Apples and Oranges - a Problem for NPS

I think we have been missing the mark when we use NPS as a critical measurement for customer satisfaction. The challenge to this metric is that it appears to be combining "apples and oranges". The metric is defined as the difference between attractors and detractors. That is the problem!

What makes this an apples and oranges metric is that we are combining two different characteristics. The assumption in the NPS metric is that attractors and detractors are measured on the same scale. If attractors are apples and detractors are oranges, you can't subtract oranges from apples. Let me expand on this notion in the following paragraphs.

According to the literature, attractors characterize satisfaction. It carries with it the notion that attractors represent aspects of the product or service that encourages the customer to continue using the product or service because the customer is satisfied. Further, detractors characterize dissatisfaction. Dissatisfaction represents some aspect or aspects of the product or service that has a negative influence on the customer.

Here is the problem. Satisfaction is not the opposite of dissatisfaction.  A low satisfaction score may lead to indifference.  As satisfaction scores drop satisfaction does not directly lead to dissatisfaction.  Nor is dissatisfaction the opposite of satisfaction for the same reasons. Another way of saying this is that dissatisfaction is not the absence of satisfaction, and satisfaction is not the absence of dissatisfaction. Each of these terms, satisfaction, and dissatisfaction, have separate scales that may have the same number of items (such as 0 to 10). You can have a scale of 0 to 10 for satisfiers and a complementary scale of 0 to 10 for dissatisfiers. The current assumption for NPS  that the units of satisfaction are the same as units of dissatisfaction. An example of this is a one-unit positive movement of satisfaction that is assumed to be identical to a one-unit negative movement of dissatisfaction.

Consider an example of a customer who purchases at McDonald's. McDonald's uses a business model that focuses on providing food quickly (it is a fast-food restaurant) and consistent quality (although it may not be the highest quality), it is designed to provide consistency for food quality). Customers would have levels of satisfaction for speed of service and quality of food. And the scale for each could be from 0 to 10. Another component of the service experience of the customer might include using the restroom facility. The restroom facility does not have a satisfaction component. Customers don't usually seek out McDonald's for the quality of the restrooms. The restrooms are considered a convenience for customers rather than a satisfier (a feature that would encourage customers to return). The restroom scale would act as a dissatisfaction scale. It does not add to the satisfaction of the customer experience.  The restroom may use a dissatisfaction scale from 0 to 10. In this case, 10 would indicate no dissatisfaction and zero would indicate complete dissatisfaction.

When we put this customer experience in perspective we see satisfiers (apples) and dissatisfiers (oranges).  A simple question might be whether one unit on the satisfaction scale is equivalent to one unit on the dissatisfaction scale.

I hypothesize that satisfiers and dissatisfiers have different scale values. Hence the assumption that the scale for satisfaction and dissatisfaction used by NPS is equivalent is most likely not true. Customers do not stop going to McDonald's if the service is slow or the quality of food is not as good as usual. Customers will most likely stop going to McDonald's if the restrooms do not appear sanitary.  Thus dissatisfiers may act significantly different than satisfiers. 

The bottom line is that the use of the NPS metric gives a distorted view of the customer relationship.  It is based on the faulty assumption that satisfaction and dissatisfaction are equivalent and can to be measured on the same scale.  Companies that use this metric are likely to be misled about the quality of the relationship with their customers by making decisions based on this metric. It's time that we provide a metric that legitimately considers satisfiers differently than dissatisfiers.

CSAT vs CES - Does It Matter

Many commercial businesses that provide after-sale services seek ways to improve customer loyalty through their customer service organizations. In addition, companies want to utilize their resources as productively as possible. Since there are many dimensions to customer service, companies want to understand how best to utilize limited resources. The basic premise that most companies rely on is that satisfied customers are more likely to increase loyalty than non-satisfied customers. For that reason a number of metrics have been developed which attempt to measure the influence of various customer interactions. While metrics can be used both tactically and strategically, this research has focused on metrics used to measure tactical performance.

Two of the more popular metrics are Customer Satisfaction (CSAT) and Customer Effort Score (CES).  The goal of this research is determine which of these two metrics will have more impact on customer loyalty. Since each of these metrics has a different objective, the question is whether they could be used interchangeably.  Is one more effective than the other, or should they be used complementary. The CSAT metric is designed to maximize satisfaction whereas the CES metric is designed to minimize customer effort.  The first issue is whether the two metrics yield the same service factors for resource allocation. The second issue is, if the metrics do not yield the same factors for research allocation, then which is preferable. The decision criteria for selection of which metric would be preferable is the strength of the correlation relationship between the independent service factors and either the CSAT or CES metric.  The metric with the higher correlation would make the most sense to use.

