NM Incite

Quick Links

The Social Marketer

Sep 6

Customer Service in the Social Media Age: A New Metric to Gauge Your Brand’s Performance

on September 6, 2012 - 13 Comments

Most marketers focus the bulk of their energy on creating campaigns that will enhance their brand health, but often overlook the importance of what people are saying publicly through social media about the organic customer service experiences they have. Through recent research, NM Incite uncovered that discussions about customer service experiences on Twitter are becoming a key driver of brand health and brand differentiation. For our first analysis we focused on the financial services industry, and the findings were eye-opening. We discovered that there is an immense opportunity for companies to track customer service performance through social media, and we developed a metric to quantify this and compare across brands. Before we get to the findings, let’s take a quick look at the metric and our process.

Customer service performance constitutes one of the key dimensions of brand health, and directly impacts customer acquisition, share of wallet and loyalty. Gauging customer service performance used to require private feedback loops with an inherent time lag. But those days are gone. Today, customers tweet publicly and instantaneously about customer service experiences they have in any channel (in-person, phone, email, online and social media – known as social care). They are also brutally honest about their feelings. Unlike traditional surveys that poll a consumer’s service experience and include a spectrum of satisfaction, tweets are predominantly binary and convey either highly positive or negative emotional views. Analyzing these posts unveils a simple and transparent barometer for how your company’s customer service is doing across the board. The Customer Service Sentiment (CSS) score, developed by NM Incite, gives companies a score that reflects the level of positive sentiment customers have toward that company’s customer service efforts. The higher the score, the more satisfied the customers are with their experiences.

Now, on to the results. For this analysis, we sampled tweets for 13 of the top financial services players to calculate their CSS scores. Surprisingly, the results reveal a significant gap between the strongest and weakest performers. This was a bit unexpected given that products and services in the financial services industry are often viewed as being commoditized. Discover, Ally, TD Bank and Citi all had a CSS of greater than 50%, with Discover topping the set at 79% positive mentions. At the low end of the spectrum was a major bank with a CSS of only 24% positive mentions, representing a gap of 55% vs. the top performer.

Customer Service Sentiment Scores for Financial Services

Despite a somewhat negative perception of banks among consumers, it turns out that some brands are actually driving a highly positive brand image via customer service.

Digging deeper, the public nature of Twitter allows us to go beyond the positive/negative split of CSS and determine the number of followers for each Twitter handle that posts a message. At a high level, this provides insight into how many followers are being exposed to customer service mentions, adding another layer to the impact of these scores. Across all the financial services companies in our analysis, we found that the average customer service mention was broadcast to over 1,800 followers, confirming the notion that an angry customer can have a serious impact that may ultimately turn a customer service problem into a marketing problem for brands. As a next step, we were able to look at the average follower counts broken out by positive vs. negative tweets, which shed light on another interesting disparity between companies.

Average Follower Count for Positive/Negative Tweets by Brand

Brands with the most positive CSS scores also appeared to have a much higher average follower count for positive tweets vs. negative tweets. On closer examination, it became clear that this was being driven by banks “re-tweeting” positive mentions to their generally large follower base. This in turn appeared to create a disproportionate improvement in the overall perception of the bank’s services. Conversely, some of the brands with more negative CSS scores were actually being undermined by influencers who had a large Twitter following and were on a mission against the company. These negative influencers were proactively seeking out posts that mentioned a negative service experience and then re-tweeting them to further their agenda.

This research pinpoints some key takeaways for brands. While the initial research is specific to Financial Services, the principles can span industries:

  • If monitored on a continuous basis, the CSS metric can provide a cost-effective, real-time snapshot of a company’s customer service efforts that can be benchmarked and measured against over time.
  • Once CSS tracking has been set up, deeper analysis of the tweets can then be used to identify and track service performance at the level of individual issues.
  • There is an opportunity for brands to become savvier and start amplifying positive mentions to a wide audience through re-tweeting. Taken even further, the opportunity can become an imperative in situations where there is an adversary focused on amplifying negative service experiences.

 
Are these opportunities being leveraged at your company to improve brand health? We’d love to hear your thoughts.

For more information on how to manage social customer service, download our white paper: The Social Care Imperative
  • matt_pierson

    “At a high level, this provides insight into how many followers are being exposed to customer service mentions, adding another layer to the impact of these scores. Across all the financial services companies in our analysis, we found that the average customer service mention was broadcast to over 1,800 followers, confirming the notion that an angry customer can have a serious impact that may ultimately turn a customer service problem into a marketing problem for brands.”

    I would think many of these customer service interactions involve tweets beginning with @mentions. Since these tweets will only be seen by followers of both the customer service account and the customer, in most cases they will not be seen by anyone else.

