As you know, we have been testing our TiLR robotic telepresence platform since Summer 2008 with various customers large and small. Our first test site, the Google Lunar challenge team over at the XPRIZE Foundation was a huge success. Since then we have had the opportunity to test TiLR in many more settings and we’ve gotten a lot of feedback and data-points on how robotic telepresence increases productivity and team cohesiveness while simultaneously decreasing travel and downtime.

Some of the data was predictable. For example, we had a good idea on how much productivity increase we can expect in certain situations. At the same time we were also surprised by some of the data we’ve been receiving from customers… and I want to share a few of those surprises with you today.

Communication Initiation

When we started back in 2003, we believed the value of robotic telepresence was in the ability to drive (the robot) to someone’s desk and initiating a conversation. This would eliminate the need for travel to that location by using the robot as your avatar in the remote office, particularly useful to people who work remotely and/or have a need to be in multiple offices at the same time. It turns out however, that 67% of all communications on the robot initiate on the robot side!

This was particularly surprising because we had always assumed that you can pick up a phone, use IM, or send off an email to the remote person. Yet it turns out that the mere presence of the remote person via the robotic avatar has profound psychological effects on how people initiate communications within teams. To put it differently, we are far more likely to ask questions or share information with someone who is physically present, and conversely we tend to delay and/or never initiate a conversation with a remote team member, even though that person’s input may be needed or even critical to accomplishing a common goal.

Digging deeper into this, we found that team members viewed formal meetings as the place to ask questions or share information with remote team members at a ratio of 2-to-1, even though no such formal procedures to this effect were in place. By “formal meeting” we mean either an actual meeting where team members get together. This can mean either a physical meeting, a virtual meeting using video and/or teleconferencing, or even a common wiki page that is updated at pre-determined time intervals.

In practical terms, this finding has some profound effects on team performance:

First, it suggests that existing tools such as IM, Email, Telephones are not being utilized as much as they could be. A team member can easily shoot off an email to inform a remote team member about a certain decision that was made at the office… but the inefficiencies with random-interval-communication in such manners is deemed not worth the effort.

Second, there  is strong evidence that physical representations of remote team members can increase both the number of conversations as well as the quality of conversations by orders of magnitude. Of course we see this everyday at the office where people talk freely, exchange information, and ask questions. It is less evident in cases where there are remote team members because generally companies don’t have informal conversation procedures between multi-geographic team members. Yet, we see a lot of anecdotal experimentation being done in this area, such as requiring all workers to be present in a virtual meeting room at all times while working (e.g. Second Life) or having  videoconferencing screens setup where they stream between offices throughout the day.

Third, in analyzing our survey results, we found that the very action(s) required to initiate a conversation, no matter how trivial (such as composing a new email and/or dialing a phone number) are by far the number one cited reason for prohibiting these ad-hoc communications in the first place. As such, removing as much of this barrier would automatically translate to a higher occurrence of informal communications with remote team members. In the case of robotic telepresence, it is very often the case when the robot (e.g. the remote team member) is passing by someone’s cubicle, and that person is then reminded to ask the remote team member a question or to inform him of something.

There were several other surprising pieces of feedback which I will cover in the coming weeks. In the meantime, I’d like to invite anyone who may be a remote team member and/or works in a multi-geographic team to chime in on these findings.

Best,
/F