Who is your insurance company thinking about?
Do you ever wonder what’s important to your insurance company? It’s an important question because an insurance company can hold your future in its hands if you ever suffer a loss. It’s important to have a feel for what’s on their minds and to find out what their priorities and strategies are. Social Media, especially Twitter, can help.
To understand the character of a company we can use Twitter data. There are many attributes of a company that can help us understand the likelihood of the company being profitable, of it maintaining financial stability and of it being able to keep its promise of paying your claim. Unfortunately, most of these attributes are financially related and Twitter data doesn’t include any financial information at all.
One key attribute that would be very helpful to have in our quest to understand an insurance company’s character, aside from its financial health attributes, is a particular collection of interests of the company. Twitter can help here. For instance, if we could determine what the current hot topics are that the company is interested in, like a hurricane, the wildfires or perhaps the increasingly aggressive plaintiffs bar, we may be able to understand the company and its character better. The information about its interests could be an indication of its innovative or creative thinking, and could help identify strategies in which the company may be interested.
One way to gather the interests of a company is using the idea of homophily. Homophily is the tendency to form strong connections with others who are similar. It’s the old adage, “Birds of a feather flock together”. My assumption is that the insurance companies I studied in my last blog entitled “Whom do insurance companies follow?”, that is, AXA XL, AIG, The Hartford and WR Berkeley, are following Twitter accounts of people or organizations similar to themselves.
That being the case then, if I analyze the List names of the accounts that these companies follow on Twitter, I should gain some insight into the focus of these companies themselves. This is of interest for two reasons. First, from a competitive standpoint, I can gain another valuable data point in my understanding of the competition. What are my competitors thinking about, what are they focused on, whose opinions are they seeking? These are all keys to better understanding their strategic thinking.
From a personal or consumer standpoint, these interests will give me a sense about the culture of the company, which is important to me. Are they a company focused on charity, are they focused on diversity and inclusion, are the focused on building resiliency and helping their customers do the same? Essentially are they focused on what’s important to me? Do they value what I value?
So, to do this I wrote the following Python code which pulls in the friends of the four insurance companies that I’ve been analyzing.
target = “Insurance Company”
total_friends = 20
# Get the first 20 friends of the insurance companies and first few of each of them
for friend in tweepy.Cursor(api.friends, id=target).items(total_friends):
g.add_node(friend.screen_name)
g.add_edge(target, friend.screen_name)
print(“Processing:”, friend.screen_name)
data = api.lists_all(friend.screen_name)
# count the number of lists
print(“The authenticated user has “ + str(len(data)) + “ list(s).”)
for i in range(len(data)):
print(‘Name — ‘ + data[i].name)
I also then pulled in the List names of each one of these friends. Once I have these lists, I categorized the lists into “Interests” as can be seen in the following tables.
These data points can now inform my opinion of the company with whom I choose to do business. Clearly, there are some additional questions that need to be asked about this analysis. The first question is, can this attribute be validated before it is used in any analysis or serious decision making? Because the idea of homophily is really based on an assumption, a clear validation is often difficult. However, one could look at recent industry press on the company in the trade journals to get a sense for what the company is up to. Once could also visit the company’s website to view recent press releases or company communications.
There are several ethical issues surrounding this effort. For instance, do people with Twitter accounts appreciate that their lists are publicly available? How accurate are the assumptions that I’ve made suggesting that someone is interested in a topic just because their Twitter List suggests it? The legitimacy of some of these assumptions is somewhat suspect.
There are also several limitations to this data. For instance, not every account has a list. This fact actually raises the question about what types of accounts create lists? Are they different from the accounts that do have lists? A study on this topic may be enlightening.
Additionally, those accounts that do have lists may not actually be interested in everything on the list in the way I have interpreted.
Lastly, the friends of the companies may not necessarily relate in any real way to the company being analyzed. We have no confirmation of why the company has followed a particular account. Perhaps it was just a one time post that caught the company’s attention.
In summary, this analysis has used the idea of homophily to identify possible interests of a set of insurance companies. This analysis, while it does have some limitations, will certainly provide the user with insight into their insurance company that they may not have otherwise had.
