I enjoyed talking to our media folks @ Gies about my mobile marketing research, what got me thinking about a PhD 7 years ago, and the next steps in my academic life. 

Interested? Read more on Gies News.

The recent growth in structured and unstructured data combined with sophisticated techniques for leveraging these data for decision-making, creates huge opportunities for marketers. All business students can benefit from learning about marketing analytics and applying analytics techniques to real-world marketing problems.

I discuss more in my "coffee chat" with Joseph Yun, Research Assistant Professor of Accountancy and Director of the Data Science Research Service, UIUC: 

https://vimeo.com/452730158

Educators all over the world are currently facing a common challenge: How do we teach in the new COVID-19 environment?

One thing we have always had on our side as educators has been time. We have time to research the topics, to research pedagogical approaches, to design courses we will teach in future way in advance. We have time to think, to deliberate, to arrive at optimal solutions. However, with COVID-19, we have suddenly been thrust outside the old world we knew.

Those who taught during the Spring semester know and appreciate how mid-way through the semester, instruction was moved completely online. You wanted more time to plan and design the online experience for your students? Well, too bad. Those of us (including me) who are teaching in Fall 2020 have more time but also a lot more uncertainty. Will classes be held purely online? Will there be an in-person element? What would a blended/hybrid course look like? By some estimates, 65% of colleges plan to re-open for the fall. However, "re-opening" will look drastically different in the new world. Colleges are already discussing how to enforce social distancing norms and masks in classrooms, with fair concerns surrounding their success in practice.

Given the uncertainty for Fall 2020, what can you do as an instructor to be prepared to the extent possible? I have been in online education for nearly a decade now. I have worn many hats during this time: A high school teacher, course designer and marketer for online exec ed courses, a doctoral student, and now an assistant professor of marketing. While these many roles silently prepared me for the challenges we face today, there are no definitive answers right now. So, I've been asking peers and mentors. I have been enrolling myself in online courses. I've been talking to elearning teams, videographers, and experts at Coursera. I've been participating in an intensive 3-week program offered by the U of I elearning team.

What am I learning about teaching in the online space?

1) Learn and own the tech:

Campus elearning teams are swamped! In the ideal world, I would walk into a studio and deliver my content. The videographer will film it. Someone on the team will work on post-production and someone else will put it all together. Voilà! My online course is up and running. The reality is these teams are swamped right now. Consult them, ask them specific questions, show them stuff and get an opinion. Other than that, learn and own the tech.

Here are some more tips from Coursera on how to record home videos. I invested in inexpensive tripod and lapel mics. It also helps to record shorter videos segments in fewer cuts so post-production is easy and at the same time, the content is nicely chunked and organized for students.

2) Survey other courses:

I have found it super helpful to talk to instructors who have been successful at teaching online (even in the pre COVID era) and surveying their courses. These courses can be MOOCs or online LMS for degree courses. These courses can be within your discipline or beyond. When you survey them, think like a designer. Reverse engineer it. Look for elements that you would never have thought of but that can be central to learning in a highly distracting online environment. For example, my colleague Aric Rindfleisch teaches a popular Coursera MOOC on digital marketing that offers a practice quiz at the end of each module in addition to a graded quiz. These practice quizzes can serve as a diagnostic tool for the instructor as well as a self-assessment tool for learners. Similarly, economics professor Jonathan Meer at Texas A&M University and his co-instructor in their online course have a full-fledged course orientation module #0 setup as a pre-requisite. Course LMS should be super intuitive, right? But as he points out in some of their videos, students can often miss a critical link to the module videos, central to their learning and not notice it as half the semester goes by. Jonathan summarizes his key lessons from online teaching in this video I HIGHLY RECOMMEND on how to set up an online course.

Organize, prepare, plan your scripts, include a co-instructor when possible, use other forms of digital media including podcasts or videos to break the monotony. Host at least a few live zoom sessions with your students so they can see you are a real human being. 

3) Include and make LIVE interactions fun:

There is value in incorporating a few live sessions over and above the asynchronous videos. Whether these are online or in-person, plan these well and incentivize students to show up and participate.

One of the best ideas I have received on garnering engagement during these online live sessions came from Elizabeth Luckman, Clinical Assistant Professor of Business Administration at Gies. Elizabeth opens a shared google doc that groups of students simultaneously work on in breakout rooms in zoom. In her own words:

"I did a “reflection exercise” at the beginning of the live session. I had four reflection questions I wanted students to address, and I created a google spreadsheet with the same number of tabs as groups. Then I put them in breakout rooms and in those rooms, they had to type answers into the doc. So that means that even though they were in breakout rooms, I could see (and they all could see) what each of the other groups was doing. I had never done it before but it worked really well!  The crowdsourcing doc that I showed is live all the time, and they can add to it at any time. My goal is that they can use that information for their final projects - and also just have this cool resource that they can download and save after the class ends. Keeping discussion going during live sessions is rough. My experience is that smaller numbers of students in live sessions, the google doc during breakout groups, and (sadly) incentivizing discussion is sort of the right mix."

