This content is part of the Essential Guide: The power of enterprise collaboration software and tools

Is CRM with social media growing less trustworthy?

Comments made on social media can sometimes give businesses a skewed perception of consumer sentiment. Learn how to correct the data to inform business decisions.

Social media is a godsend to marketing departments everywhere. For decades, they employed endless, semisuccessful strategies to goad customers into sharing information about their likes, dislikes, preferences and wish lists. In the last 10 years, the marketers' gift of social media grew into a forum where customers log on and talk -- all at once, all the time.

CRM with social media followed shortly after, exploiting those conversations to give marketers invaluable insights into the customer mind -- spontaneous and unguarded insights.

If only those insights were reliable.

What goes on online

Acknowledging that social media platforms, such as Facebook, Twitter and Instagram, provide oceans of useful data about customers -- how they think and feel and what they want -- it's important to understand exactly what's happening in a social media forum when the enterprise brand is mentioned and to interpret the data accordingly.

Put simply, what is said about an enterprise brand on social media can't always be trusted because the context of the comments being mined is often unclear and that can invalidate or even corrupt the data. Here are some reasons why:

  • Social media discussions tend to be loosely structured. More often than not, when a brand surfaces in a social media discussion, it does so incidentally. That is, it isn't the object of discussion; it just happened to come up. For instance, a bicycle manufacturer's brand is mentioned in a group discussion by cycling enthusiasts, and a particular bicycle model is spoken of negatively. But neither the brand nor the bicycle is the topic; the discussion was about extreme mountain biking.
  • Social media participants are socially diverse. Using CRM with social media, businesses take whatever they get. They can be so excited by the spontaneity and presumed honesty of the input that they pay little or no attention to the profile of the customer doing the talking. Sometimes, social media participants jump in just to participate, with little concern for the content of their post -- it's being included that counts. In the above example, several cyclists might join in dissing the brand just to be present, when they don't really dislike that bike at all.
  • Social media omits the most meaningful metrics. In marketing surveys of old, at least a cursory customer profile was attached to the survey data: the age and/or gender of the customer, if nothing else. That's usually absent in brand sentiment analysis. Worse, no gradations are available in social sentiment data -- nothing like the "strongly agree ... strongly disagree" Likert Scale with which everyone is so familiar. In the biking example, there is no indication of how strongly anyone feels about the hapless bicycle being criticized, with the possible exception of the first participant.

This leaves a negative brand sentiment analysis, due to keyword-based mining of dialogues that really weren't about the brand -- both uninformative and misleading. Suppose, for instance, that the original poster used that particular bicycle out of necessity, rather than choice and that it's a bike that wasn't appropriate for the terrain in question?

Benefits of social CRM

Expressing opinion vs. piling on

Most people spend enough time on social media to understand there's a difference between writing a post expressing something important vs. piling on when a bandwagon starts rolling. The original opinion can be trustworthy, and by going beyond the keywords into the surrounding text, you can often get a measure of the intensity of the sentiment.

The amount of dissent or counteropinion found in a discussion is often a random function of the personalities speaking.

Beyond that, the amount of dissent or counteropinion found in a discussion is often a random function of the personalities speaking. Some people go along because they authentically share the same opinion; some pile on just because piling on is fun. But some people express a different view because they truly feel differently, and others have trouble doing so, despite what they actually think, because they are, by nature, conformists. Social media discussions can be too random, too spontaneous and too unstructured to be trusted.

The ideal middle ground

Though social media can often be an unreliable source of information, it can still be a great place to look for authentic brand sentiment -- with a little extra effort.

  • Deploy your brand ambassadors. Many CRM packages now boast the feature of brand ambassador management -- the identification and management of brand enthusiasts who can be deployed on social media to promote the brand. Find those people, and make use of them. Have them begin focused, controlled discussion of the brand on the popular platforms as neutrally as possible, and make those discussions your primary sentiment source.
  • Ask a neutral leading question. Have your brand ambassadors kick off their discussions with no pro or con bias. An anecdote is often a good starting point, and one that solicits similar experiences or stories typically spurs candid -- as well as diverse -- responses.
  • Make the discussion a safe place. Make certain the brand enthusiasts who lead your brand discussions take care to keep those discussions free and open, with no slippery slopes of disproportionate emoting or personal comments. You will get better sentiment data if all participants who feel like expressing an opinion feel equally safe, whether saying something positive or negative about your brand.

Social media is the marketer's dream come true. The trick is to not get too excited by the superficial outcome using CRM with social media and to do a little extra work to boost its integrity. Down that path, reliable customer input can be found.

Dig Deeper on Customer experience management

Content Management
Unified Communications
Data Management
Enterprise AI