Chapter 11

The Social Network Index

NET PROMOTER SCORE: THE NEW AND IMPROVED SURVEY METHOD

Most of us consumers are surveyed to death, and we delete the majority of e-mail surveys we receive, hang up on the phone surveyor, or throw away surveys that come in the mail in an important-looking envelope. To combat this, a consultant from Bain & Company developed a brilliant solution: a one-question survey. The survey asks you to rate the product or service on a 1–10 scale as to whether you would recommend it. Simple, eh? Predictably, more people respond to these one-question surveys than to longer ones. There are four major downsides, however. First, it is still a survey of someone’s opinion, and happiness with a product or service does not necessarily mean I will buy it again. I would rate the Kia I had as a rental car as a 10 out of 10, but I would never buy one and give up my classic British car. I might rate my flight on United as a 3 out of 10, but I would fly United again because I want to earn my 1K status again to get free upgrades. Second, I don’t want to take up my time to give your business feedback when I am already mad or even if you did an okay job—I have better things to do with my time. Third, most people don’t respond to any survey unless they are really mad or really happy, so survey scores may not reflect the entire group of customers. Fourth, if you get a low score on a one-question survey, you have no idea why, unless you follow up with more questions, aggravating a disgruntled customer even more. Therefore the data from these one-question surveys is not actionable. There is no way to tell from a one-question survey why people are happy or angry about your product or service.

TYPES OF ORGANIZATIONS WHERE THIS METRIC IS APPROPRIATE

Any organization that sells products or services to consumers needs to think about using the social network index. Even though you may rely on others to sell and service your products, you need this index on your dashboard. What this means is that any organization that manufactures products or provides services to consumers needs to track this. Stores, restaurants, toothpaste manufacturers, car companies, package delivery companies, online retailers, and hundreds of other types of organizations need real-time data on what consumers are saying about them. What about business-to-business? There a lot of big companies that do not sell directly to consumers. Cargill, the largest privately owned company in the world, sells almost exclusively to industry. It produces and sells food products, everything from flour to corn to meat. It sells raw materials such as tanker cars of corn syrup to Coca-Cola and also sell eggs to McDonald’s for its Egg McMuffins. When a consumer drinks a Coke and scarfs down a couple of Sausage McMuffins with that Coke, the consumer often has no idea he or she is consuming Cargill products. However, it might be interesting to see what consumers are saying about the products made with your ingredients. It is also pretty important to hear what Coca-Cola and McDonald’s are saying about your products. Businesspeople use social media to communicate as well, and they are consumers also.

In short, just about any size or type of organization needs to have this metric. Schools, hospitals, government organizations, and others that tend to have more of a captive audience of customers should be monitoring what people are saying about their experience with your organization. The beauty of this metric is that you don’t have to spend hundreds of thousands or be a big corporation to track and use this analytic. The data is available, sometimes free of charge.

Another useful approach to measuring customer satisfaction without surveys is to monitor what customers are saying about your business on sites like Twitter and Yelp. Yelp is a site that started in San Francisco but has migrated to other cities; it gives consumers a way to review businesses using 1 to 5 stars and provide detailed comments on their experience. A small business that cannot afford surveys or simply doesn’t want to do them can monitor its reviews and use average rating as an overall metric. The reviews often provide a wealth of useful details on what people liked and did not like. The sad thing for many small businesses is that a few angry consumers can post things that may or may not be true about your business. Twitter or Facebook might be a better bet, but there is no overall numerical rating system like Yelp’s 5 stars, so it is harder to get quantifiable data. Yes, the words and comments are interesting, but it is tougher to sort and categorize them into useful summaries. There is software available to help you make sense out of all the tweets. Online surveillance of social networking sites has become the latest tool in the arsenal of techniques for measuring and predicting customer buying behavior. According to an article in USA Today, Indiana University has developed techniques for capturing daily tweets, analyzing them with mood measurement tools, rating the mood of the tweets as positive or negative and calm versus anxious, and then quantifying the data in summary form.1 So many people use Twitter and other social networking sites to comment on their experiences that this is a wealth of data that might be used to measure customer satisfaction without having to do surveys.

Data from tweets is viewed as unstructured data that is harder to quantify and hence less useful than ratings of 1 to 10 on survey questions, or even counting numbers of delayed flights, lost packages, or power outages. However, firms that have figured out how to tap into this data, quantify it, and use it for performance assessment and decision making are often leading their competitors. Several investment firms are even using data from daily tweets to make buy or sell decisions and beat the market. Thomson Reuters’s news analytics service measures how positive or negative a news story is about your firm and, combined with circulation information, it can use this data to predict actions such as a drop in stock price. Market Psych is a company that is selling data feeds from its proprietary mood measurement software to investment firms.

