In markets where manufacturers use dealers to sell their products, it is important for the manufacturer to strive for a good relationship with its dealers. Not only because a good relationship within the business chain is desirable for a smooth cooperation, but also because the brand retention decision of the consumer may depend on the dealers’ performance. The dealers’ performance is specifically relevant when it is expected that dealers add substantial value to the brand they sell. An example is the new car market, where the relationship between the consumer and the dealer might decide whether or not the consumer stays loyal to the dealer, and thus loyal to the car brand. In any market where the service-experience is an important aspect of the perceived product quality, the performance of the manufacturer and its dealers are interconnected. Other examples are internet connections, electronics, and durables, like machines, boats, and cranes.

This article provides guidelines to create a framework to improve the performance of the dealer and the manufacturer, based on close cooperation and data sharing.

A framework for gathering useful insights
The framework helps with creating new propositions, adjusting strategies and segmenting different customer groups. It is executed by first inventorying all variables and determine whether or not there is cohesion between the variables. Secondly, the framework examines and evaluates the detected impact of each relationship. Finally, the results are interpreted and conclusions are drawn.

The framework is currently used by The Next Organization in a case about luxury car dealers. In this case, the effects of dealer services on the customer satisfaction and the retention rate of the dealers are analysed. In other words: how do different dealer services affect the customers’ choice of whether or not they stay loyal to their automotive service & repair provider. The case is used throughout this article to function as an example to clarify abstract components.

To make use of the framework, a straightforward step-by-step approach is to be followed.

Step 1: set a clear, measurable goal
Determine the goal of the project. Questions to be answered are:

1. What needs to be achieved with the project?
2. Which variable to use?
3. How to measure the variable (which scales and dimensions to use)?

What needs to be achieved with the project?
Defining a clear goal helps the organisation to maintain the focus, prevents complicating the process and helps to determine the further compositing of the project. Therefore, it is important to set the goal beforehand and make it specific. In the case of the car dealers, the manufacturer wants to improve the service retention rate of the dealers, by improving the service package of the dealers. With the knowledge derived from the research, dealers create tailor-fit service packages for themselves, based on their current customer base.

Which variable to use & how to measure the variable?
Defining the goal often automatically determines which variable is used to measure the outcome of the research. This variable is called the “outcome variable”. Frequently used goals are improving customer satisfaction or a higher Net Promoter Score rating by the customers. The outcome variables of these goals would be customer satisfaction and the Net Promoter Score.

In the case concerning the car dealers, the goal is to examine which dealer services affect the service retention rate (and whether or not there are differences among customer segments). This makes the service retention rate the outcome variable.

Service retention rate is an objective construct, measured in percentages from 0% to 100%. This automatically answers the last question on how to measure the outcome variable. However, when measuring more abstract constructs like perceived service quality or price fairness perception copious techniques, scales, dimensions, and combinations are possible and available. (For instance, the quality of a service is the customer’s expectations of the service, minus the customer’s perception of the received service. This basically means that the quality of a service is subjective and open to interpretation. Luckily, social scientists developed user-friendly scales to make such abstract constructs measurable).

Step 2: determine the input variables
Input variables are factors which might affect your outcome variable. It is important that these input variables are measurable and selected carefully. This step in the approach explains three criteria to make the selection.

1) The input variable must be manipulable by the organisation conducting the research
A good example in the car dealer case would be to use price fairness perception of the customer. This variable can be manipulated by the dealers via the price and by influencing the customers’ perception of the price used (e.g. giving free merchandise, advertise with view small services for very low prices or avoid round prices and end with .99). Being able to manipulate the variable is important since it is pointless to examine the effect of a variable which the organisation cannot change. For instance, it would be fruitless to examine the effect of gas prices on the retention rate of car dealers. Is it interesting? Maybe. But neither the dealers nor the manufacturer can change the gas prices. Therefore, it is not a good input variable for that research.

2) There should be differences in the values of the variables
A difference between the values of the variables is crucial. If there is none, there is nothing to compare and nothing to measure. A good example would be examining variations in the offered services among the dealers. These variations might explain the variance in the performances of the dealers.

