Collocated Teams: 7 Problems With Face-to-Face Communication
XP emphasizes radically collocated teams sitting together in a shared workspace. Some of the XP thought leaders claim that anything less than that will cause significant reductions in productivity. In some situations that may be true but it hasn’t always been my experience when I’ve worked on distributed teams.
The productivity claim seems to be based on several issues: the speed of communication, the effectiveness of communication, and team bonding. The counterpoints for these issues are that physical distance doesn’t necessarily slow down communication or make it significantly less likely to happen when people want it to happen (see the speed and quantity of communication through instant messaging and various online forums, for example). I agree that team bonding can be an issue but I haven’t seen it be a big problem if the team has already bonded prior to being distributed. The communication effectiveness issue is the topic I’ll expand upon. Again, some XP thought leaders claim that face-to-face communication is always preferable to other forms of communication. Of course, a “truth” is seldom always true. As a thought experiment, let’s look at some of the potential problems with face-to-face communication.
- Misinterpreted body language. Some people compare face-to-face communication to a broadband network connection while other forms of communication are something like 300 baud (extremely slow) phone modem. The idea is that you receive more information per unit time when communicating face-to-face. That’s true. The question is whether than information is helping the effectiveness of the communication or not. The scowl on the other persons face might be something they had at lunch. You may not realize that her shoes remind you of your first grade teacher who ridiculed you in front of class. You have all this extra “bandwidth”, but how much of it is noise that hurts effective communication rather than helps it.
- Less time to think about the message. Face-to-face communication also means real time communication. The human brain is particularly fast about processing and generating language. Some people are relatively good at it and some aren’t. Face-to-face communicate will give an unfair advantage to the person who is fast. The slower person may be forced to say things they don’t really mean or haven’t considered deeply just to stay in the conversation. With other forms of communication, like email for example, you can spend more time to think carefully about what you want to say and how to say it. I’m not saying people always use that time, but it’s there if you need it.
- Less formal communication. I recently moved to a different country where I don’t speak the language. Fortunately, the people at work can speak my language so it works out well. However, being here I notice that I use many very informal ways of communicating that can easily be misinterpreted (idioms, slang, and so on). When I think about it, this is common in most spoken communication. There is nothing about forms of communication like email that make it more formal, but it easier to be precise than when speaking face-to-face in real time.
- Interruptions. Between mobile phones and the emergency-of-the-hour there’s often something that will interrupt the face-to-face conversation before we complete. Even if we resume later, chances are we will not fully remember what we discussed because we… less likely to have a record of the conversation.
- Less likely to have a record of the discussion. Most people don’t record or transcribe their spoken communication. That’s fine, because most of that conversation not effectively communicating information because of the reasons we’re discussing. With many forms of electronic communication, there is automatically a tangible record of the discussion. This could be good or bad depending on whether you invested some time to think about your message first.
- Synchronous. To have a face-to-face conversation we must both be there. I know, it’s obvious. But, think about it. If we try to find a time when we are both available isn’t it possible that it will delay the communication unnecessarily? With asynchronous communication I can send my message now and you can read it whenever you have time. It’s very possible that you will have time available sooner than we can both be physically available to have the synchronous face-to-face communication. Worse yet, the communication may never happen if one person isn’t available immediately and the other isn’t persistent about finding a common time and place for a face-to-face discussion.
- Expensive. By expense, I don’t necessarily mean money. Physically moving people so they can have face-to-face discussions can be extremely difficult and expensive, if not practically impossible. Imagine what it would cost and how much effort it would require to physically collocate the 7000+ members of the XP forum at Yahoo so they could communication “more effectively” face-to-face. Even if it were possible, I’m skeptical that the communication would be more generally effective.
I like face-to-face conversation and I’m a strong believer in radically collocated teams. However, I also recognize there are tradeoffs and sometimes distributed teams augmented with other forms of communication can be productive and cost effective.