HOW DOES THIS IMPACT PERFORMANCE?

The impact of this analytic on all other measures of an organization’s performance cannot be underestimated. In the past, companies could easily ignore disgruntled customers or placate them with a small gift or apology. Hotels always use the ubiquitous fruit basket as the standard apology for any error or problem. I like Marriott’s new approach much better. When the valet was loading my luggage in the back of the car at the JW Marriott in Palm Desert, he knocked the suitcase into a glass vase my wife had back there and broke it. She had purchased it a few days before and we had forgotten about it. The valet was appropriately horrified and offered to refund the close to $100 for valet parking that the Marriott was charging for our three-day stay. Since the vase was $6.99, we thought that was a fair deal, so we gladly accepted his offer and waited while he credited my card for the parking fees. We left the Marriott feeling good rather than being mad about the vase. My wife immediately announced this to the world on Facebook as we drove away, no doubt causing a few of her friends to place inexpensive breakable objects in their trunks during their next visit to the Marriott. The impact of this valet’s quick action is that my wife and I are probably more likely to stay at this Marriott again, and I am more likely to seek out Marriott properties during my many nights on the road for business. So the action of the valet not only helped ensure my future business, but this episode was broadcast all over Facebook to my wife’s friends and contacts, further enhancing Marriott’s reputation for excellent customer service.

Being able to monitor good and bad qualitative data like this is invaluable to organizations. Not only can you learn whether a customer is happy or angry, but you can learn exactly what made them feel this way. The best part about this type of data is that it does not require you to ask anything of the customer. For some reason people hate doing surveys, but they love posting their opinions on Facebook or Twitter. I think I know why. When you fill out a survey, there is zero reward and zero feedback. Even if you do get feedback it is almost always delayed. Psychology 101 teaches us that behavior must receive immediate reinforcement or reward if it is to be repeated. There is no reward for filling out a survey. If you have negative ratings, there is often a negative consequence—more questions or someone calling you to ask about the reasons for your low ratings. Posting comments on social media often results in immediate feedback and positive reinforcement. Friends and contacts read your comments, and often provide their own comments about what you wrote. Write a comment about the new restaurant you went to last night being horrible, and several friends might respond that they loved it or agree with your perceptions. Whether people agree, like, or dislike your comments, you are getting feedback. Most of us like to be listened to, and we really like it when others laugh at what we say or agree with it. Social media provides a lot of immediate positive reinforcement for the behavior of posting our comments about a product or service. There is zero feedback and no positive reinforcement provided for filling in a survey other than an e-mail saying, “Thank you,” which just doesn’t do much for most of us.

As the way we communicate with one another continues to evolve, social media is quickly becoming the dominant force in customer buzz. According to a January 23, 2013, news release from CFI Group in Ann Arbor, Michigan, people shared their experiences regarding a business with others 47 percent of the time within the 2012 CFI sample, up from 45 percent of the time in 2011.2 Of those people who shared, a full 91 percent of them used social media in some form to do their sharing. Facebook dominated the social media channels, representing 33 percent of the “sharing volume.” What this data tells us is that almost half of consumers regularly comment on businesses they frequent, and that almost all of them use social media as the method of proving this feedback. What they don’t do is fill out surveys!

COST AND EFFORT TO MEASURE

The cost of using social media data to measure customer satisfaction can range from zero up to hundreds of thousands of dollars. If you own a restaurant with 13 employees and 40 seats, chances are you have to work as the host, busboy, waiter, manager, accountant, and possibly sometimes the chef. There is no way you can afford a company like Gallup or J.D. Power to do your survey, and you don’t have the time or expertise to do it yourself. Yet you are very interested in getting feedback from customers. Asking people about their experience as they leave might be one way to get real-time feedback, but many won’t want to give you negative feedback to your face, so they may just say it was fine and never come back. Also, data like this is hard to quantify, and you might tend to overreact to one negative customer who might be the type of person who complains about everything. A better way is just to go on a web site like Yelp where consumers post daily feedback about their experiences at stores, restaurants, and other types of businesses. Not only are there numerical ratings that can be viewed as quantifiable metrics, but there are comments to support the ratings as well. This can provide you with a wealth of information and you can ignore comments that may be unfair and just from one person and look for trends and common themes. So by reviewing the data on Yelp on a daily basis you have a way to get real-time quantifiable and qualitative data from your customers without spending any money. The one caution is that not everyone takes the time to go on sites like Yelp to write reviews, and not everyone goes on Twitter, Facebook, or other social media sites to post their comments. So you are not getting feedback from a lot of your customers, but the same is true of surveys, and I would guess that a lot more people use social media to post comments and ratings of a business than those who fill out surveys.

