When Boeing chose to consolidate its five separate Facilities Management groups into one unit?the Facilities Asset Management Organization (FAMO)?it was Lovette's responsibility to quantify and document the resulting improvements.

"When FAMO's new leadership team was formed we determined that, in order to understand the impacts of the consolidation, we needed to adopt standardized performance metrics," Lovette says.

Team First, Then Measurements

The first step in establishing a measurement program was to develop a metrics team that could coordinate operations with district representatives from each of the five regions. The goal was to give local representatives a voice in the process, and gain regional cooperation. Lovette was recruited as the team's statistical consultant.

The metrics team resolved to measure three "macro-processes" within the facilities organization: acquire and modify (building construction/modification), maintaining assets (both buildings and machinery), and transport (in-plant and across plant). Boeing Airplane Group mandated that the team measure improvements under four categories: customer satisfaction, cost efficiency, cycle time, and defects.

"During the consolidation, regional management teams made more than 200 suggestions for specific measurements to study within these categories," Lovette says. "It was a tremendous amount of information to consider. Because five different business units were coming together and everyone had their own terminology, it was very important to get the specific definitions of the measurements down carefully."

The team ran the 200-plus "measurement candidates" through a criteria-matrix to determine which ones would provide the most pertinent feedback, while still being relatively easy to gather using the company's established infrastructure.

The team's primary goal was to formulate a set of metrics that would be functional for use at the vice presidential level. For this reason, the matrix dictated that measurements be straightforward and give a realistic reflection of the macro-processes from beginning to end, rather than from a daily management perspective. They also had to be efficient to collect, and, most importantly, easy to validate.

"Developing this metric system was a constant selling operation at all levels of management so data integrity was highly important," Lovette explains. "If people don't believe the data, they're not going to suit up for any kind of improvement."

The matrix allowed Lovette's team to zero in on 17 measurements regarded as key performance indicators. After the specific measurements were selected, the next step was to develop consistent data collection methods that could be implemented across all five districts.

Education, Access, and Support:
Formula for Success

Extensive, ongoing one-on-one training turned out to be a major component in the team's success. Representatives from FAMO's metrics team went out to each district and instructed management personnel and analysts on how to collect data, read the resulting statistical process control charts, and create a district-level monthly rollup of the information.

"We use statistical process control methodology so there's a fair amount of training that has to take place," Lovette says. "We didn't want to promote a practice where I was making charts then going out and saying 'this is what the data reflect, can you please improve this or that procedure.' We wanted the divisional people to start looking at their own data and making decisions about ways to improve their processes."

Because FAMO consists of 80 service groups spread out across five geographic regions, it can be exceptionally difficult to disseminate information and streamline organization. The team mitigated these difficulties by storing all historical data and charts on a server connected to Boeing's intranet. The centralized location makes the information readily accessible and allows the metrics team to provide remote support for training and troubleshooting. The team also publishes its current metric information on the corporate intranet.

Case Study: Measuring Results

Implementing macro-process metrics is an ever-evolving process. As an example, Lovette presents a metrics case study in the Acquire and Modify category that analyzes the delivery of Triage-Zero (T-0) services, projects involving less than eight hours of craft time that do not require permits or design?such as supplying a whiteboard to a cubicle or office. Within FAMO, the target time for delivery of a T-0 service is five days from the time the request is submitted until the day it is fulfilled.

"Initially, when we first started collecting measurements, the overall cycle times improved because people were starting to understand that this metric was important. After the initial reductions leveled off, we could see there was still a lot of variation in cycle times between districts. Upon investigation, we found that some districts had policies and procedures in place that required long lead times, like ordering parts, or obtaining lots of signatures, and so forth. Using metrics allowed us to uncover these factors and adjust them to improve the process," Lovette says.

These adjustments lowered the average cycle time from 140 days to 48. Next, to continue driving down the range, Lovette's team focused on "cleaning up" the data.

"We found that in most cases the projects had been completed without being closed out of the database system." Lovette says. "When they were finally closed out weeks later it bumped the cycle times way up. Once we got project managers to consistently close out jobs when they were finished, the average cycle time dropped to 30 days."

The team also went through specific one-on-one training with project managers to make sure they understood the definition of T-0, and knew how to change the triage level of a project if it turned out to be more than just ordering a whiteboard. These and other system improvements, such as training service group managers in root cause analysis, brought the average T-0 cycle time down to 13 days, with some districts at or under the five-day target.

The data were analyzed with the use of X-Bar-R statistical process control charts. The graphic depiction allowed the weekly average cycle time (in days) to be viewed with the X-Bar chart alongside the range of variation (maximum cycle time days minus minimum cycle time days) of the weekly data on the corresponding R chart. This helped the team begin process improvement by reducing the cycle time variation to zero so that when process improvements were implemented, a clear movement in cycle time could be seen.

Management training, clarity of terminology, and database cleanup proved to be key factors in a second case study as well. In this instance, the team analyzed the scheduling of equipment maintenance, specifically measuring the percentage of planned jobs that were completed on or before the scheduled due date.

After the team went through the process of cleaning up data and claifying terminology, the average was right at 50 percent. Then the Maintain Asset Process Team implemented a "plan of the day" to improve the completion rate, holding a daily meeting with the maintenance personnel service group manager during the first 30 minutes of each shift to discuss the status of jobs scheduled to be completed that week. These jobs are then slotted into the "plan of the week." Once these programs were established, the timely completion of planned jobs rose to 80 percent

Lessons Learned:
It's all in the System

According to Lovette, solidifying upper-management commitment, allowing plenty of time for progress, and properly training all parties involved are essential steps for successfully implementing a metric analysis system. It's also important that, to ensure data accuracy and collection efficiency, companies use their existing infrastructure.

"There's a lot of resistance when you're trying to create a system like this, so system efficiency and management buy-in are essential," she says. "When we were determining what our measures were going to be, we had managers and analysts write comments or changes on the charts themselves and sign off on them. I like to have people sign things because signatures go a long way when you're out trying to get other people in the company to buy-in."

Lovette also recommends that "while it's important to have an expert statistician involved behind the scenes, those employees responsible for the results should be the ones presenting the data. Because, even with a total management buy-in, the people actually involved in generating the data on a daily basis have to understand why they're doing it. If they don't, you won't get either the data integrity or the results you need."

By Johnathon Allen