Steve Wampler wrote:
Hey Steve,
I agree (mostly) with your conclusions. In the risk project we’re using Skype and VNC to do pair programming. between the Austin and Dallas sites. It’s quite effective. In fact, it may be better than actually setting next to each other because it’s easier to see and to switch who’s programming vs. who’s watching. Distance pair programming may actually be better because it encourages both programmers to stay focused (at least I think it does). At the beginning of each iteration though, I’ve found we have to get together in the same room. Those meeting have more to do with keeping the relationships fresh than actually getting anything done, although we do that too. I guess I would also say, that from a relationship point of view, that I would always prefer to be in the same area/office … as long as the members of the team are compatable. I have one guy on my team that turning out to be somewhat of a problem. Luckly he’s a contractor, so I’ll likely let him go when the project is over. That alone will improve the team’s dynamics. By the way, I now have a bunch of Skype addresses based on which machine I’m currently using. Just do a search on swampler and you should find all three of them. Give me a ring sometime when you’re free.
Posted 06 Feb 2006 at 6:03 pm ¶
Steve Bate wrote:
It sounds like the project is going well. We don’t do frequent pair programming at my current job but sometimes I work with the developers in Paris to resolve some issue. We also use Skype, VNC, and IM. However, I’ve never liked the speed of VNC very much, even over a fast LAN. I wish it were as fast as Windows’ Remote Desktop Connection (which unfortunately doesn’t support shared desktops, as far as I know).
Posted 09 Feb 2006 at 8:01 am ¶
Dave Nicolette wrote:
Your comments make sense generally, but in the specific context of XP you miss the mark a bit. The thing about XP is that the various practices it calls for complement one another. Quite often, when a team omits two or three of the XP practices the whole thing starts to unravel. The problems you cite with face-to-face communication could indeed be problems in general, but if you are applying XP correctly most of them are non-issues.
For example, misinterpreted body language is always possible, but less likely than misinterpretations caused by the absence of body language. Besides, a misunderstanding is easily and quickly resolved between two people who are physically together.
Less time to think about the message. What prevents a person from saying, “That’s an interesting point. Let me have some time to think it over.”
Synchronous / interruptions. With XP the team is working in pairs and sitting in the same general area. Individual team members are dedicated to the project full time. They are always available, except perhaps when visiting the toilet. The common time for a discussion is, by definition of XP, “now” and the common place is, by definition of XP, “here”. XP is the cure for, not the cause of interrupted discussions.
Expensive / 7000+ users. No one in his right mind would attempt to run an agile project with 7000+ team members. It would take a year to do a daily stand-up. Agile methods in general (including XP) are not scaled by adding more and more people to the team. They are scaled by decomposing a large initiative into several smaller projects, each of which is then staffed with a small team.
Foreign language / customs, etc. I worked on a distributed project a few years ago that included team members in China, Canada, and the US. It was not unlike the situation Steve Wampler describes in his comment. We worked “together” by using IM, whiteboarding tools, and the like. The Chinese team members’ English skills were quite good for text communication. At the end of the project we all got together in NY for a celebration. The Chinese team members’ spoken English was absolutely unintelligible. So I can sympathize with your comment that face-to-face communication can be problematic when you have a multi-cultural, multi-lingual team. But this is not a problem inherent in XP or any other methodology. When you compose a team in this way, it is a reality you must deal with somehow. If you fail to deal with it, is it really a failing of the methodology?
Posted 30 May 2006 at 3:05 pm ¶
Steve Bate wrote:
Dave,
Thanks for the comments. I agree that some of the issues are less likely with people who are relatively self aware and highly skilled communicators. For example, it is sometimes difficult to realize we are being affected by body language or other extraneous factors like hair style, clothing, smell, accent or whatever. /If/ we know we are affected we can possibly control our reaction to these factors (at least there is the possibility of it). I don’t know of anything in XP that lessens the possibility of these types of subconscious reactions. If anything, like flow interruptions in a busy team room, it might be more likely to happen than less with a collocated team. Even if that’s true (and there’s no way to know for sure) I’m not claiming that the potential problems of collocated teams outweight the benefits, at least in general. I prefer collocated teams, but I also acknowledge the benefits of distributed teams in certain contexts.
Posted 30 May 2006 at 4:23 pm ¶