Diminishing Returns of Collaboration

Artist Alessandro Villa http://www.flickr.com/photos/unsuono/
Artist Alessandro Villa http://www.flickr.com/photos/unsuono/

Having an economic foundation Haque approach seems for me rational. And having gained some life experience I have to admit that – and this is not age related – one can not connect to everyone as a friend! And what about the networked computers and so on. Something sometimes never change


Naumi Haque June 15th, 2009

While generally a believer in how collaboration can lead to better insights and greater efficiency, I continually see examples of where it is neither effective, nor terribly efficient – and in the worst cases totally counter-productive. I work in a highly collaborative environment and study many others, and my experiences have led me to two areas where problems typically emerge:

  1. At an individual level people suffer from cognitive overload. As people get busy and collaborate across a multitude of projects, the brain gets distracted, and the quality of the output suffers. In short, one person can only do so much.
  2. At a project level where you run into a situation of ‘too many cooks spoiling the broth.’ In short, only so many people can do one thing.

If you put the two of these together, the worst-case scenario is that in an individual could join a project as the Nth person who ‘spoils the broth,’ while the time they dedicate towards doing so distracts them from their other work – which, continuing the cooking metaphor, leads them to burn the toast as well.

The problem is, it’s very difficult to apply a scientific approach to measure exactly how many people per project, and conversely how many projects per person is optimal. The most well-known study around this is Dunbar’s Number, which sets “a theoretical cognitive limit to the number of people with whom one can maintain stable social relationships” at 150. In terms of collaborative overhead, Dunbar speculates that “as much as 42% of the group’s time would have to be devoted to social grooming.” Now that might be acceptable for the hunter-gatherer societies described in Dunbar’s anthropological study, but I would imagine this amount of “grooming” time would be extremely unproductive in an enterprise context.

This post is recommended for you  How CIOs can use machine learning to transforming customer service

In his book Collaboration, released this month, Morten Hansen, a professor at the University of California, Berkeley and INSEAD, identifies two costs related to enterprise collaboration. The first is the opportunity cost collaborating (i.e. the opportunities individuals could have been pursuing had they not been collaborating), the second is the cost associated with fostering co-operation. In both cases, as the number of projects or the number of individuals grow, so too does the potential for diminishing returns.

At the project level, I feel as though most people have general understanding that there is a certain point at which there are simply too many stakeholders and collaboration breaks down.


However, at an individual level, I think we are less cognizant of – or less willing to admit – our own limitations. I’ve seen many cases where an enthusiastic and eager collaborator was clearly overburdened and well past the point of optimal effectiveness. Incidentally, my personal hypothesis is that this point of optimal effectiveness is a fairly small number of projects per person. My main “proof” for this is anecdotal, but I notice that the busier one is, the more likely they are to quickly skim a topic and provide feedback in short (sometimes valuable) chip-shots without contributing to a better in-depth understanding of the topic space. Worse, in some instances perceived value comes from dissenting, so instead of constructive feedback, you get wildly varying opinions with no one working towards a coherent solution.


On the subject of cognitive overload, a recent Deloitte report notes, “Even a Sunday newspaper contains more information than the average 17th century citizen encountered in a lifetime. Add to that the stress of decision-making amidst uncertainty, corporate change, and a tidal wave of tasks. Never before in history have workers been asked to absorb and make sense of so many data points.” One more sensational study even suggests that information overload is more damaging to the brain than smoking pot. I think we can certainly make an argument that where collaboration is most likely to break down is at the individual level.

This post is recommended for you  The secret to digital success? Design thinking

This brings up another point: What about the virtues of solitude? Are we losing our capacity for individual decision-making? Moreover, who’s actually doing the deep thinking needed to solve complex problems? We talk about the multitasking Net Gen brain that is not actually doing multiple things at once, but rather switching more efficiently. Does constant switching allow for deep analytic thought?

So what is the solution? Overall, I’m wondering if there’s a Dunbar Number for the optimal number of simultaneous projects per person (small and large). How is this number affected when you take into account broader ecosystem participation and places where quick feedback from multiple participants is actually desired over in-depth participation?

As a start, I think collaborative technologies can help by streamlining different types of feedback. So, for example, a project can have 1,000 collaborators if they are providing feedback via a prediction market. Conversely, if only three people are collaborating on a document, perhaps a wiki is most effective.

One possible model for managing cognitive overload is letting individuals self govern – i.e. everyone decide where they can add the most value. Of course, this also raises many issues, including: people, especially in high-performance cultures, tend to overextend themselves; people tend to pick project that interest them, but that may not add the most value to the organization; and people tend to be social and so will gravitate towards the same projects, thus contributing to project inefficiency.

In order for this to work, you would have to architect a system that would allow people to allocate their own time in a structured way (similar to the Freiburg budget example). I’m envisioning a system where resources are finite but can dynamically allocated; where employees are guided by decisioning logic that identifies the projects that provide the most value to the organization; and where limits are set that prevent projects from being staffed by too many people and that stop people from taking on too many projects.

This post is recommended for you  Chatbots: Should you waste your time? Top Minds Chime in


Reblog this post [with Zemanta]