What makes Networks tick? Learning from (a lot of) experience

January 26, 2017

Reframing climate change: how carbon reduction can also reduce poverty and inequality

January 26, 2017

Handy NGO Guide to Social Network Analysis

January 26, 2017
empty image
empty image

Social Network Analysis has been cropping up a bit in my mental in-tray. First there was my IRC coverChristmas reading – Social Physics, by Alex Pentland. Then came yesterday’s post from some networkers within Oxfam. So here are some additional thoughts, based on a great guide to SNA by the International Rescue Committee.

Complexity and Systems Thinking seems to push people into two divergent sets of conclusions. One camp says ‘planning and toolkits suck, stick to close observation and response to any given context, and work from instinct, judgement and fast feedback’.

Another group cries ‘whoopee, a whole new set of toolkits!’ The problem with the first is how to ensure rigour and accountability, when everyone is just taking decisions according to their ‘instinct’ (what if someone’s instinct is disastrously wrong?) The problem with the second is that it can create exactly the kind of alluring but false certainty that complexity and systems approaches criticise elsewhere (logframes etc).

I lean towards the first group, with qualms, but like to keep an eye on the second, both to see if it is being oversold, and in the hope that it really does contain the seeds of a more rigorous approach to complexity. Social Network Analysis (SNA) looks like one of the most promising type 2 approaches. I’ve been worried that it is primarily being conceived as a tool for big guys with a lot of resources (see this post on Ben Ramalingam’s work, which prompted an impressive set of comments), so I was happy to come across the IRC’s excellent ‘SNA Handbook’, which provides a practical guide that seems well suited to NGO capacities.

In 9 pages, the guide sets out how to run a relationship mapping exercise, how to examine the influence of the different network members over the issue that is being addressed (pretty much the same as Oxfam’s Power Analysis approach). It then gets on to how to analyze the network, with some examples from an SNA by IRC’s team in Sierra Leone.

‘To understand social networks it is not only important to analyze the relationships between actors, but to also consider their location within the network and the overall structure of the network. The following questions may help to analyze the structure of the network:

  1. A) Are there any actors with a high number of connections?
  2. B) Are there any actors that appear peripheral to the network?
  3. C) How centralized or interconnected is the network?
  4. D) Are there any fault lines between or separate parts of the network?
  5. E) Are there any actors that link significant parts of the network together?

In trying to manage and mitigate risks the following are common issues to look out for in your network:

Examining the Influence of Network Members

Examining the Influence of Network Members

Dependency: The network maybe highly dependent on a single actor or a funding source, which can create bottlenecks and sustainability concerns.

Dysfunctional / conflicting relationships: There may be certain key broken relationships which

impede the entire network. New actors or interventions can also introduce conflict for resources or control.

Marginalization: Certain actors or groups of people may be excluded or marginalized within the network, perhaps owing to gender, ethnicity, status, income or other factors. Analysis of the reasons behind the structure of the network may help to uncover the reasons for marginalization and how best to overcome it.

Disincentives for change: Certain actors may have disincentives to support the proposed change and may try to actively oppose it. Pay particular attention here to how the intervention would change resource flows or change the levels of influence of each actor.

Like-me relationships: You may notice that actors (people / groups) who share certain attributes, such as gender, age, education, ethnicity, religion, status, tend to have many ‘like-me’ relationships and fewer relationships with people different from themselves. This is a common pattern in many networks. It may be worth considering how this affects the specific issue and how to overcome it.

Structural challenges: Structural risks may include an overly centralized network or a structural split within the network.

The Sierra Leone team identified risks at the community level related to the disincentive of

Analyzing the Network

Analyzing the Network

pharmacists, traditional healers and secret society heads to support community health workers, who represented a threat to their livelihoods and status.

In trying to capitalize on opportunities the following are common issues to look out for in your network:

Critical relationship building: There may be some very simple wins that you identify during development of the network map. For example, you might identify two actors who are positive and have influence, but these champions may not be connected. Facilitating relationship building between key actors may prove beneficial.

Tap into under-utilized support: You may identify actors within the network who are very positive about the change you seek to bring about, but who have not been given a role or sufficient voice within the proposed intervention. Give voice to these ‘champions’ and empower them to play a more central role.

Building networks within the network: There may be the potential for coalition building to raise the voice and influence of those who are positive about the proposed change. This can be done through more formal partnership arrangements or could be through organizing events to give a platform to those who share your ambitions.

The Sierra Leone team identified opportunities to build important relations between traditional and administrative leaders and health services providers to better coordinate support for community health workers.’

It then uses the SNA to think through possible scenarios:

‘SNA is useful for analyzing what the network looks like now. It can also be helpful for considering how it might change in the future. Participants may wish to consider how different scenarios would affect the network, for example:

  • What would the ideal network look like and how could this be brought about?
  • What would happen to the network if conflict were to resume?
  • What would happen to the network if funding ended or the IRC transitioned out?

If a funding source ends or key actor leaves, a functional network (A) can quickly break up (B). It is therefore a helpful to consider which relationships to invest in (C).’

Smart. Any other SNA tools that people can recommend for overstretched organizations on the ground?

And just because I recently watched it again, here is ecologist Eric Berlow talking (very fast) about the value of visualization in finding the ‘simplicity lies on the other side of complexity’, using my favourite complexity diagram – the Afghan military mind map, and lots of other stuff, all in 3 minutes.


  1. Fantastic, thank you Duncan. Look forward to digging into this. For software programs that map systems and networks well, I recommend:


    I have used it for evidence mapping and for showing connectedness, distributed leadership in networks. The base, like many of these, is free to use if you allow your maps to be public. There’s a good example of how Hewlett Foundation has used it to map the deliberations and dynamics of the America Congress (sigh).

    Nanci Lee

  2. I´ve heard NetMap is a workshop/interview-based tool, so less reliant on lots of fidly surveys, data entry and software packages https://netmap.wordpress.com/about/
    In a very interactive and visual way, it helps determine:
    – what actors are involved in a given network,
    – how they are linked,
    – how influential they are, and
    – what their goals are.

    I think there´s potential to combine it with some of the power mapping tools we´re already using to do a useful but fairly quick network analysis. I´d be keen to know if anyone has used it.

  3. I agree with the above comments, but in my experience it takes a huge amount of computing power. About 4 years ago we tried to map the diffusion of HIV prevention messages across NGOs of HIV prevention programs, but the computing power at UCLA then wasn’t sufficient. Berlow seems to have managed, though.

  4. How have others sorted out relationships between individuals vs. relationships between organizations? For example, organization A and organization B may have an MOU with each other, clearly pointing to a relationship between the organizations. However, staff member X at organization C may work personally with staff member Y at organization D on an initiative. How would one decide whether to characterize the relationship between X and Y as individual relationship between X and Y or an organizational relationship between C and D?

Leave a comment

Translate »