Sie sind auf Seite 1von 2

Sigma Brew Coffee : Analyze Phase

Summary:
It is the phase in which the statistical analysis of a problem starts. During this phase, statistical analyses are
performed on the variation groups to help determine the project owners who are the major contributors to
the output. This phase focuses on recognizing and analyzing the root causes of imperfection. The phase
typically starts by arranging the schedule of experiments to determine which of the possible KPIV's (key
performance indicator variables) are the main source of the defects. There also needs to be a sufficient
amount of data. This collection of data is analyzed using an appropriate statistical method. The goal of this
step is to determine the KPIVs which makes the main effect on "Y" also known as output. Any misuse
noticed for change should be reported and given priority.

Tools:

Create VA Flow Charts by Product


Anything that does not add value to the process and consumes time, space or costs money can be termed as
waste. In the analyze phase, the first thing that needs to be worked on is to identify waste also called Muda
(Japanese word for waste) in lean culture and eliminate or minimize it. To analyze waste, we purchased
Value Added charts for all 20 products. One thing that needs to be addressed immediately is that there has
been approximately 21% decrease in the value-added ratio of new products (42%) compared to old products
(53%). It was also observed that writing orders on a cup and walking up to the counter takes ~10s on
average which is common for all products. The manual handwritten activity can be replaced with pre-printed
stickers and orders can be served via conveyor belt thus eliminating motion created by the movement of
people and equipment. However, in addition to these activities, it is important to notice that walking to
retrieve different items contributes to a major proportion of the non-value-added time.

Analyze Service Time - Multiple Regression


After analyzing waste, we tried to understand which inputs (activities) are significant for predicting service
time by purchasing Analyze Service Time – Multiple Regression tool. This data represents a series of
potential inputs collected by on-site store teams during a busy period (6:00 am to 9:00 am) across multiple
stores. Using multiple regression analysis, it was observed that the dependent variables like the number of
drinks ordered during a busy period, number of builder employees used by the store and nu mber of espresso
makers are statistically significant with respect to independent variable service time. Using tollgate as a
reference point, it shows that few more tests such as t-test, simple linear regression, time study, and 2-
proportions were performed to identify additional significant factors. The results of these tests show that the
complexity of drink, the complexity of sub-inputs - work time, walk time and remake rate have a significant
effect on service time.

Detailed Demand Analysis


In the previous Measure phase, we analyzed average customer demand data across multiple locations for a
single day. But to accurately measure Takt time i.e. time taken to start preparing for the next order once the
previous order is completed, we need to consider hourly data of a single location. Thus, we purchased a
Detailed Demand Analysis tool that represents hourly data of the 28 days to understand day to day, day of
the week, and hourly variation in customer demand. It was observed that the takt time for walk-up customer
orders is 82 seconds i.e. 44 customers per hour and for builders, it is 48 seconds i.e. 75 customers per hour.
The highest number of drinks are served during a busy period, 6-9 AM. During this period, the largest delay
and queue results due to drinks that are yet to be built. Thus, we can say that the Non-Value-Added and
Value-Added time required to build these drinks exceeds takt time.
!1
Analyze Remakes vs. Product - 30 Day Sample
In the previous phase, we analyzed the remake of product type for a single day at a single location. But it is
important to analyze a larger sample size over time to understand the number of remakes stratified by
product, year introduced, and by espresso vs. non-espresso drink. Thus, we purchased Analyse Remakes vs.
Product – 30 Day Sample tool. This data represents drinks sold and no. of remakes for a period of 30-days
across several SigmaBrew locations. It is observed that complex drinks like frozen and espresso contribute
to the majority of remakes and the mean remake rate is 7.6% which results in long service time.
Additionally, using tollgate as a reference point, a two proportion test states that simple drinks (3.9%) have
lower remake rates compared to the complex drinks (12.9%) by rejecting the null hypothesis (both drinks
have the same remake rate) with 95% confidence level.

Results :

Conclusion : The team has gained valuable knowledge from the Analyze phase of the Sigma Brew Coffee
simulation. We got a thorough understanding of the tools and Service time was found out to be the key issue
with potential X being Demand, Build time and Remake rate. In this phase the team found bottlenecks and
failure modes. The tools helped us understand and answer questions to a great extent. The phase was
challenging and exciting as the team looks forward to the Improve phase.

Initial plan for next phase: The objectives for the next phase would be to identify and evaluate potential
solutions for the bottlenecks identified. The team would identify systemic indirect effects and unintended
consequences of improvement solutions while working towards implementing short term countermeasures.
The team would also establish operating tolerances for the new processes.

!2

Das könnte Ihnen auch gefallen