Communications of the ACM

Economic and business dimensions

Geography Is Alive and Well in Virtual Teams

By Jonathon N. Cummings
Communications of the ACM, August 2011, Vol. 54 No. 8, Pages 24-26
10.1145/1978542.1978551
Comments

Though members of a virtual team can work anywhere, virtual does not end the role of geography. Geography is alive and well. Geographic distance and lack of overlapping work hours impose large coordination burdens on a team. Research has begun to suggest what circumstances generate the largest burdens and why.

For example, members physically separated by equally great distances experience fewer coordination challenges when they are within the same or a few time zones than different time zones. Thus, not surprisingly, members from New York find it easier to work with members from Buenos Aires than Tokyo. However, for team members within the same or a few time zones, physical separation still reduces awareness of what other members are doing. Thus, in order to benefit from working virtually, members need to allocate enough time to the task to overcome these additional coordination obstacles.

Back to Top

Geographical Space and Time

Geographic dispersion in teams is a multidimensional construct. Michael O'Leary and I noted that while people often think of teams as being dispersed or not, dispersion is much better characterized by the degree to which members are in different geographic locations.³ Spatial dispersion should be treated as a separate dimension than temporal dispersion. Together, geographical space and time yield important insights about working at a distance, including the benefits (for example, accessing member expertise) and costs (for example, constrained member communication); see the accompanying figure.

In the figure, Quadrant A reflects teams with high spatial dispersion but low temporal dispersion. For example, members could be split entirely between North and South America, or between Europe and Africa. Quadrant B shows teams with high spatial and temporal distances. These far-flung teams could have members distributed across North America, South America, Europe, and Africa. Quadrant C reflects teams with members who are co-located and work at the same time while Quadrant D shows teams with members in the same location but split entirely between two work shifts (for example, 5 A.M.–2 P.M. shift vs. 2 P.M.–11 P.M. shift).

Time zones and coordination delay. How should this more nuanced view of geography in teams affect thinking about where members work spatially and temporally? For teams in Quadrant A (distributed north-south) versus Quadrant B (distributed east-west), consider the consequence of coordination delay, which occurs when it takes longer than expected for one member to receive progress reports, communication clarification, or issue resolution from another member. Over time, coordination delay among members can lead to worse team performance in terms of completing the work on schedule, completing work within budget, and meeting final requirements. In a study of 108 teams in a multinational corporation (projects included software development, hardware development, and systems integration), this is what Alberto Espinosa, Cynthia Pickering, and I found.¹

Using company database information on the locations of 675 team members (in 54 cities across 22 countries), we determined the extent to which each pair of team members was separated by spatial boundaries (same city vs. different city/same country vs. different country). To measure temporal boundaries (non-overlapping work hours), we used time zone data associated with each city along with work hours in a 24-hour time period (and accounted for members who shifted their hours to overlap with other members). Some pairs of members had at least one hour of overlap during a nine-hour workday and some pairs of members had no hours of overlap during a nine-hour workday.

Based on surveys, we also assessed the extent to which each member used synchronous communication tools and asynchronous communication tools with every other member of their team. Lastly, we asked about coordination delay with other team members.

The results of our study showed that, as expected, working across spatial and temporal boundaries were both significantly related to coordination.¹ That is, coordination delay was greater for pairs of members in different countries compared to pairs of members in different cities (in the same country) compared with pairs of members in the same city. And coordination delay was greater for pairs of members who had no overlap in their workday compared to pairs of members who had at least one hour of overlap.

We also found that while synchronous tools (such as Web conferencing) were effective in reducing coordination delay when pairs of members were separated by space but not by time, asynchronous tools (such as email) were also effective when pairs of members had high spatial boundaries but low temporal boundaries. Thus, temporal boundaries were more difficult to cross than spatial boundaries given that none of these communication tools reduced coordination delay for pairs of members in different countries who had no overlap in their workday.

Further interviews led us to conclude that teams with better alignment of geography to the work being done (that is, maximizing access to member expertise while minimizing communication constraints) were more effective compared with teams that tried to work through high spatial and temporal boundaries. One virtual team member even reported reorganizing teamwork to take in account being separated by many time zones: "One of the problems that we have when working with people that are eight hours away in time zones, is the coordination of large projects...the engineers need to work together to talk through problems. So, when there are significant time differences, they just can't make good solid progress without being able to talk. So, what we started to do in the last couple of years is starting to divide up the project to where each site has a completely separate job, with some minor overlap with the other sites. But in general, they have one complete section of the project all to their own. Otherwise, what we found in trying to manage one functionality across multiple sites is that people were having to work a ridiculous number of hours in order to maintain that communication." (interview cited in Cummings et al.¹)

Physical separation and time allocation. Now consider teams in Quadrant A (distributed north-south) versus Quadrant C (co-located), and how members allocate their time to the team. It is common today for members of teams to work on several teams concurrently. As a result, virtual teams are often composed of members who vary widely in the amount of time they work on the team.

To investigate how geographic dispersion impacted the relationship between time allocation and team performance, Martine Haas and I studied 285 teams in a large global company (projects focused on product innovation, operational improvement, and customer service). The 2,055 members were located in over 50 countries. We measured the extent to which pairs of members were on the same hallway, different hallway but same floor, different floor but same building, different building but same city, and different city but same country, or different country. We then aggregated this fine-grained location data to the team level to capture geographic dispersion among members.

---

There was an advantage for more dispersed teams when members allocated more time to the team.

---

We also measured the percentage of time that each member contributed to the team, and used independent ratings by company executives to measure performance (for example, value delivery, goal alignment, and tangible results). As expected, we found that teams with members who allocated a greater percentage of their time to the team performed better, on average, than teams with members who allocated less of their time to the team.²

More interesting for the discussion of geography, we showed that there was an advantage for more dispersed teams when members allocated more time to the team. That is, teams with members spread across different buildings, cities, and countries benefited more from allocating more time to the team relative to teams with members spread across different hallways and floors in the same building.

Greater physical separation (for example, different floor vs. different building vs. different city) took a greater toll on member attention in terms of keeping track of what others were doing in different locations and managing the communication required to stay on task. This raises an important question: If virtual teams are more likely to have members who allocate less time, yet virtual teams benefit the most when members allocate more time, how should managers resolve this paradox?

We suggest that when designing teams, managers should explicitly take into account where members work when determining how much time members should allocate (to maximize attention focus when dispersion is greater). For distributed work to be successful, the task should not be added onto what members are already doing, since just because the work is distributed does not mean it is going to be better.

Back to Top

Conclusion

The realities highlighted here regarding the persistence of geography in virtual teams are not intended to offset the enthusiasm around advanced collaboration technologies (for example, Cisco's Quad, IBM's Lotus Connections, Salesforce.com's Chatter, Microsoft's Lync). Rather, they are intended to clarify where certain trade-offs still exist when working across space and time.

Moving forward, technologies with the best odds of success in assisting virtual teams will need to increase member communication as well as help manage and coordinate their work through better partitioning tasks by location, managing dependencies among tasks that bridge locations, and synchronizing how tasks are integrated across locations. Otherwise, there is a greater chance of an obituary being written for particular technologies than for geography.

Back to Top

Back to Top

Back to Top

Back to top

The Digital Library is published by the Association for Computing Machinery. Copyright © 2011 ACM, Inc.

---

No entries found