Competency Model Construction, New

Wednesday, March 24, 2010

 

There are two ways to approach the creation of almost anything, build new from the foundation up or purchase something already created and modify it to meet your needs. This is true for houses, organizations, software systems, and competency models. What would be required to build a new fully functional, validated competency model for, let us say for a long distance truck driver? We would need to know what a long distance truck driver does, when, where, how, why, and who does the work. Furthermore, we would want to know the frequency of the work and the importance of the work to the overall success of driving a truck. As a designer, we would want to talk to drivers, their managers, their mechanics, and others. A good idea may be to ride along with drivers and observe them in various situations, times of day, locations, and weather conditions. Lastly, all of this “data” would require analysis and formulation into an initial and tentative model, with preliminary testing for reliability and validation with subject matter experts.

 

Using the example of a long distance truck driver we would collect data on the type of truck, engine and transmission specifications, trailer type, number of drive axles and non-drive axles, mileage driven, tonnage delivered, geographical area driven, type of driver assignment, i.e., single or team, typical weather and traffic conditions, number of stops or trailer drops, hazardous vs. non-hazardous cargos, safety records, and types of cargo. It is possible that much of this data is available from the organization’s data management system(s).

 

Interviewing drivers could prove to be challenging as many may be on the road and inaccessible for long periods. Some type of online (many drivers now file their logs electronically) questionnaires could prove to be helpful when interviewees are not housed in a central location. Interviewing drivers after returning to their home depot might be an alternative, if you allow for the drivers to rest after driving 5,000 miles in five days. As with any job, managers would be a key subject for interviewing. Mechanics could be useful in identifying the types of drivers who take better care of their trucks, i.e., lower maintenance cost, down times, repairs … etc. Interviewing hundreds of drivers could be time communing and expensive, requiring a team of interviewers. The very nature of an interview could alter the behaviors of the drivers, for the better or for the worst. The option of hosting focus groups with drivers could be as much of a challenge as interviewing individual drivers. Job diaries and journaling might be an alternative means to collect specific details about examples of certain events, traffic, accidents, weather … etc.

 

Riding “shot-gun” with drivers could certainly add a personal perspective to the boredom, lonely hours, and monotonous conditions of driving thousands of miles each trip. It would further help to understand the issues associated with traffic stops, weight stations, inspections, accidents, city driving vs. interstate, weather conditions, trailer drops, cargo loading and unloading wait times. However, this could also alter the driver’s behavior. The use of remote video recordings could support the observation of drivers at their home depots as they enter, prepare for, and leave for another run.

 

Once all these data points and observations have been collected and organized, some form of analysis will follow to identify and separate outstanding, average, and poor performers. Following the initial data collection, organization, and initially analysis, preliminary validation of the data will require review by managers and other subject matter experts. “Others” might include individuals such as, driver recruiters, mechanics, diver trainers, dispatchers, depot managers, material handlers, forklift operators, and other drivers. The goal, at this point is to gain a consensus among the various groups as to what constitutes outstanding, average, and poor performance of a driver. With this consensus, we should have enough tentative information to construct a model of driver behaviors to support selecting, training, evaluating, rewarding, and managing drivers.