Want Old Ideas? … Then Keep Talking to Your Friends


If you have been reading this blog for a while you will know that a lot of the research work that Tim and I do looks at the link between networks and innovation. When we talk about networks, we mean all sort of ways that people and businesses can connect to each other. For example, we have studied networks of project engineers, a virtual network within a multinational corporation, networks of entrepreneurs in Taiwan, and networks of academics – just to mention a few. Tim’s doctoral thesis was on the evolution of the world trade network and that had some important implications for free trade agreements. We think that networks are a really cool thing to study!

When we do this research we try to link network structures to outcomes. In the innovation context this means that we want to know what network structures are associated with better innovation performance. This morning Tim sent me a link to an article in Slate Magazine that summarizes a famous research study of networks and creativity in the Broadway musical industry. The main finding is that a particular class of network stucture called a ‘small world’ is correlated with the appearance of blockbuster musicals. When this structure appears in the networks of scriptwriters, choreograhpers and librettists, then it is more likely that a highly successful musical will be produced.

Small-world networks have a particular signature of clusters and sparse links between the clusters. They are called small worlds because the handful of links between the clusters create shortcuts between anyone in the network. The general idea is that this makes it easier for new ideas to flow within the network.

One obvious question to ask is why doesn’t a network with more links between the clusters work even better? In the Broadway study they found that improving the connectedness of the network with more shortcuts between clusters resulted in less musical success rather than more, so what is going on here?

My explanation comes from another classic network study that is over forty years old. In a famous paper called “The Strength of Weak Ties” Mark Granovetter highlighted the importance of weak network links. This was demonstrated in his study of how people find a job. It wasn’t the strong connections that were the source of information about a new job, it was actually the weak connections to people you don’t know very well that were the most common source of an opportunity.

When you think about it, this makes perfect sense. Tight networks circulate existing information. It is unlikely that you will get anything new out of your closest friends. You know them and what they know, and vice versa. It’s the bridges into unfamilliar networks that are the better sources of opportunities and ideas. In terms of the broadway musical study, this means that better connected networks don’t necessarily result in more innovative musicals.

This also has a very practical implication for our own networks. Are you in a densely connected cluster? How can you get out and make just a few random connections?

Note: Small-world network image is from Six Degrees by Duncan Watts.

Please note: I reserve the right to delete comments that are offensive or off-topic.

11 thoughts on “Want Old Ideas? … Then Keep Talking to Your Friends

  1. Very good point, John!

    I too was mulling over the Slate article yesterday after Tim’s tweet pointed me to it.

    What I find particularly interesting is that the findings are often interpreted ‘locally’, e.g. with regard to team performance, such as: for innovation your team shouldn’t have too much experience working together, but they shouldn’t be total strangers either.

    The article mentions:
    “Uzzi and Spiro found little or no relation between success and local measures, such as the extent to which team members had worked with each other on previous shows. The success or failure of a Broadway show had less to do with the relationships between the names in the Playbill than the shape of the broader network at the time the show was produced.”

    I think this is an important point as not immediately intuitive.

    Cheers, Ralph

    • Hi Ralph

      It’s a really remarkable paper becuause it suggests that broader networks trump local effects. Diversity in network connections is more significant than team diversity.
      Thanks for bringing that point out more clearly.


  2. That’s an interesting article, I wonder why ‘ideal’ small worlds lead to innovation. Perhaps it has something to with intermingling after isolated evolution – each small world evolves its own tropes and each Broadway show forces the now different worlds to collide and collaborate.
    I think you’re right about weak ties John. Strong ties help maintain the status quo. Weak ties on the other hand allow for change, adaptation and learning. A good example of the ‘strength’ of weak ties can be found in the behaviour of swarms and animal packs.

    • Hi Matthieu
      The original paper from 2005 by Uzzi and Spiro talks about the tension between diversity and cohesion. Too much cohesion with your friends being friends of friends tends to block out new ideas. I think your point about preserving ideas in different network clusters is right.

  3. It’s very possible I’m missing something fundamental here… The Small World mechanism makes sense to me but I’ve not got why the Weak Ties thing is the reason more connections leads to less success. Surely Weak Ties is more of a local effect? It’s more likely we will find a job through our Weak Ties because we can have lots more of them. The more ties (of any sort) we have – the better the chance of finding the right job. It seems to me that the reason weakly connected Small Worlds work is because ideas can leap between them but groupthink can’t.

    • Hi Mark
      the institutional effects of network clusters as they block out new ideas is the reason given in the paper. If you Google Uzzi and Spiro 2005 you can get the full paper.
      I’ve been interested in the interaction between personality, local network effects and global network effects. For example, one study of an industry that we are looking at suggests that risk averse people like to surround themselves with contacts who can provide information, this aggregates into a clustering effect in the network.
      Similarly, weak ties are more likely to bridge disparate groups because there are not the same social forces for group convergence. Some people have particular personalities that can bridge groups and they tend to hold brokerage positions in the network.
      I need to write another post about the ‘strength of weak ties’ argument. I’ll try to do that next week. Thanks for the comment- this an important point that needs more discussion.

Comments are closed.