Tom’s Ten Data Tips – February 2007
Customer Satisfaction & Loyalty
Customer Satisfaction is an important driver of business success because it embodies value creation for the customer. The assumption is that satisfaction results in repeat business, and as positive experiences accumulate, the relation will strengthen. This will immunize customers from alternative offers, escaping a competition on price.
When you measure satisfaction do just that. Quality perceptions, recommending, trust and product usage, are something different. Loyalty is another such term that is used ambiguously. They mean different things in different settings, and hence require context relevant measurement. If you confuse these concepts you are bound to misguide your strategic initiatives.
1. The Meaning Of “Satisfaction” Depends On The Context
Satisfaction is context dependent. Satisfaction with a retail bank differs from satisfaction with dining in a restaurant or satisfaction with the dentist. This warrants the development of context specific definitions of satisfaction, and corresponding measurement procedures.
If you want a valid measure of satisfaction, you have to cater the measuring instrument to the specific context. Psychometric research has conclusively demonstrated that there can not be one universal satisfaction instrument. “Generic” satisfaction measurements are consistent in one respect: they do a poor job.
2. Customer Loyalty Can Not Be Measured With One Single Item
Despite what Bain & Co and Frederick Reichheld have been claiming recently, it is utter non-sense that customer loyalty can be measured through one single question: “Would you recommend our services to your friends and colleagues?”
The reason for this is that customer loyalty is a psychological construct that refers to a domain of behaviors. One single instance (recommending to friends) within that domain can hardly be an accurate measurement. Messick wrote as early as 1989 that “single items yield moderate measurements of constructs because they almost certainly reflect a confounding of multiple determinants.”
Apart from this “technical” default, recommending is associated with various other traits, and therefore reflects loyalty very poorly (as we have also found in proprietary research).
3. Expectation (Dis-) Confirmation Does Not Work Well As A Measure Of Satisfaction
Following the Service Quality research strand, some practitioners consider the comparison of expectation with performance as satisfaction. This typically leads to poor measures.
There is no consensus whether to measure performance and expectation separately, or combined in one single item (although probably the latter is preferable). But what is worse, which expectation exactly should one ask for: the ‘perfect’ or ‘true’ expectation. The latter is grossly influenced by the context. Instead, satisfaction should be measured on the basis of items reflecting different aspects of overall satisfaction with the company.
4. Satisfaction With A Company Is Not The Same As Satisfaction With Its Services Or Products
Satisfaction can be defined and measured at varying levels of aggregation. It is adamant to distinguish between overall satisfaction with the company and satisfaction with the products or services. Consequently, the measurement of overall satisfaction differs from the measurement of satisfaction with some product or service.
By analogy, there is a distinction between transaction specific satisfaction and summary satisfaction. Whereas transaction specific satisfaction is based upon a single service encounter, summary satisfaction is based upon the accumulated encounters.
5. Customer Loyalty Is Repeat Business Plus “Something”
There is no universal agreement as to what constitutes “customer loyalty”, nor how best to measure it. This is because “being loyal” means different things in different businesses. It is clear though, that “merely” repeat behavior is not enough to qualify as loyalty (although it is quite often measured that way). To distinguish inertia from true loyalty, the measurement needs to take this into account.
Loyal customers are not only repeat customers but also:
- Display empathy: the
reciprocity in the relation makes customers also guard the supplier’s best
interest
- Give unconditional
credit: even if something goes wrong in the service delivery, the relation
is not immediately questioned
6. “Average” Satisfaction And Loyalty Are Inflated
It is striking that when satisfaction is measured, the typical outcome tends to be 7 on a scale from 1-10. How good or bad is this? We typically do not observe much variance, which makes any such score even less informative.
Because satisfaction and loyalty are measured among customers, your typical sample will contain few unhappy customers because they have already left. Such censored distributions can be deceptive.
Benchmarking may help, but valid sampling is a tremendous challenge. How subjects are invited to participate, wording of items, method of data collection (paper/internet), all influence the outcomes.
7. Don’t Merely Measure Satisfaction – Include Its Drivers
A persistent problem with many reliable and valid instruments to measure satisfaction is that the company is stuck afterwards with the feeling: “now what?”. In itself it is merely nice to know if a good or reasonable score was obtained.
Client organizations need to know what they can do to improve satisfaction, and preferably get an indication which drivers are most important to elevate satisfaction. Too many surveys fall short here. This emphasis on actionable outcomes rarely works as an afterthought, but instead needs to be woven in from the outset (when you design the questionnaire). The assessment of drivers is far from trivial. What drivers should be included in the survey? What measurement model fits the data? Complaint registration may provide valuable insights, too, here.
8. Low-Involvement And High-Involvement Products Have Different Satisfaction Drivers
For low-involvement products (which are more common than product managers are tempted to believe J), satisfaction is often the absence of dissatisfaction. One needs to attend to hygiene factors and accept that satisfaction is in fact a pre-conscious state involving little if any cognitions. Consequently, providing information may have little impact.
High-involvement products have a far greater likelihood of getting satisfaction evaluations on the basis of cognitive and emotional elaboration. Communication to set proper expectations can have a profound effect here. An example might be an “apologies for the inconvenience while we renovate” sign in a hotel, or notification about delays in a travel schedule.
9. Manage Satisfaction, Loyalty And Retention In Conjunction
Loyalty is more than merely repeat behavior, which should be more appropriately labeled “inertia”. Loyalty is a necessary but not sufficient condition for retention. But the soft, “attitudinal” component in loyalty is hard and impractical to measure for each and every customer on a repetitive basis. You manage retention by using behavioral drivers to trigger marketing actions.
Satisfaction plays a key intermediating role because at the customer level dissatisfaction precedes disloyalty and switching behavior. All this illustrates why these KPI’s are best placed under central management to ensure alignment across business functions.
10. Customer Satisfaction Is A Requirement For Sustainable Success
There is more to running a profitable business than “merely” making customers happy. If that were so, why not simply give all products away?
However, without satisfied customers, retaining customers will be very difficult. One simply cannot “buy” loyalty, so the result is bound to be a continuous throughput of customers. Customer acquisition then needs to make up for churning customers. It is fundamentally impossible to keep such a “model for success” up indefinitely: at some point there will be no “fresh” customers left in the marketplace to disappoint.
Special thanks and attribution to Maarten Terpstra for providing
invaluable feedback and comments. Maarten Terpstra’s draft thesis and suggestions were instrumental to this issue of Tom’s Ten Data Tips.
Further reading
Some excellent books on Customer Satisfaction and Loyalty:
Jacques Horovitz (2000) Seven Secrets of Service Strategy. ISBN# 0273635778 
Giese & Cote (2000) Defining Customer Satisfaction,
Richard Oliver (1996) Satisfaction: A Behavioral Perspective on the Consumer. ISBN# 0070480257