Saturday, July 26, 2014

Can Social Network Analysis Be Implemented For Social Change?

In our rapidly changing world, traditionally effective marketing strategies have to be adapted to meet the needs of a very diverse set of demographics.  This is particularly true when dealing with the underserved markets at the bottom of the socioeconomic pyramid.  How do you penetrate a market that struggles with low literacy rates, lack of trust, and lack of access to traditional media channels?  I believe that social network analysis is the first step.

The dynamics in many slums are based on relationships and unspoken hierarchies.  Following the advice of key community influencers are how many people make their purchasing or lifestyle decisions. The problem is, how do you identify these influencers?

Assuming that money is no object, we would need to survey community members in order to find out where the social dynamics gravitate.  This is particularly difficult to do because, unlike other surveys, you need a significantly larger sample size in order to get an accurate picture.  This is because there are many networks present, and you can'r rely on confidence intervals to determine your levels of accuracy in this case.

For the purposes of this scenario, let's assume that we were able to get an 80% response rate.  This is probably extremely ambitious, but it's always best to aim high.  This survey asked the respondents who they socialized with, who they took advice from, who do they respect, where do they live, occupation, etc.  All of this information will be useful when mapping out the social analysis map later.

Once all of this information has been compiled and the connections have been made, the data would need to be studied.  First, it would be important to determine which people have the highest eigenvector centrality for both "respect" and "advice".  This will identify the key influencers.

Afterwards, it would be helpful to try to find any correlations between those with high eigenvector centrality and any of the other descriptors mentioned (age, occupation, address, etc).  This is important because it helps to answer the question of "why?" rather than just who.  If you can determine WHY someone is a key influencer, you can create a better communication plan because you better understand the needs and wants of a community.

All of this is the first step in crafting a marketing strategy in an urban slum.  Once you have been able to map out the social networks and perform an in-depth analysis, you can make an informed decision about how to move forward with your business strategy.  This will also prevent you from rushing in blind to a potentially hostile business situation.


2 comments:

Christopher Tunnard said...

80% response from a BOP survey is indeed ambitious. But, let's say you could get that. I wonder whether a question about purchasing patterns is the best approach? Or whether eigenvector is the appropriate measure? There are other ways to get at the "why?", which would have to do with where they live, mobility, tribal/family relationships, access to technology (or even electricity.) Depending on the country, you'd want to get at use of mobiles. There's a lot to consider, but you could put together an SNA that might lead you to the outcome you want, which is...?

Christina Filipovic said...

Interesting idea: using SNA to understand the BoP. Deciding what you plan to "sell" might help to better inform your "advice" and "respect" networks. For example, the friends we ask for career advice are different from those we ask for romantic advice. The friend I would ask for advice on which cell phone to buy next is different than the one I might ask about a health or beauty product.

Also, there might be a more effective way to determine who is "respected" than ask outright.

In addition, I would like to get a sense of the size of the community you would want to study. For a very large slum, you would need more focused data (purchasing patterns?) to find important insights.