    • http://twitter.com/NMIncite_Erin Erin Tavgac

      This is only true for tweets whose text begins with
      the @mention, but not for tweets where the @mention appears somewhere other
      than the beginning of the text. For the data that we sampled, the
      majority of tweets did not begin with an @mention, and were therefore broadcast
      to the full follower base of the sender. The sample tweet shown in the
      above post is actually a good example of such a tweet that is seen by all the
      followers of the person tweeting. These follower counts also don’t
      reflect the amplification of tweets that occurs through iterations of
      Re-tweeting, which would drive additional broadcasting. Taken a step
      further, this same metric could be used to analyze customer service related
      tweets that don’t include an @mention at all, but do refer to the brand (e.g.
      Citibank or Citi instead of @AskCiti). The volume of such tweets is many
      multiples higher than those including an @mention (in most cases close to 5
      times greater) The point therefore remains that messages mentioning
      customer service experiences do indeed get broadcast to a wide audience.

      • matt_pierson

        ” For the data that we sampled, the
        majority of tweets did not begin with an @mention, and were therefore broadcast
        to the full follower base of the sender. ”
        I am extremely skeptical of this. This does not line up at all with my experience. A quick Twitter search on @AskCiti for the last 24 hours shows a much lower incidence. Of those tweets that did not with an @mention, many were retweets by the @AskCiti account. Those would only go to its own followers and inflate your average follower counts, since the service handles likely have big audiences.

        I’m also curious why you decided to take a sample instead of using the full Twitter firehose? You should be able to export the full list of search results and filter them in Excel to see how many start with @.

        These follower counts also don’t
        reflect the amplification of tweets that occurs through iterations of Re-tweeting, which would drive additional broadcasting.
        Can you explain why you chose not to consider retweets? Again, filtering search results by “RT” and “RT@” should give you the answer.

        Taken a step
        further, this same metric could be used to analyze customer service related tweets that don’t include an @mention at all, but do refer to the brand (e.g.
        Citibank or Citi instead of @AskCiti).
        How relevant would this be for a common term like Discover? It will be hard to filter that to relevant content given 140 characters to work with. And would be challenging to differentiate a customer service tweet from a generic brand mention. Citi is actively searching for broad mentions and other brands are not. Also not everyone who mentions a brand like this _wants_ to be contacted; some people find it creepy and intrusive.

        • http://twitter.com/NMIncite_Erin Erin Tavgac

          Matt
          - the sampling/research that we did is
          based on sound methodology backed by our measurement science team. If
          you’d like to discuss further please feel free to reach out at contact@nmincite.com

          • matt_pierson

            Unless you know how many followers the average Twitter account has (which would be a fascinating statistic), it’s bogus. Sorry Erin. Sampling mentions is not the same as sampling follower counts.

          • matt_pierson

            If I’m off the reservation you can feel free to correct me. There are lots of folks there with my contact info.

  • http://twitter.com/RobertoBattist8 Roberto Battistoni

    CSS is another name for net promoter and indeed a valuable metric. You make a very good point when you say that disparity in CSS can be the real USP in commoditised markets. I for example take every opportunity have to share my negative feed-back on O2 F-T-F, on FB, on Twitter, on LinkedIn: reach is massive!

    • http://twitter.com/LCafferty Leslie Cafferty

      I couldn’t agree more! But there is massive upside for marketers to promote positive messages through social and get that kind of reach as well – it just seems like many don’t leverage the opportunity as much as they can or should. Interesting comparison to net promoter. Since net promoter looks more broadly at brand loyalty, I wonder if there is a correlation between CSS and net promoter scores.

  • http://samwhiteman.me/ Sam Whiteman

    Does this score then penalize brands for using Twitter as a service channel? If Brand J above is using Twitter for complaints/service requests/troubleshooting actively then their n will be higher (hopefully on both +ve and -ve, but almost certainly on -ve).

    • http://twitter.com/NMIncite_Erin Erin Tavgac

      That’s a very interesting question. We didn’t look specifically for a correlation between CSS scores and usage of Twitter as a service channel. However, I think the score would be driven more by the “quality” of the customer service experience that the brand provides on the Twitter channel (as opposed to whether or not they use Twitter as a service channel). Our research focused on Tweets where the user was commenting on a positive or negative CS experience after it happened. The score doesn’t count the initial complaint/service request/troubleshooting Tweet itself. A brand therefore shouldn’t be penalized for participating in the Twitter channel unless the underlying customer service experience is itself more negative than positive. Having said that, this would definitely be an interesting correlation to test for in follow up research.

  • http://twitter.com/Melissahanfordc Melissahanfordca
  • http://www.salesportal.com/ SalesPortal

    Keep in mind that customers might be more likely to take their grievances and complaints to social networking sites more often than their good customer experience stories, so you might not be getting a complete picture. However, if the social sites are littered with angry customers that should be a red flag that something is amiss.

    • http://www.nmincite.com/ NM Incite

      Thanks for pointing out this nuance. It’s for this reason that we benchmarked financial service institutions against each other rather than look at one or two in isolation.