novel and important area for prediction: enterprise risk.  In particular, we will investigate the power of Google search data, combined with other data sources, to predict risk likelihood in two areas:  counterparty credit default risk, and cybercrime scenarios. 

    Distilling Social Media Effects on Brands

    Using Dynamic Models to Enhance the Development, Application and Performance of Electronic Sales

    Dr. Michael Siegel

    Sponsor:  MassMutual

    The last decade has seen a huge rise in the ability of organizations to support a distributed workforce. Along with this trend, organizations are creating value from the large amounts of data collected from this distributed workforce. Mass Mutual has developed a new set of applications to which comprise its Electronic Sales Program (ESP). The goal of this research is to develop dynamic simulation model(s) to enhance the performance of ESP in the Mass Mutual agent ecosystems. Our analysis will use System Dynamics Modeling (SDM) methodology to develop and simulate quantitative models that linked software dynamics, key actors, and management systems to estimate the costs and benefits of various business approaches. The project will utilize complementary research techniques, including interviews with multiple constituencies, to develop a holistic view of the business approach and challenges. We believe the interactions and results of this research will have a direct effect on the evolution in the agent ecosystem. 



    Measuring Synergy Between Traditional and Digital Media

    Professors Glen Urban and Renee Gosline

    Sponsor:  MassMutual

    MIT has conducted a number of research projects on new and old media and their interactions.  The focus of this research has been on allocation between TV and Google/Facebook/banners and storytelling on social media.  MIT is interested in extending this stream of research by modeling the synergistic effects of TV and social media.  In this project we will undertake a controlled behavioral experiment in a consumer panel.  This is a five cell design and a sample of 1,500 consumers with forced exposure to alternative media in a natural environment.  This would be a pre/post and test/control design.  These measures will be integrated in a customer behavioral flow model which calculates the effects of the flow from awareness through brand attribution to trust and finally to consideration of MassMutual. 


    Mobile App Marketing  

    Professor Glen Urban

    Sponsor:  Suruga Bank

    In the fall of 2013 we will finish the simplified iPhone Smart Move app and will build a series of Android apps to meet the requirement of the Android phone suppliers.  We plan on a sample size of 1,500 movers in Japan in the winter of 2013.  We will conduct the field study, analyze the results and examine the marketing implications for Suruga in the spring of 2014.  Measures will include familiarity, consideration, brand ratings, preference and probability of purchase.  The change of these measures due to exposure to the app or advertising will be recorded.  The spending implications of implementing the app will be researched.  In addition to the Smart Move research, in spring 2014 we will begin a “story telling” project which will combine new media with the classic power of storytelling in a new app Suruga can use to generate new business.  The initial idea would be for consumers to share stories with each other about moving (or mortgages) in an effort to provide convincing information that Suruga is a trusted supplier of loans and should be high on the consideration list.  The design will include the app features and a prototype to test feasibility of storytelling by consumers in the social media world.

    Prediction – How Search Engines Foreshadow Financial Products Markets

    Professor Erik Brynjolfsson and Shachar Reichman

    Sponsor:  Suruga Bank

    We intend to develop a model that incorporates search query volumes into financial products (e.g. loans and deposits) demand predictions. We will train the model on public financial data (interest rates, income distribution, geographic location and etc). We will quantify the improvements offered by our model in predictions of overall sales and of the demand for certain financial product.  Specifically, we will use three different techniques to automatically identify correlates among searched keywords:  algorithmic methods, thesaurus, and crowdsourcing.  Our plan is to compare the performance of these methods in general and for specific categories of search queries and create a hybrid model that will efficiently combine the best method for each type of term.

    The Trajectory of Academic Papers and Researchers

    Professor Erik Brynjolfsson

    Sponsor:  Thomson Reuters

    In this research project, we will develop a method for predicting the trajectory and future importance of an academic paper; and an academic dashboard that predicts the trajectories of individual researchers.  We plan to utilize data from Thompson Reuters, Google scholar, SSRN, and the Mathematics Genealogy Project.  We intend to create an index, called “paper-rank”, that takes into account not only the number of citations for a paper, but also the importance of the papers, who cited the paper, and related metrics, based in part on ideas from the Pagerank algorithm that Google uses to rank websites. We further plan to make predictions about outcomes regarding papers by using information on citations over time with data mining and machine learning methods.  For example, we believe we will be able to predict which papers will receive major awards, which papers are destined to become important in an academic field, etc.  The key in this task is to look at the overall network of citations and the dynamic evolution of those citations.   Our plan is to build a dashboard that predicts the trajectory of individual researchers early in their careers.  Such a dashboard could become an important part of some of the major decisions in the life of an academic institution:  who to tenure, who can be expected to attain academic stardom, who to hire and correspondingly make available significant resources, and many other decisions.  Today such decisions are made using a combination of subjective and objective evidence factors.  However, it is our belief that the predictive power of such evidence has not been validated scientifically.

    Trademarks and the Internet

    Professor Catherine Tucker

    Sponsor:  Thomson Reuters

    The idea of this project is to explore how trademarking activity has changed as a result of the internet.  How has the evolution of the internet changed the companies' demand for trademarks and trademarking activities?  The evolution of search engines and the importance of internet search has increased companies' need to identify unique trademarks and protect them.  As stated in a recent New York Times article (Strom, 2012), “brand experts and trademark lawyers say the value of simple, easily understood brand names has escalated in the Internet era because consumers are more likely to find such products while doing searches on the Web.”  The research will use statistical methods to:

    • Relate how the spread of usage of the internet has affected the number of trademarks are registered, and the number of firms are searching for information about prior trademark registration.
    • Understand how this change in trademark search and registration patterns differs between national and local firms.  Has the advent of the internet made trademarking relatively more important for national firms or local firms?
    • Understand how changes in trademarking activity have affected patterns of economic activity and growth at the county-level.