A bad example would be examining a service which all dealers offer. Since there is no variation in the offering of a service, this cannot explain the variance in retention rates (e.g. if all dealers offer a replacement vehicle to their customers, there is no point in measuring this service, since this service will explain zero variance in the retention rate among the dealers). In most cases, it is not difficult to avoid such a situation. By discussing the input variables with the parties involved, fruitless variables are identified easily and can be deleted from the analysis.

3) Use logic reasoning
Simple logic will often help in determining the input variables. There is always a chance that variables correlate, despite the lack of logic and causality. This is because most statistical software cannot reason with logic like a human being can. It will simply report a relationship when it finds a correlation between variables. A good example of a study where logic and causality were clearly overlooked is a study by the University of Alberta in 2016. According to this study, one glass of red wine affects physical performance and muscle strength in the same way as going to the gym for one hour. Of course, it is possible that there are similarities in physical performance and muscle strength between people that go to the gym for one hour and people that drink one glass of red wine. However, that does not imply that drinking one glass of red wine has the same effect on your body as one hour at the gym.

It is important to keep in mind that correlation and causality are very different constructs and that they should be interpreted differently. Hence, it is important to reason whether a possible link is logical or not. A good question to remember is: is it logical that the one event is the result of the occurrence of the other event?

Another example of a study in which logic and causality is clearly overlooked is: Late-Night Snacking May Have A Surprising Effect On Your Memory

Step 3: gather the data
There is a handful of methods to gather the data. The type of data needed determines which gathering method is most appropriate. However, in most business settings, conducting a survey among consumers is a suitable manner. This is the step where the cooperation between the dealer and the manufacturer becomes crucial.

Dealers have much direct customer contact and can easily ask the customer to (for example) fill-in an (online) survey after engaging in a customer touchpoint. But, on the other hand, this gives dealers the opportunity to influence the results dramatically, whether this is on purpose or not. In the automotive industry, many dealers receive bonuses and privileges from manufacturers when they score high on predetermined performance metrics. This might tempt a dealer into purposely influencing the customers’ rating, by rewarding the customer when it gives an excellent rating on the predetermined performance metrics. To overcome this, the manufacturer should brief its dealers beforehand. Explain the goal of the research and how it can help them in improving their performance. This also stimulates the feeling of involvement and might improve the response rate of the questionnaire.

In most cases, the use of a survey is a sufficient method to gather consumer data. However, when needing data on customer behaviour, (e.g. behaviour in a showroom or on the website, the route a customer walks in a store, or how customers react to certain employee behaviour), other methods, like observation or experiments, are more suitable.

Step 4: running analyses and interpret the results
After running the analyses and interpreting the results, it is time to formulate new propositions and strategies. It is advisable to formulate and discuss the new propositions and strategies with managers from the dealer network. Combing the new insights with the practical knowledge and work floor experience of dealer managers will create a strong combination of theoretical- and practical knowledge. Ideas derived from this team are often effective and easier to implement. The involved dealer managers function as representatives of the dealers. The involvement of these representatives will lead to less resistance of the dealers, and a faster, easier acceptance of new propositions and strategies. This is important since motivated dealers will execute the new propositions and strategies better then unmotivated dealers.
Finally, after implementing the new strategy or propositions, it is wise to measure whether or not the changes have the expected effect.

Making it tailor-fit
Dividing the respondents into customer groups, based on demographic, behavioural or personal variables, provides the opportunity to split the data and detect possible differences among customer groups. This might sound abstract, but an example will clarify the principle. The survey used in the car dealers case contains questions which ask the customer to rate their service encounter on several service aspects. But additionally, the customers are asked to answer questions about themselves, their car, personal interests, and needs. This enables the software program to split the data into customer groups based on these customer interests, needs, demographics and many more. As a result, it is easy to detect differences among customer groups and create a tailor-fit service package for each customer group. For instance, owners of a certain brand, with a lower average age, expect the dealer to be reachable by using an app, while other, older customers ask no such thing. It is highly recommended to use this technique since it will create more depth in the analysis.

Summary
Set clear and measurable goals, communicate these goals with the parties involved, determine the variables carefully and translate the results into usable insights. This will enhance the performance of the dealer network and subsequently, increase the performance of the manufacturer.