This study included 6381 customers from the medical electronics and IT industries. Surveys were taken during the months of April and May 2015 and included customers in both North America and Europe. The results of the surveys indicated a relatively low correlation (0.51) between the two metrics, CSAT and CES; however, the results yielded similar relationships between the independent service factors and each metric when the total sample was considered. These results also held true when the total sample was examined independently for field service and technical support. However, when the sample was examined separately for medical electronics and IT equipment the most highly correlated independent factors  with CSAT and CES were different for the two products.

As a result of this research we notice that maximization of satisfaction and minimization of customer effort only provide similar rankings for the independent variables. Since the two product groups which were included in the study yielded different rankings for the independent variables, a generalization seems inappropriate for concluding that the two metrics are interchangeable.

 Because this research focused on two very different groups of technology products, the solution for businesses wanting to increase customer loyalty should be to consider the use both metrics since it is not clear what the criteria should be for using one metric in lieu of the other. When both metrics are used the question is which of the independent variables are most appropriate for the products.  Until an individual product area is evaluated, there is no obvious preferred method. To choose one over the other without the understanding of the relationships between the individual independent variables and CSAT or CES may lead to inappropriate allocation of resources.

Loyalty Model – part 2

In the previous blog, the general form for the construction of a loyalty model was postulated.  A base equation hypothesized that loyalty could be described by measuring the strength of the relationship between the company and its customers.   A sub-model was proposed to show that the strength of the company-customer relationship could be explained by various factors.  In this blog, the two major components of the model will be discussed.

The two components that describe the company-customer relationship are those that strengthen of the relationship and those that diminish the relationship.

Factors that enhance and strengthen the company-customer relationship may be considered relationship builders; “satisfiers.” Satisfiers represent activities or involvements that yield positive experiences by the customer and also provide value to the company. In the service business, response time tends to act as a satisfier.  For example, when services are requested, the time to respond will strengthen the relationship between the customer the company as long as the response time meets or exceeds the customer’s expectation. Consistently meeting a customer’s expectation of response time for service has been shown to strengthen the relationship with the company. More discussion of satisfiers will be provided in a separate blog.

Factors that diminish the company-customer relationship may be considered “dis-satisfiers.” A dis-satisfier is generally not the opposite extreme of a satisfier. As an example, consider customers who frequent a fast food establishment. Customers expect a reasonable quality of food to be delivered quickly. If either the quality of the food or the service delivery time does not meet the normal expectations, the customers may be disappointed but probably not dis-satisfied. However, if the restroom at the same fast food establishment has not been adequately maintained, many customers may refrain from returning to the restaurant due to a concern that lack of good hygiene in the restroom may be an indicator of lack of good hygiene in the kitchen. In this case, the quality of the hygiene in the restroom can be a dis-satisfier; but the hygiene quality of the restroom is surely not the major attraction of the fast food establishment and is not considered a satisfier. More discussions of dis-satisfiers will be discussed in a separate blog.

Company-customer relationships are not linear.  They are not a function of adding the satisfiers and subtracting the dissatisfiers, which is similar to the NPS metric when it subtracts the detractors from the promoters.  This simplification of the loyalty model makes no sense since it is equally equating positive values (satisfiers) to negative values (of the satisfiers or possibly dissatisfiers).  Beyond the basic NPS score, many analysts mistakenly assume that the relationship between each component of the model has a linear relationship with the strength of the relationship. 

As we examine the components that increase the value of the customer relationship do not assume that each component is independent and has a linear relationship with the strength of the customer relationship. There will be some satisfiers that may increase the strength of the customer relationship dramatically, whereas other satisfiers will only provide an incremental increase of improvement in the strength of the relationship.

The same logic also holds true when we discuss the dis-satisfiers in the future blogs. Always challenge the assumptions! Be curious.   

Building a Loyalty Model

The kind of loyalty model I will be discussing in the next several blogs is a basic business model that is often used in strategic management. The basic premise of the model is that customer loyalty leads to profitability. The purpose of this model is to develop an understanding of the business components that contribute to the loyalty of the customers.

In this blog I will describe the general loyalty model and a sub-model which feeds into the general loyalty model.