To better teach online, we must first learn to learn ourselves. Overall, don't fret!

Accept that teaching in the new world will look different. Accept that it will require unlearning and relearning, and higher investment of time and energy. Be patient with yourself but open to learning new tools and techniques. Try to place yourself in the shoes of your students and wondrous new insights will emerge!

Five years. 2015-2020.

When I started my PhD five years ago, like any new incoming student experiencing the academic world for the first time, I had naive expectations about research and doctoral education. In this article, I share my own experience and challenge five myths new students often have about getting a PhD:



​1) It is all about grand ideas. 

Grand ideas are a good start. However, execution of a few good ideas is what gets you through. When I started my PhD in Marketing, I had the grand vision of researching how the mobile-first world impacts consumers and firms in different domains, including retail (how we shop), education (how we learn), and healthcare (how we access health services). Many first year PhD students come in with a grand vision and think they must pursue it all. When defining the scope of my doctoral research, it helped to keep three things in mind: 

1. Your "ideal" data may not be available. Often for data-driven marketing research, quality of data is critical. Ideal data with all possible variables (P) and large enough number of observation (N) are often hard to find. Sometimes, the data you do find may not be usable for each of your grand ideas, e.g., for mobile-based education, I approached an ed-tech company providing virtual mobile coaching. Their data were limited to their clients from a handful of companies with no granular information on users' actions in the app. Another company I cold-called gave me a data dump of 8GB with no labels or explanation and 70% missing data for several columns. Not the best data.
2. All your questions may not be answerable. Your data may not allow you to answer all possible questions you started with. In my case, the data I could access for shopping had information on loyalty program members of a retailer only. So, I could not necessarily compare how mobile marketing impacts loyalty vs. non-loyalty members differently. I could certainly examine how app users differ from non users within the loyalty program members (and this resulted in a successful co-authored peer reviewed publication).
3. Your initial scope may be over-ambitious. In addition to the feasibility of specific projects, an overall explosive portfolio of projects with too many ideas can distract away from the core focus and reduce the quality of research. In my case, I actively decided to shelve my healthcare and public-policy related research interests for future and focus on the retailing piece first. 

2) The dissertation is the be-all-end-all of my PhD research and must have lasting impact on the field.

There are differing perspectives and philosophies on what a dissertation is and should contain. I personally found it helpful to conceptualize my dissertation as a set of related papers that together fairly represent my broad interest areas, demonstrate my empirical skills, and introduce sufficiently new ideas or insights to an existing body of work. Often, PhD students get caught up in the "impact" trap -- where we want to create impactful research right away. Consequently, no manuscript or draft seems "good enough." The danger is a resulting research paralysis, a kind of a "writer's block" that prevents progress. Along my PhD journey, lots of people gave me good practical advice that helped get my first drafts done. Three senior academics at different times told me:

1. Often, it's not a single paper that creates impact but a collective body of work or set of papers. 
2. Impact is too heavy a burden to carry at your level.
3. Work on interesting problems and focus on publishing -- your work will get noticed. 

​
3) Unless I go to a top school or find a well-known advisor, it doesn't count.

Incoming new students often place tremendous weight on top schools and popular advisors. While these factors can help, ultimately your peers, colleagues and recruiters are all looking for signals about YOU. Who are you independent of your advisor or school? What do you care about? Have you demonstrated research acumen, perseverance and initiative? Over the years, as I have seen PhD students who entered a PhD program only to find their advisor leave their job at the university or department management and budgets severely changed, this perspective may be a good reminder to take ownership of their projects despite all odds as they work to improve their circumstances. 

​4) Unless my advisor contributes, I cannot make progress.

Let's face it. Advisors are busy. They are constantly juggling teaching, traveling, multiple projects, and sometimes, even consulting. Waiting for your advisor to do the work for you is dangerous even if they have the best of intentions. Do your part. Do as much as you can. Go out there and find the help you need -- online, offline, at the library. Most of the problems you are waiting on your advisor to solve for you may have been solved before. Take your best solution to your advisor instead of just the problem statement. You will be surprised how much faster things are able to move. 

5) More number of manuscript will help me succeed.

