Best PracticesDealership Operations & Processes

How Weather Impacts Your Dealer’s Service Repair Orders

Every fixed operations department across north America thinks that weather has an impact on the number of completed repair orders (ROs) each day. Every fixed operations department across North America thinks that certain days of the week are slower (i.e., fewer completed ROs) than other days. We’ve uncovered some unexpected results.

More than 100,000 service records from 2015 and 2016 were used to construct 12,000 data points for analysis in this study. Various locations from the U.S. were used to account for differences in regional reactions of the same weather pattern.

For example:
People in Chicago, Illinois may react differently to 3 inches of snow than people in Atlanta, Georgia react to 3 inches of snow.

Since one repair order (RO) at a small shop accounts for a larger percentage of the total than one RO in a very large service department, the data were normalized to account for this variation in service drive size. After standardization, the RO counts per day were expressed as a number between 0 and 100.

Here is the breakdown: 50 = Average, >50 = High Level

The daily RO count was analyzed using the day of the week as our independent variable. This article will refer to this as weekday factor.

Figure 1 is the boxplot for analyzing weekday factor. It shows that weekday factor has a significant effect on the number of ROs. Monday has the highest number of ROs compared with all other days of the week while Saturday has the lowest.

Figure 1

WHY?

  1. More research is needed to fully understand this significant difference, but we have a few theories:
  2. Fixed operations departments are usually open for a shorter period of time on Saturdays (i.e. 8 hours instead of 10).
  3. Monday RO count is higher perhaps due to procrastination of service customers during the weekend.
  4. People often desire to spend days off (Saturday and Sunday) with family and friends. Folks often need their car for personal events, errands, or travel.

For some, it may be more desirable to take time away from work on Monday rather than one’s own personal time on Saturday.

“Actually, the effect of weekday is so overwhelming that in the result of regression tree-based machine learning method for predicting number of ROs, weekday has twice importance than the second most important factor, as shown in Figure 2.”

Figure 2

What does this mean?

Predictability is important in any statistical analysis. In fact, one might say determining predictability is the reason for analyzing any data set. If a variable is more predictable it typically means it has a larger and more consistent effect on another variable. In this case, we compared ROs to the following:

  1. Weekday
  2. Sea Level Pressure
  3. Humidity
  4. Dew Point
  5. Temperature (Fahrenheit)
  6. Wind Speed (MPH)
  7. Precipitation (inches)
  8. Visibility (Miles)
  9. Normal (No measurable events)
  10. Rain
  11. Thunderstorm
  12. Snow

Weekday factor has DOUBLE the influence as the second most influential factor (Sea Level Pressure) in predicting RO count.

This tree (Fig. 3) uses a significance test procedure to select variables. These significance tests are computed at the beginning of each algorithm. We calculated all possible values of the test statistic under rearrangements of the labels on the observed data points.

 

Figure 3

What does this mean?

Each variable was compared to every other variable in every possible arrangement. After this, it was determined that Saturday is, by far, the most predictable and influential factor including all weather events and other days of the week.

Speaking of weather events, they were analyzed separately from weekday factor to isolate weather events from the powerful influence of weekday factor. Examining t-test results, it seems like rain and fog do not impact the number of ROs, but snow and thunderstorms do have a noticeable impact on RO count. This is a general observation using all data, but does this same pattern persist when we compare different regions in the U.S.?

 

Figure 4

Analysis of dealerships in North Carolina revealed that rain has a statistically significant impact on RO count. When it rains, dealerships should expect 0.35-0.45% decrease in number of completed ROs. It is worth mentioning that snow does not have a significant impact on RO count for dealerships in NC.

 

Figure 5

Figure 5 shows the scatter plot of number of ROs spread over temperature and humidity and grouped by whether it’s raining or not. We can observe that if there is no rain, there seems to be more percentage of data points with high level number of ROs.

Figure 6

Analysis of dealerships in Chicago showed that snow, not rain, negatively impacted the number of ROs. In fact, when there is a snow event (snow accumulation), we expect to see a 4.5-8.8% decrease in RO count!

Moreover, Figure 6 shows that snow doesn’t have significant impact on number of ROs for them.

Figure 7

This analysis was performed at 3 other locations and the results can be seen in Figure 8. Upon closer examination, IL and OK have more frequent days of snowfall than the other 3 states. Therefore, snow impacts daily activities more frequently in these states than in the other 3 states. Likewise, GA, AL, and NC have more frequent days of rain compared with the other states.

Since these rain events occur more frequently in these southwestern states, the effects of the rain occur more commonly, and therefore rain’s impact on RO count in these states is more far-reaching and impactful.

Figure 8

In conclusion, there are two noteworthy findings:

  1. Weekday has a profound impact on the number of closed repair orders.
  2. Weather events have a different impact on service shops in different locations.

So how much does weather impact your dealership?

It depends on what day of the week it happens. If the forecast calls for bad weather weather on a normally busy weekday you can expect a relatively normal day. Also, different parts of the country respond differently to adverse weather.

Jon leads Driven Data’s vision, strategy, and growth, providing dealerships an analytics platform that serves as a long-term strategic asset. He ha...