If you are AT&T or Nestlé, you probably are willing to invest some money to measure customer satisfaction, and because you have millions of customers, dealing with all this data requires the help of some professionals. Companies like Verizon can track daily tweets and Facebook posts that mention Verizon and can then detect whether the comment is positive or negative as well as the amount of emotion. If they find that most of the comments are about dropped calls and there is a lot of the negative emotion of frustration, Verizon can zero in on the geographic region where the comments originated and work on solving the problem. While this kind of data is clearly going to cost some big dollars to collect, its value is huge. Being able to pinpoint problems with real-time customer data has immeasurable value for an organization. Being able to identify exceptional service and tie it to individual employees is also useful for employee recognition and for figuring out how this great service occurred so it can be duplicated and deployed to other locations.

HOW DO I MEASURE IT?

One of the first steps in developing a measure of customer satisfaction via social media is to first decide whether you want to count all customers the same. Yes, all customers are important, but for most organizations some customers are much more important than others. This analytic is usually more meaningful if it includes both what people are saying about your products or services and who is saying it. For example, what an important restaurant critic says about your new menu is probably more important than what the average customer says. However, looking at the importance of an individual customer can be deceiving. The relatively unknown country singer whose guitar was broken by careless United baggage handlers became very well known when his ballad and YouTube video “United Breaks Guitars” went viral. The social media commentary of a single customer caused a drop in United’s stock price. Individual consumers have an unprecedented level of power these days. Through social media and eventually going on television shows like Ellen, a single teenage girl got Pepsi to remove a chemical from Gatorade that caused her health problems.

SORTING OUT THE BEAUTIFUL FROM THE UGLY CUSTOMERS

Lenders have been scoring the relative attractiveness of customers using credit scores for years. We all have some number from 400 to 800 that influences the likelihood we will get credit as well as whether we will get a preferential interest rate. Well, it turns out that we are being rated on more than just our creditworthiness. We are being rated on how influential we are. If you have an account on LinkedIn, Facebook, or Twitter, you are already being rated.

Influence scores range from 0 to 100. On Klout, the dominant company doing these ratings, the typical score is in the teens; a score of 40 would indicate a strong niche following. A score of 100 would indicate you are someone like Oprah or Lady Gaga. Marketers of the influence scores claim that over 2,500 companies are using this data. Audi recently announced that it would be offering promotions based on Klout data. Last year Virgin America offered highly rated influencers in Toronto a free flight to either Los Angeles or San Francisco. The good news from all this is that your level of influence is based on more things than how good-looking, rich, or powerful you are. The bad news is there is now another number we are getting labeled with that will impact how easy or hard life will be for us. A Klout score on its own is not a measure of customer satisfaction or dissatisfaction but a factor that allows you to assign weight to the importance of social media feedback. A 5-star rating and enthusiastic positive comments from a highly influential person are valued at many times more than a similar rating from an unknown person.

The second decision that has to be made is which social media sources you want to include in the index. Based on the study from CFI Group mentioned earlier, Facebook is the most common way for people to voice their opinions about products and services they buy. Yelp is something to consider, because it already includes numerical ratings that can be tracked, but I don’t believe that Yelp is used in all cities. Amazon is another place where people can write reviews about things they buy and give numerical ratings. However, this won’t work if you are a hospital or a dry cleaner. Each organization has to review the various social media sources where customers might comment on the organization and pick three to five of them to include in your social media index. Once you have selected the sources that are most appropriate for your organization, you need to assign a percentage weight to each one, based on its potential to be viewed by others.

A third decision you need to make is whether you want to quantify the verbatim data or just use ratings. A good part of social media data is not 1- to 5-star ratings but the comments people make. Not including the verbatim comments certainly makes it easier and less expensive to track social media data, but the feedback is not very actionable if you don’t review why someone rated you 1 star out of 5. Without any numerical data you will need software to analyze the posts for whether they are positive, negative, or neutral, and to assign a numerical value based on the degree of emotion displayed.

