Share on Facebook Tweet on Twitter Share on LinkedIn Share by email
The Embodied Social Proxy Backstory

-Gina Venolia

Remote Teammates Embodied as Hallway Artifacts

In 1997 I was part of a product development team at Silicon Graphics Incorporated (SGI). Most of the team was collocated within a block of offices and cubes in Mountain View, California, USA, but two of the senior members were based out of their homes. Jim lived in Columbus, Ohio, USA; Helga lived in Reykjavik, Iceland. We had the usual mechanisms for staying in touch - phone, email, weekly status meetings with a speakerphone for the remotes, as-needed meetings, check-in and bug mailers, etc. Despite Jim's and Helga's central roles in the project, our manager, Jackie, felt that they were, to some degree, out-of-sight and out-of-mind. She put two clocks on the wall in a common hallway and, significantly, labeled them not "Columbus" and "Reykjavik" but "Jim" and "Helga." Over the course of the project other things got pinned to the wall with the clocks - postcards from the locations, photos of our distant teammates and their families, and postcards of their vacations.

The clocks and the constellation of memorabilia around them became a constant reminder of our remote colleagues. Beyond just telling the Mountain View teammates the time at the remote locations, the physical manifestations of Jim and Helga as collections of artifacts in the team space were a constant reminder of the social presence of Jim and Helga. It's hard to say whether they caused a measurable change in behavior, such as increasing the amount of ad-hoc communication with our remote teammates, but it certainly felt like they did. Jim and Helga were now in sight and in mind.

Erickson et al. developed the concept of social proxies, which they defined as "a minimalist graphical representation of users which depicts their presence and their activities vis à vis the conversation" [Erickson 1999]. There is an analogy between social proxies and the function of the physical embodiment of Jim and Helga in our team space in Mountain View.

In 2006 I performed a series of interviews with Microsoft software engineers based in Beijing, China, and their teammates in Redmond, Washington, USA. I realized that many of the engineers had no collocated teammates - they were satellites to a collocated hub elsewhere. These satellite workers and their hub teammates faced the well-known factors of language, cultural and time zone differences. In addition they faced the same kind of out-of-sight, out-of-mind challenges that Jim and Helga faced. (Indeed their challenges were greater because many of them were relative newcomers to the team, and lower on the career ladder, as opposed to my senior and well-established colleagues at SGI.) I wondered if the same kind of embodiment with artifacts in the hub team space might make them more "socially real" to their distant teammates.

Embodied Social Proxies

Technology had advanced and commoditized such that it was reasonable to think of replacing the collection of artifacts with a computer display. The display might convey the same things - identity, local time, and mementos - to convey a social presence but also go far beyond it by providing an awareness display and acting as a communication hotline the satellite teammate. I call this device representing a remote teammate an Embodied Social Proxy, or ESP for short, since it takes the on-screen representation of a person that Erickson et al. identified and gives it a physical manifestation.

The concept of an Embodied Social Proxy is appropriate for a team that is predominantly collocated but has one teammate that's remote (or perhaps a few, with a proxy for each). It would be appropriate for a team that needs frequent communication to coordinate and negotiate highly-coupled and dynamic work.

As I envisioned it, the ESP would have two modes of operation: awareness and communication. Normally the device would reside in the team hub space with the display showing information about the satellite teammate's identity, availability, location, and work activities. A tap of a button would change to communication mode where the device would become a videoconferencing "head" devoted to the satellite teammate, which could either be used in situ or taken to a conference room or back to a teammate's desk. The satellite teammate would have complementary views of his hub teammates, though it would typically be impractical to give each of them a physical manifestation - it's more likely that existing screen resources would be co-opted as simple awareness displays, with a progressive UI to reveal detailed awareness information and enter into communication. In October of 2007 I created a simple mockup of what an ESP might look like and how it might work.

Almost a year later I learned that George Robertson, a long-time MSR-Redmond colleague, was planning to move to Northeast Harbor, Maine, USA and to continue his many collaborations from his home office there. He worried about his ability to continue to be an effective collaborator. His specific concerns were about his presence in meetings and the general out-of-sight, out-of-mind problem. I showed him my ESP mockup, and we decided to work together to make an ESP representing him in the MSR building in Redmond. Subsequently John Tang was hired into George's group, working mostly from a Microsoft facility in Mountain View, California, USA. We built a second ESP for him.

Research Questions

The two modes of ESP suggest two research questions:

  • RQ1. Can we create an almost-as-good-as-face-to-face experience for meetings?
  • RQ2. Can awareness displays increase ad-hoc communication with a satellite teammate?

The original spirit of ESP as a collection of hallway artifacts inspires a third:

  • RQ3. Does the physical presence of the satellite teammate’s cart increase his social presence in the hub team?

We hope to publish a paper soon which will describe our deployment and answers these research questions.


[Erickson 1999] Erickson, T., Smith, D. N., Kellogg, W. A., Laff, M., Richards, J. T., and Bradner, E. 1999. Socially translucent systems: Social proxies, persistent conversation, and the design of “babble”. In Proc. CHI 1999. ACM, New York, NY, 72-79.