Continue in 2 seconds

Customer Service as Marketing through Statistical Modeling

  • November 14 2006, 1:00am EST

Often an automotive manufacturer's sales, service and parts functions appear to be polar opposites. The corporate managers responsible for vehicle sales oversee a revenue source with a high per-unit profit. The managers responsible for service and parts oversee a cost center with a low per‑unit profit.

However, service and parts can have a significant impact on sales. From the time of an initial vehicle sale to a customer's next sale, service and parts exerts a zone of influence. Deliver poor service, and the probability of the next sale decreases, regardless of what country singer is singing about your latest line of pickup trucks. Deliver excellent service, and the probability of the next sale increases. This zone of influence also has a viral marketing effect: excellent service creates positive word of mouth, creating the potential for exponential growth in an automotive brand's influence in the market as a reliable vehicle.

Automotive manufacturers and suppliers have worked extensively to streamline product design and manufacturing. Highly efficient supplier partnerships and manufacturing operations are no longer a competitive advantage but a minimum requirement to compete in the automotive industry. In numerous ways, automotive executives and managers of dealership operations are just beginning to apply similar efforts to dealer channel performance.

In the manufacturing industry, Six Sigma and just-in-time (JIT) production are two well-developed concepts that improve quality (Six Sigma) and reduce costs (JIT). Six Sigma is a measurement of process quality using statistical procedures to continually improve manufacturing processes. Both Six Sigma and JIT continue to significantly impact the automotive industry and can be applied to dealership customer service.

Six Sigma improves the quality of each product that rolls off the manufacturing line and reduces variability in quality from one manufactured product to the next. A Six Sigma for dealer operations would improve the customer service of each dealership in the channel in a way that positively impacts business results. To achieve this positive impact, any statistical performance initiative must reduce customer service variability between one dealership and another. Good to excellent service must be delivered again and again. As soon as a customer has a bad customer service experience, the seeds of doubt are planted. The probability of the next sale decreases, erasing gains made from previous positive experiences.

Customers will form their customer service opinion around the product or brand - not necessarily a specific dealership's service department. This dynamic underscores the importance of maintaining good performance across all dealerships in the channel. Furthermore, the manufacturer must account for the entire customer service picture, which may extend beyond the service department to sales, financing and other departments.

An understanding of a dealer channel's causal networks provides the necessary and often missing basis to apply statistical process control to continually reduce variation and improve dealership processes. Often, people make inferences about simple causal relationships, focusing on a single cause and effect, for example, how reducing the number visits required to diagnose and complete a repair may affect customer satisfaction. A causal network represents a more sophisticated set of relationships, providing deeper understanding and control with the ability to predict the consequences of actions that have not yet been performed.

Structural equations modeling (SEM) provides a statistical method to develop a causal network of exceptional service and quantify how each relevant variable affects service. It is an advanced statistical technique to study the simultaneous impact of several independent variables on a specific outcome variable. Each outcome variable of significant interest would have its own model. For example, if the goal of dealer operations is improving customer service, a model might be created for the customer satisfaction index outcome variable.

The way structural modeling works is as follows:

  • Set initial model according to expert judgment.
  • Collect data for all variables. Data may already exist in company databases or may be commercially available.
  • Run the statistical software to estimate the model parameters.
  • Review and revise the model according to data results.

The resulting output of SEM - a path diagram showing which variables cause changes in other variables - could represent a dealer channel causal network that describes what interrelated factors affect customer service. Not only do dealer operations managers get an idea of what variables directly or indirectly affect customer service, they also get coefficient values that quantify how the variables affect one another.
For example, an initial model for automotive customer satisfaction may include a wide range of possible relevant data values. Large amounts of data already can be collected from third-party dealer management systems and evaluation systems used by an automotive manufacturer's field organization to score dealerships on a wide range of metrics. Using statistical software, the initial model is tested. Some variables will have an impact on the customer satisfaction outcome and others will not. Variables that appear relevant are retained and the model parameters are run again.

In one study performed for an automotive manufacturer, the first fixed visit (FFV) was an outcome variable used to measure the effectiveness of parts and service. This variable measured how often a vehicle was repaired correctly after it was first serviced. In the resulting causal network, ratings for the technical infrastructure, the service manager, and the dealership itself directly impacted FFV. Each of these variables was affected by additional unknowns. The variables affecting the service manager attribute may include specialized training that the manager has taken, the employee satisfaction and the manager's relationship with the parts department. The model can be refined further with multiple causal networks. For example, an SEM could be run on employee satisfaction to see what variables can be indirectly leveraged to boost the service manager rating and affect FFV.

Causal networks built by SEM can provide corporate and dealership managers with predictive models that can help them optimize resource allocation and maximize service and parts outcome variables. One approach would be to create incentive systems and certification programs that encouraged all dealerships (reducing variability across the dealer channel) to allocate resources and effort in those areas that will positively affect customer service. For example, in the previously mentioned example, a certification program may be put in place that requires the specialized training for service managers that affects FFV. The dealership owner may be given incentives to participate in the program.

The causal network may also be applied to a Web-based scorecard that can be accessed by the dealers and the automotive manufacturer's corporate staff. By compiling performance metrics and subsequently linking causes and outcomes, the scorecard not only aggregates performance metrics, but also provides steps that dealers can follow to improve weak scores. In the end, the customer receives better service and is more likely to make that repeat vehicle purchase.

Register or login for access to this item and much more

All Information Management content is archived after seven days.

Community members receive:
  • All recent and archived articles
  • Conference offers and updates
  • A full menu of enewsletter options
  • Web seminars, white papers, ebooks

Don't have an account? Register for Free Unlimited Access