The last few years has seen a gradual erosion of the lone genius approach to innovation, with a preference for a connected and collaborative approach whereby innovation is a gradual improvement with recombination at its core.

In many ways, this is perfectly logical.  Technology today is incredibly complex, so the chances of lone individuals knowing enough to improve on what we have is slim, especially in terms of the great leaps forward we have perhaps seen in the past.

Diverse groups

This then feeds into the drive for diversity in those groups, at least in terms of thoughts and ideas, for the wider range of perspectives you can bring to a problem, the more likely you are to find an innovative way of tackling it.

Connect up these diverse groups and you have the prospect of networked super-intelligence.  That’s the way things work, at least most of the time that’s how things work.

A recent paper by researchers at Arizona State University reminds us that it doesn’t always work out that way however.

The study saw participants asked to try and find a virtual remedy and stop the spread of a virus.  The game offered players the option of innovating via incremental improvement or recombination of existing cultural traits.

They were placed into one of two sets.  The first saw them fully connected with each of the other five members of the group, with full transparency on the contributions of each member.

The second set, by contrast, were only partially connected into three groups of two members, with the contributions only available to other members of the sub-group.

Too much connectivity

When the output of each set was analyzed, the results revealed that the connected group did well initially, but then rapidly converged towards homogeneous solutions.

As a result, this connectivity dampened the diversity inherent in the group and created a homogeneous whole that was less than the sum of its parts.

In the less connected group, the innovations appeared more slowly, but they were infinitely more diverse and creative.

“In fully connected groups, the individuals’ propensity to learn from successful cultural models–a common strategy that allows us to copy efficient solutions from others–quickly reduced the diversity of solutions. Partially connected groups are more likely to produce diverse solutions, allowing them to innovate further by combining different solutions,” the authors say.

The results suggest there is a fine line between being connected to diverse ideas, and being too connected and suffering from groupthink.

Suffice to say, the authors don’t explore the functioning of the groups and whether steps can be taken to ensure that groupthink doesn’t emerge in connected groups.

Nonetheless, it does seem that there may be a connectivity sweetspot that we need to aim for.