The hypothesis upon which the loyalty model is based is that loyalty is a direct function of the strength of the relationship between the company and its customers. Logically this makes sense since a strong relationship between the company and its customers should produce greater loyalty than a weak relationship. The heart of the loyalty model is built around understanding the components that make up the strength of the relationship. There are many variables that contribute positively or negatively to the strength of the relationship. Recent experiences between the customer and company would be an obvious component of the strength of the relationship.  A single experience between customer and the company may not significantly influence the strength of the business relationship. In fact, there is a “zone of tolerance” that ranges from minimally acceptable to extremely exceptional. As long as the other factors exist even an incident that was less than minimally acceptable may not change the strength of the relationship. Of course, this assumes that any negative experience may be resolved by the company. Otherwise, there may indeed be a negative change to the strength of the relationship between that individual customer and the company.

Therefore the simplest form of the loyalty model is as follows:

            Loyalty = constant x (strength of the relationship).

The value of the constant associated with the strength of the relationship will vary with respect to product, geography, and possibly other variables. Thus, a company with multiple products may have a different strength of relationship with their customers based solely on the product. An example would be the strength of the relationship between the Apple iPhone and its customers versus the strength of the relationship between the Apple iPad and its customers. iPhone users are known to have a very strong loyalty connection with Apple; whereas, iPad users may not have the same level of loyalty.

The strength of the relationship (loyalty) is another way of describing the business relationship between the company and its customers. The model proposed would consider variables such as level of satisfaction, recent experience, product quality, commitment to the relationship by either the company or, the customers, and the bonds may exist between the company and its customers.

The bonds that exist between the company and its customers may fall into a number of categories; such as, legal bonds (contracts), technology bonds (shared or licensed technology), knowledge bonds (shared information), social bonds, geographical bonds, cultural or ethnic bonds, and economic bonds. There may be other bonds as well.  The fact is there are many bonds that may exist and give strength to the relationship between the company and its customers (such as loyalty points).

The general form of the model for the strength of relationship is:

            S (Strength of relationship) = function of (recent experience, level of satisfaction, product     quality, service quality, commitment to the relationship, and appropriate bonds).

This model presents a challenge of quantifying each variable and assessing the strength of the relationships between each variable. This model can be simplified by assuming minimum interaction between the variables contributing to the strength of the relationship and assuming the variables are linearly related.

This discussion will be continued in the next blog where a simplified version of this loyalty model will be presented with an example. At this point, a simple loyalty model is presented that includes most of the significant variables that are usually included in loyalty discussions. Be curious.

Another look at your customers

In my previous blog I discussed the tyranny of the urgent and why is such a limiting perspective in understanding customers. I closed the blog by pointing out that the focus should always be on all your customers not just the ones who have urgent concerns. In this blog I will point out several techniques which may be useful in understanding the needs of our customers.

Several years ago I was working with the company that had a very complex product which required significant service support. The concern of the company was that the users of the product were not the decision-makers but had significant influence about decisions regarding the product. The product was a complex computer that used very sophisticated software. The end users of the product were scientists. The computer was managed by the IT department. Neither the users of the product nor the IT department made the decisions regarding the product. The decisions were made at a management level above the IT department and the users department (which included all the scientists).

The challenge with this assignment was that there was no department that was responsible for evaluating the performance of the product. Each of the three groups (users, IT, and management) were involved in the decision-making process regarding the performance of the computer. The measure of satisfaction for any one group is not sufficient to understand how well the product is meeting the needs of the customer. This product required a multidimensional model of customer satisfaction that incorporated satisfaction metrics from each of the three departments. The objectives of the metrics were to assess the satisfaction with each organization department and evaluate any inconsistencies in the measures of satisfaction between them. In other words, although satisfaction of each department was important, it was equally important to determine if there were inconsistencies or discrepancies between the scientists, IT department and upper management.

While the metrics for a current customer is important, and is the basis of most customer surveys, some additional areas of interest include the following:

1.       1. measurement of concerns from lost customers,

2.       2. specific measurements directed toward ultra-valuable customers, and

3.       3. measurement of the gaps between customer expectations and the performance delivered.

Reflecting back on the previous blog, the intention here is to provide some areas of interest beyond the basic customer satisfaction survey. The previous blog pointed out the need to separate normal survey responses from responses to customers with urgent needs for support. Curiosity is the watchword for surveys. There are many dimensions of involvement between the company and its customers. Not all contacts between the customer and company are from the end-user.

As an example, consider a previous assignment with another technology company led to the surprising conclusion that the Accounts Receivable department did not have customer skills training and was the primary reason for customer irritation. In fact, the Accounts Receivable department put IBM on credit hold because their payment was overdue. Needless to say IBM was not happy with this treatment.

If your curiosity is great enough, you may find more connections between your company and the customer that may be worth exploring. You may also find that many of the employees involved in those connections do not have customer management skills training.  Be Curious!

