Listen to the podcast:
Teresa Lunt, who directs the computing science laboratory at the Palo Alto Research Center, is involved in a wide range of activities, including ubiquitous computing, security and privacy, and ethnography for organizational environments and technology design. During a talk with Knowledge@Wharton at the recent Future of Publishing Conference in New York, she discussed a few of her current projects — such as research into workplace efficiencies, a study on mobile advertising and the creation of a rich media information service for a customer in Japan.
An edited version of the transcript follows.
Knowledge@Wharton: Teresa, thank you for joining us today.
Teresa Lunt: Thanks very much for inviting me.
Knowledge@Wharton: Earlier on the panel about the future of content, we were talking about the rule of context. I understand that at PARC you have some really interesting research about the rule of context and content. I was wondering if you could tell our listeners a little bit about that research.
Lunt: Let me tell you about a few projects. We did one for a customer in Japan who was looking for a rich media information service they could offer to their customers. Their customers today are in the print business — magazine publishers and things like that. So we worked with our social scientists to try and run through about 60 different possible services the client could offer and analyzed what those might be. And then we did some ethnography on the streets in Tokyo, interviewing and observing people. Finally, we arrived at a certain demographic and a definition for the service and then we worked with them to elaborate on that.
Our social scientists went out and discovered this market was about young people who needed recommendations on things to do and places to go with their friends. So it was all about context. The young people were going to converge on some part of Tokyo that they were not familiar with and they needed to find places to go out to eat, places to go after dinner, movies to watch, places to shop, and things like that.
So to know what activities would be appropriate, we needed to understand the time of day, the people involved, and the personal tastes of those people. We took that information from a lot of sources to infer what they were going to do next so that we could recommend content to them. They would never have to search. They would just look at the device and get a recommendation for a place to eat that would be local and that they could walk to. Basically, it is about understanding the task at hand, what specific information is going to be valuable and important to them, and then presenting them with only that information.
Knowledge@Wharton: What are some of the other contexts in which you find content might be relevant?
Lunt: We have a big project right now [based on] several years of research, including a lot of technology development and ethnography, to understand work practices and the things that we are missing or that maybe are not happening efficiently in the workplace. We’re calling the project Meshstro. We are having a private beta [test] right now … and we will probably have a live beta later in the year.
The first product [involves] your email. As you are dealing with your email, there is a sidebar that will give you lots of relevant content about what you are doing now, who you are communicating with now, or what group or company you are involved with now. For example, on the sidebar you will see related email. You will see links out to social networking sites. You will see links out to web content. You will see links to other content. Basically, this is to help you quickly get in one place information that you really need to deal with.
We have plans to extend that in the future to be able to index information throughout the enterprise and beyond so that you can index it not just by a deep understanding of what’s in the content, but who has been using the content, what activity it is associated with, and what task it is associated with. So you can get very specific and precise [with things like], “I am writing a proposal to company X on this type of topic,” and it will bring up all kinds of related activities that are going to help you with that task [such as] related content.
Knowledge@Wharton: So based on the research that you have done on both of the projects you described, what kind of general conclusions have you drawn about the relationship between content and context?
Lunt: One observation is that it is extremely difficult to develop a precise understanding of what the user needs at the moment. If you can’t do it well, the user will quickly learn to just ignore all the suggestions that you give them. So it is very critical to do a good job learning those inferences, making the correct inferences, and being very precise about it.
Knowledge@Wharton: Have you looked into the issue of whether the value of content changes depending on the context?
Lunt: Right now we are doing a study on mobile advertising because we see that the advertising industry is very interested in location-based advertising and we are trying to broaden the definition of context beyond location. Yes, we think it will be very valuable to know when somebody is particularly interested in receiving certain advertising context or making a purchase because they are hungry, because they are shopping for a particular item, or because it is somebody’s birthday in the family and you know the preferences of that person. So we believe all that information combined could be very valuable to advertisers.
Knowledge@Wharton: So do you see context primarily in geographical terms or is it different than that?
Lunt: The location is one piece of the context. The fact that it’s your mother’s birthday or that you have been driving for three hours and you have hungry kids in the backseat are also pieces of context. The time of day and the people that you are with can determine part of your context. All of those things are useful in figuring out what is valuable to you to know right now.
Knowledge@Wharton: What are some of the unanswered questions that you would like to research in the future around these themes?
Lunt: Unanswered questions for me have to do with kind of a vision of what the world will look like in the future. We can gather a lot of information about what you are doing, who you are with, what information you are accessing, what you are purchasing, what you are looking at, and what you are thinking about, but how can we really harness all of this? … We can collect all of that information, but what are all of those use cases where we can create value through leveraging all of that context? That for me is a big space to explore.
Knowledge@Wharton: Thank you very much for speaking with us today.
Lunt: Thank you.
For more coverage of The Future of Publishing Conference: