Turning Social Media Into Business Intelligence
Social media is becoming a powerful BI tool that enterprises can leverage to create personas, identify trends, and build communities.
- By Jasmine Morgan
- August 15, 2017
The emergence of social media has created a new reality as well as new tools for understanding human behavior. The applications are varied -- from marketing to crime prevention -- and the phenomenon is still growing. One of the best descriptions of social media is "a snapshot of the real world." This almost poetic definition is useful for determining the ways information retrieved from social media can be of use in business.
Because it is a snapshot, social media captures a certain moment in time or a brief set of close moments, when the user is focused on a particular goal or is trying to highlight an achievement. The lifetime of a post is only a few hours -- a few days at most -- until it is replaced with a new one. This rapid succession shows why companies need a real-time response: you either seize the moment or the window of opportunity closes.
The Social Media Landscape
The first challenges for organizations striving to harness the power of social media are the variety of environments available, the volume of data being quickly updated, and the number of daily active users. This is big data displaying the four Vs (volume, variety, velocity, and veracity) at their finest.
There are important structural differences and discrepancies between different social media platforms, and any thorough analysis should take these into consideration for accurate results. A recent study highlighted the difficulties in retrieving business-relevant data from reviews and proposed an analytical framework for comparing data across multiple platforms that can be used with a few modifications.
Beginning Your Analysis
When studying information from social media, it is helpful to note its five defining traits: scale, immediacy, heterogeneity, duplication, and semistructure. We've mentioned the first two. Heterogeneity refers to how information on social media contains a mix of data types, including text, hyperlinks, hashtags, pictures, and video. Duplication describes how viral content is shared many times; this can be useful in detecting trends and influencers. Semistructure means that although some information can be organized in a table, other data is more free-form and needs to be broken down to be classified.
Data retrieved by Web-scraping tools from social media must be divided according to the information type and the text items filtered by linguistics, semantics, and source. The information is then combined to determine the prevailing sentiment and can be further filtered by gender, age group, income bracket, and other demographics.
Depending on the goal of the organization commissioning the study, social media can answer questions related to the most important vertices in the network (influencers), create clusters of users with common traits, or highlight the perception of a brand by a particular demographic group.
Such an analysis creates dashboards with various drill-down levels that can answer questions such as: "What do females who are 25-35 years old with a median income care about the most?" Having a list of 10-20 answers ranked by other factors is pure gold for marketers. Starting from simple words, they can design a selling proposition that is congruent to the lifestyle of the target clients.
Before social media, the only way of creating buyer personas was by applying surveys and retrieving information from focus groups. Now, with the right tools, trends emerge naturally and clusters can be used to create detailed descriptions of profiles.
It is paramount for organizations to understand that social media is not about selling. As the name implies, it is about building meaningful relationships and finding your tribe, both people and companies. For example, by looking at hashtags, InData Labs found out that Instagram users aged 25-34 are interested in fitness, while users over 45 are most impressed by landscapes and city views. These results could be used by wellness companies or travel agencies to connect to the users most likely to be interested in a relationship with those enterprises.
Some people use social media as a tool for measuring success or to keep themselves accountable; they can have in-platform goals such as getting likes or shares or real-life goals they talk about, such as running a marathon. Companies can use both types of user-specific goals to reach their economic targets. Their marketing representatives just need to filter leads correctly by using appropriate interests and demographic/sociographic determinants.
These platforms are bringing people together in a new way. Before social media, individual communication was mainly one on one, and companies broadcast to general crowds. Now when individual users post, depending on their number of connections, popularity, content, and even the hour, the ideas could develop into a conversation or even a phenomenon. At the other extreme, many posts are not seen or commented on at all.
By looking at the patterns of these types of online interactions, companies can understand what makes viral content and what will end up in the Internet's basement. Companies can also target specific content to specific groups and tell whether anyone is viewing or interacting with their content -- something difficult in the old days of print and television advertising.
The London Example
Due to the scale of the information available, it is advisable to study just one to four variables at a time, draw conclusions, and develop adjacent studies if necessary.
For example, a recent InData Labs project wanted to know which brands are dominating London Instagram. Their study retrieved about five million posts. InData Labs analyzed the hashtags to identify brands and associated sentiments. Author demographics (including geolocation) were included in the research to create a heat map. The content of the pictures related to these posts was ignored. Sponsored or free types of posts were also not considered part of the experiment, to mention just a few of the possible extensions.
By limiting the elements studied, InData Lab's heat map revealed preliminary results that were used to direct further research.
Speak Directly to Your Customers
Social media offers a glimpse into potential clients' lives, goals, and desires -- as well as the clans they identify with. Making sense of information retrieved from social media by Web-scraping requires employing big data analysis tools to obtain real business value.
Including the findings in your BI tools can help you advertise more accurately by precisely striking the sensitive chords of your target audience and speaking their language. Think of the posts as peeking into your customer's private life with their approval. However, the tone must be genuine and consistent because users detect a company's fake tone much as they do with friends; they penalize this behavior quickly.
Jasmine Morgan is the technology consultant at GFT Group, where she is responsible for providing architectural leadership and support on large transformation programs for global financial services organizations. You can reach the author at [email protected].