Who are you listening to?

The “tyranny of the urgent” is a phrase that is commonly used when working in a service environment.  The phrase reminds people that “urgent” requests often take priority over “important” requests.  Service managers often find themselves in the role of “firefighting” - putting out “the fires” of customer complaints.  Before the manager realizes it, the firefighting consumes all of his time. The executives will note what a wonderful job the service manager is doing by managing customer complaints.

In this situation the only customers whose voices are heard are those who are complaining. The large and small customers who were not complaining are overlooked. When the urgent requests consume most or all of the service manager’s time, there is little time remaining to consider what is not urgent.

In his book Exit, Voice and Loyalty, Albert Hirschman argues that unsatisfactory conditions (such as service problems) may lead customers to (1) to exit or leave the company without trying to resolve an issue, or (2) speak up and try to remedy the situation.  The “nice” customer will exit without telling you why and never return.   A “good” customer, on the other hand, is willing to speak up and tell you what is wrong and is willing to work with you to resolve an issue. A “good” customer is really your best consultant because he will tell you exactly what is wrong.

It is often said that the urgent requests are like customers pounding on the chest of the service providers and managers while the other customers who do not require immediate assistance or attention are ignored.  The fact is that customers’ lack of urgency requests may actually have more important information to share with the company. Many companies spend too much time listening to the urgent customers while ignoring other customers.

Urgent needs of some customers inhibit service managers’ time to anticipate, plan and develop future service requirements. Too many managers spend most of their time in a reactive mode. To get beyond the “tyranny of the urgent,” companies, especially service organizations, need to fundamentally change the way in which they listen to customers.

From a strategic perspective, a balance is needed between urgent problems and important problems. The urgent solutions generate “at-a-boys” for the service personnel while the non-urgent solutions pave the way for long-term success.  A simple first step to managing the “tyranny of the urgent” is to set aside a specific amount of time each week or each month to solve the problems of how to increase productivity, review service skills, parts inventory, and technology.

The bottom line is to focus on the needs of all your customers.  There are three tactics that will provide the first steps to leave the “tyranny of the urgent”.

Step one: Conduct interviews with customers who have left your company.

Step two: Engage those customer segments that are not being addressed by those customers with urgent needs.

Step three: Understand why customers left you in the past. 

Be curious! You may discover/uncover challenges and opportunities you never realize existed.

A Different Perspective

I found this short comment from the Brisbane, Australia newspaper.  I am not sure when it was published –but it is reasonable to assume this aspect of loyalty is still around.  I doubt this has happened only once – it may be more pervasive than we want to believe.  There is a naïve perspective that perfection of loyalty programs is near – maybe not.

Be a loyal customer, be taken for a fool

I was a passive consumer, believing loyalty is virtuous and appreciated by business. Having chosen products and providers I stick with them, anticipating that long-term allegiances will have their rewards.

I have learnt, however, that a more apt description of my conduct was ''mug punter''.

For more than 15 years I have happily coughed up for cable internet that served its purpose adequately. The cost disappeared painlessly from our bank account each month and we rarely troubled our allotted usage limitation.

Then we did. On the last day of the month our internet connection slowed to a crawl. Thinking it a technical malfunction, I rang a fellow at an Indian help desk only to learn that I had been ''throttled back'' for the day, having finally burst through a 12-gigabyte glass ceiling.

After 15 years of good behavior I thought that I might have earned a little leeway. Why, even the RTA and the courts have a heart. Perhaps a small draw down from the huge bank of unused gigabytes built up over the years.

I started to look around. My limit had become minuscule with the passing of the years and technological advances. Non-customers were being tempted with gigabytes beyond imagining, at breakneck speeds, for a pittance. Whither my perceived reward for loyalty?

I spoke to another friend in India. He immediately more than quadrupled my limit, increased my speed fivefold and knocked $10 a month off my bill.

Rather than feeling a surge of gratitude, I felt cheated. They had been ripping me off, probably for years. I decided to test all the waters in which I swim. My gas and electricity tariffs dropped. My insurance renewal premium quote was instantly reduced. All goodwill I might have felt towards those I once thought of as preferred suppliers evaporated. Worse, I now have ill will towards those who took advantage of my loyalty.

I now realize one can't sit idly by and expect anyone to realize the absolute value of a lifetime customer. So I whine and moan and price check. It's a waste of my time and ultimately an expensive way for business to do business, but at least I know that if it chews up the gigabytes, I only have to blast a call center somewhere to rectify the problem.