Advanced and finished projects with a fair pipeline are much better than a bunch of loosely connected works-in-progress or working papers. Prioritize prioritize prioritize. Yes, we all need some "fun" projects and starting new projects is always fun. As they age and get further along in the review process, projects tend to become drudgery for anyone. That's the time to push through. Instead of constantly starting new projects, finish the ones you start. Every research has weaknesses and the review process is geared to bring them up. Preempt them. Acknowledge them. Fix them to the extent possible. Turn those manuscripts back in. If you are going to get it rejected, might as well find out sooner than later.

Finally, along the way, have fun! Make friends. Explore your campus. Reach out to people. Talk to people at conferences. Talk to people outside your field. Ask lots of questions. Help others. The journey is so much more memorable that way. Best!

The PDMA (Product Development and Management Association) Doctoral Consortium provides doctoral students an opportunity to connect with a community of innovation researchers and scholars. The 2019 Consortium was hosted by the Gies College of Business in August. This year, 24 Ph.D. students and 22 faculty attended. The event sparked several interesting conversations on research in innovation as well as perspectives for young scholars to manage their academic career and research program well.  

Source: https://www.pdma.org/news/466705/PDMA-Doctoral-Consortium

In several product categories, such as electronics, video games, computer hardware and software, and other hi-tech products, backward compatibility feature--the property of a current generation of hardware to allow previous generation of software or accessory to work with it--is an important strategic decision for firms introducing hardware upgrades. We empirically investigate the effect of Microsoft Xbox’s decision to make its new generation console (NGC, Xbox One) backward compatible with selected games for its previous generation console (PGC, Xbox 360) on sales of video games for both PGC and NGC. We assemble a unique dataset using data from a large proprietary game retailer and data scraped from gaming sites during 2013-2017. We analyze the effects using a difference-in-differences approach that includes the use of a synthetic control group. Our results show that when a video game console firm makes its NGC compatible with PGC games, unit sales of PGC games do not change. However, dollar sales of PGC games increase due to a price increase effect. Importantly, sales (units and dollars) of NGC games increase due to a spillover effect, driven primarily by console upgrades. The increases in dollar sales of PGC games result from greater sales of classics, sequels, and first-person shooter genres than sales of non-sequel and other genres of games. Our research offers managerial insights on the backward compatibility feature decision in product upgrades.

(Complete working paper with @Venky Shankar available on request)

1. Mobile: 

• Focus: How does mobile impact shopping in online and offline channels?
• Extensions: (1) Mobile "last mile" -- where are the returns on mobile coming from? Advanced user-profiling based on location and movement data, for improved targeting in other channels and analytics on visitation, purchases, competition (2) Beyond shopping: Mobile's impact for health, and education
• Student researcher-in-focus: Boram Lim, UT Dallas

2. AI and robotics

• ​Focus: How do consumers perceive human-likeness in robots? Investigating the "uncanny valley" in its upward trajectory
• Extensions: (1) Contexts and settings when the upward trajectory might not exist at all, e.g., robots as co-workers (2) Beyond physical aspects of human-likeness, how would consumers respond based on 2-way interactions with robots? (3) Affective computing and implications for consumption and choice
• Student researcher-in-focus: Noah Castelo, CBS

3. Search and Segmentation

• Focus: How can offline search and learning improve our extant models of simultaneous search and choice? Segmentation is still cool, with new cutting-edge machine learning approaches to segment consumers based on purchase behavior
• Extensions: Online and offline search modeled together, with a view to profiling shoppers / segmentation based on search might be a promising domain
• Student researchers-in-focus: Dan Yavorsky, UCLA and Milad Darani, Mays Business School

4. Pro-social and gift-giving behaviors 

• Focus: Consumer preference between charity and cash registries, consumer valuation of products based on object history value
• Extensions: The importance of trust and privacy in these settings might be an interesting avenues to study in future, e.g., in pro-social behavior, leveraging online efficiencies to ensure the last dollar donated is benefitting the persons intended. 
• Student researchers-in-focus: Michelle Daniels, ASU and Charis Li, UF

This code uses the MatchIt package for propensity score matching to demonstrate with and without replacement Nearest Neighbor matching. The additional extension it offers is to create panel data using matched sample in both instances (particularly non-trivial for with replacement matches).

There isn't an easy way to export messy R output for regressions into a usable table. You're probably okay manually doing this for 1-2 regressions, but what if you have to estimate hundred of models for different Ys (dependent variables). This code allows export of regression results produced from regressing different Ys ('outcome list') on a set of Xs into a usable CSV file. You can flexibly incorporate clustering, fixed effects, etc. as needed in the code. Happy regressing!