As with many of the analytics described in this book, your approach to constructing this index needs to be customized based on your budget, number of customers, and the type of organization where you work. A small community hospital or restaurant would be expected to have a very different approach than a big corporation with 25,000 employees. The straw man model presented here is for medium to large organizations with thousands of customers and probably at least hundreds of employees:

Each individual rating (assumes comments converted to 1–5 rating) is multiplied by the Klout score or some similar measure that evaluates the importance of the reviewer to the organization and its customers. Klout scores range from 0 to 100 points, which is a good scale to use even if you are creating your own metric. Rather than multiply the ratings by the importance of the customer, a more common and probably better approach is to code and analyze the data according to broad categories of customers in bands of 20 or 25 on the 100-point Klout scale. A third and even simpler alternative is just to compute average rating or number of stars like they do on Trip Advisor and other sites. That way each customer review is counted equally and the organization just looks to improve its average rating. Whether the Klout score is a factor in your analytic depends a lot on the type of business you are in. Many industries have identified people they view as thought leaders, or individuals whose opinions are important for the industry’s success. For a pharmaceutical company, these might be important doctors and researchers at prestigious universities or clinics. For a film production company, thought leaders might be reviewers and members of the Academy who vote on Academy Awards. For a bank these might be high net worth individuals, as they call them. If your employer is a medium-priced hotel chain, you might just want to count all guests’ feedback as being worth the same value and not bother with the idea of a Klout score.

If you decide that the concept of factoring in the importance of the customer providing the feedback is worth including in your social network index, then each individual rating (assumes comments are converted to 1- to 5-star ratings) is multiplied by the Klout score. As mentioned, Klout scores range from 0 to 100 points. You want to set up the metric so a strong positive or negative review by a really important customer has a big impact on your overall performance. Therefore, a few really important people giving you 5 stars and lots of kudos on Facebook can offset bad reviews by 15 unimportant people. Of course, the danger of this is that readers of Facebook posts or Trip Advisor ratings look at the overall number and percentage of positive and negative reviews, and don’t really care who did them. If you are a five-star hotel, it might be important for you to know that a billionaire frequent guest was not happy with his or her stay versus a middle-class family who used all their frequent flyer miles to pay for the stay and will probably never come back.

Regardless of how you decide to calculate the social index data and combine it into an analytic, this data should be tracked and reviewed daily in most cases. By reviewing this data every day you can quickly detect small problems, isolate them by location or even staff member or product, and take quick corrective action. If you wait until the end of the week to read the horrible reviews about the rancid walnuts in the arugula salad, you would have served a lot of bad salads and angered a lot of guests who might not come back.

VARIATIONS

The easiest and most common approach to getting started using social network data as a customer satisfaction metric is to simply use average ratings done by customers on sites like Yelp, Amazon, and Trip Advisor. Since this is the same data prospective customers will use when deciding whether to buy your product or visit your establishment for some surgery, this is probably a good metric to track.

Another variation is simply to use the Klout rating as a measure in and of itself. If you are a professor, writer, consultant, doctor, director, investment advisor, advertising executive, or are in any other profession where your image, credibility, and reputation are key, the Klout measure by itself is probably the best social network metric to use. My colleague Bernard Marr reports on this metric in his excellent book, Key Performance Indicators: The 75 Measures Every Manager Needs to Know.3 Like your FICO or credit score, the exact formula for calculating the score is not widely publicized, but it is comprised of three variables:

Marr suggests that his own score is 68 out of 100, whereas Barack Obama’s is 88 and Lady Gaga’s is 93. I haven’t bothered to check mine, but perhaps I should. Another common metric used by researchers and writers is citations—number of times others have cited your work in theirs. I just gave my friend Bernard a few points by citing his book.

TARGETS AND BENCHMARKS

General targets for a social media index would be to receive at least an average of 4 stars out of 5. Many people think twice about buying a product or giving their business to an organization with a rating of 3 or lower. Some will only buy things that are rated at least 4.5 stars. Overall, I would set targets as: 4 stars or above is green, 3 stars is yellow, and 1 or 2 stars is red.

Your targets for a Klout score or something similar like citations should be set based on studying competitors you admire. A benchmark for me is Bernard Marr, with his score of 68, since he is an author and consultant and we are in the same field. Another benchmark for me might be Dr. Robert Kaplan, who has written extensively on activity-based costing and balanced scorecards.

BENEFITS OF DATA

The benefits of getting real-time data from almost half of your customers on a daily basis are enormous, and include:

• Saving money by not having to survey your customers.
• Avoiding making customers angry by sending them surveys.
• Gathering verbatim comments that can be quantified as well as read for details.
• Diagnosing the precise reasons why customers are satisfied or dissatisfied with your products and services.
• Getting early feedback on new products or services before broad-scale introduction.
• Detecting minor problems before they become bigger problems.
• Showing your customers that they matter and that you listen to them by responding to their feedback.

NOTES

1. Adam Shell, “Wall Street Traders Mine Tweets to Get a Trading Edge,” USA Today, May 4, 2011.

2. “Social Media Best Used for Damage Control by Call Centers,” news release, CFI Group, January 23, 2013.

3. Bernard Marr, Key Performance Indicators: The 75 Metrics Every Manager Needs to Know (London: FT Publishing, 2012).