Epsilon's Noticeable Results, 1991-1994
This list of noticeable results was collected and amended through the learning history interviewing process. People were asked to comment on items, their significance, and if it was familiar, describe how it was accomplished and what if any role they and others they knew had in it. The items in this list are observable events or objective measures which provide data on Epsilon program progression.
- Mechanical Prototype (MP) build (8/91): The Mechanical Prototype is a
production level prototype for the underbody and front end of the car. The
Epsilon MP design represented a considerable stretch from the previous AutoCo
vehicle; it incorporated multiplex wiring, all new suspension and accommodation
for electronic navigation systems.
In part as a result of earlier delays the MP drawings were sixteen weeks behind, but the first MP build was completed only four weeks behind the original schedule. The quality of the MP prototype build and maturity of its design allowed extensive testing to be done much earlier than is normally possible.
- Team collocation [1] (10/91):
Although the Epsilon team had not been designated to be collocated, program
management pushed for it. The Epsilon team collocated 37 months before Job One
(the date when production manufacturing was set to begin).
- Stage 8 Market research clinic (4/92): Forty engineers from the
development teams participated in a market research clinic in California. This
was said to be the first time engineers formally talked directly to customers
this early in a vehicle program.
- Harmony buck complete (1/93): The harmony buck is a mechanism to
review early designs and design changes prior to the periodic prototype builds.
The harmony buck was an idea proposed by engineers on the Epsilon team.
However, the $2 million cost to build a harmony buck was not covered in the
Epsilon program's budget. Program management supported the concept and lobbied
Senior Management (Vice Presidents) to gain funding support. AutoCo now uses
the harmony buck in other programs.
- Evaluation Prototype (EP) build (4/93): The EP brings all vehicle
systems together, so that integrated testing can occur. The program team
completed the first EP on April 1, 1993, making up for earlier delays and
meeting the original program timing plan. Eighty-five percent of parts were
available for the EP build (setting a company record; other car programs have
had between forty and sixty percent of parts available at this point in their
programs).
- Change Requests (CRs) reach 500 (7/93). Change Requests (CRs) are
documents which engineers write to indicate the need for alterations in parts
or technical specifications. CRs indicate that rework is needed; thus, senior
management uses the count of CRs to evaluate program performance at any moment
in time. Following the EP build, the Epsilon had 524 outstanding CRs,
ordinarily a sign of very poor performance. (A more typical number would be
200.) Product development and manufacturing management said that they had never
seen a program recover from such a high level of CRs.
- Validation Prototype (VP) build (10/93): Validation Prototype vehicles
are built to test changes made after the EP build. The VP design was frozen in
July of 1993 -- three weeks ahead of plan. Ninety-three percent of the VP parts
were on time to the material requirement date (MRD). According to
manufacturing management, the quality of the VP prototypes was the best any
vehicle program had ever accomplished. The subsequent engineering release (ER)
was completed in August of 1993, four weeks ahead of plan. Ninety-eight percent
of the ER parts were delivered one month ahead of plan, with the other two
percent known and accounted for. Four VP prototype vehicles were built on the
regular assembly line at the Mission Hill manufacturing plant.
The new owner vehicle assessment (NOVA) scores for VP were 96, compared to an average of 108 for other vehicle programs (lower scores mean higher quality ratings). The NOVA scores for the earlier build had been substantially worse than average; they were 145, compared to an AutoCo average of 105. Top AutoCo managers made what were described as uncharacteristic acknowledgments that the Epsilon program was performing well.
- Accelerated 1PP build (6/94): The 1PP (final) prototype build began
one week early. The team had 70 percent "production status" parts (normally 50
percent). The new owner vehicle assessment (NOVA) scores for 1PP were 28, a
company record. The previous best NOVA score was 35, and the average score was
55 for other vehicle programs.
- Job One accelerated by one week (11/94): Production builds began one week ahead of the scheduled date. Starting production early was previously unheard of, and thought not feasible given the normal chaos that surrounds a vehicle launch. The program was able to return an estimated $65 million of the $90 million budgeted for late changes to parts. Based on the "18 panel" reports submitted at the end of product development, the Epsilon met or exceeded all forecasted goals (quality, weight, fuel economy, performance, functional image, customer satisfaction, variable costs, investment, and vehicle profitability).
- Final quality results: The final Nova C quality assessment for the
Epsilon was 5.8 -- significantly lower (better) than the average Nova C
assessment for the last six recent launches (which averaged a score of 9).
Subsequent quality rating by an independent market research organizations
(Competitive New Vehicle Quality) showed a 30% improvement in quality as
measured by things gone wrong, rating AutoCo's Epsilon in second place for
automobiles in initial customer quality.
1 "Colocation" means that people from the diverse engineering functions developing the car (i.e. chassis design, air conditioning, suspension, alternator, sound systems, dashboard subsystems, etc.) work from offices in the same location. Team members are physically placed together, during the time it takes to design the car, instead of coming together only for formally scheduled meetings. [Back]
[Back to the Abbreviated Table of Contents]