stats 3

Home

Six Sigma
Way

BI
Quality

BI Capability
Maturity Model

BI Comprehensive
Improvement Model

Six Sigma BI Statistics

Should your goal be Six Sigma?

Selecting Six Sigma as the goal in the original systems work was based primarily on the complexity of the products being measured. Six Sigma was used as a measurement of each characteristic within the whole and only a very tight goal for each of these myriad components would allow for the final product to be acceptable. The Business Intelligence Asset Base is also a very complex system and it is recommended that you use Six Sigma as the goal for each component. The Six Sigma methodology for improving business intelligence includes the monitoring of the Y Critical to Quality factors, i.e. the desired results, along with the analysis of the causal characteristics, i.e. the X factors, using a business intelligence audit and tracking system. The first level, the quality, is based on metrics related to the business information product and interface/interaction with the customer. These measure business information quality, provide tracking, monitoring and triggering for the second level analysis of X, i.e. the causal and impacting, factors. The second level of metrics involve identifying and analyzing the causal factors and includes an audit of the product, as well as, the underlying infrastructures, software, hardware, systems and processes, i.e. the business intelligence asset base. This business intelligence asset base is extremely complex and there are 100’s of characteristics and impacting factors, i.e. X factors.

Costs vs Reward: Even so, you may find that a product which is of good to excellent quality and acceptable for the customer can be satisfied with a goal less than Six Sigma. The first level of metrics for six sigma business intelligence , i.e. the CTQ characteristics for the customers, are the Y factors. A goal of four or five sigma may be sufficient to satisfy your needs at this level. These Y factors result from the cumulative effects of the X characteristics of the business intelligence asset base. In order to reach the standards of 4 or 5 Sigma, the specifications for the X characteristics must necessarily be set higher. You will need to judge the effects and the specific requirements for your organization. Keep in mind, also, there are some strong psychological reasons to set Six Sigma level goals.

In any case, if you elect to set a Six Sigma goal, then a case must be made to pay the cost and take the risks which are associated with setting the specifications at that Six Sigma level. It is extremely costly to provide for the metrics, the data collection, tracking, analysis and feedback necessary to reach those goals. Remember, these are extremely detailed metrics and goals. You must carefully assess the costs and determine if the incremental gain is worth the additional resources. (note - the methodology is always identified as a Six Sigma program, whatever the goals established.)

Whether you meet your six sigma goal or not, you will reap the benefits from the process of defining and understanding all the business intelligence components and iterating through the Six Sigma improvement process. You will see definite improvements and there may be breakthrough changes. There definitely will a better understanding of the product and the organization.

The Breakthrough Equation

Six Sigma Continuing Improvement Model for Business Intelligence is based on a simple to understand equation:

Y = f(X) + E

Y is the outcome, i.e. the result you need
X represents the inputs - the factors which create or impact the outcome;
f stands for function; i.e. y is a function of all the x factors. These X factors create and impact the Y/Quality.
E represents the presence of error. The formula needs to account for the uncertainty factor, i.e. the fact that the X factors combine to produce the Y result has some uncertainty.

Essentially, this basic formula means that a certain set of inputs is transformed by a process into an output. The formula represents the principle of determination: The independent variable(s) is operated on by a deterministic function (f) to transform it into the outcome (Y), i.e. the dependent variable. The Y results from, or is a function of, the Xs.

Every variation is caused by something. If you are going to control the output, you must control the cause. That means understanding the causal factors. You may be sick with the flu. You want to get well. The coach says you need to take vitamins (X1), jog three miles a day (X2), and practice 2 hours a day (X3) in order to become skilled enough to make the team. Great! But is it ever that simple? There is always some uncertainty in the outcome. Will you get sick or break a leg? Maybe 2 hours of practice is not enough for you. Everything is deterministic. An outcome is the result of inputs and a process which acts on those inputs. No matter how well the inputs and process are designed and executed, there is always some margin of error. There will be some variation.

We use the principle of determination as a basis for finding cause and effect. All outcomes are the result of the inputs and the process that acts on them, plus the error that creates variation. If you want to get better results, then you analyze the outcome, identify all the inputs and the processes involved in creating that outcome. If the outcome was good, then you will try to replicate the inputs and process. If bad, i.e. needs improving, that analysis should allow you to pinpoint some of the reasons for that less than perfect outcome and try to make it better next time.

Remember - the improvement process is iterative -try, try, try again.

Introduction

Statistics

The Six Sigma BI
Initiative

Should your goal be a Six Sigma level of Quality?

Maybe -- but then again, Maybe not!

How much will reaching that goal cost? Can we attain a slightly lower goal for appreciably less money, time and other resources? It is very expensive to create and manage a six sigma initiative for BI. Collection and management of metrics may be complex and cost extra resources which could be used on developing business intelligence applications. Watch the BI asset costs --- but also carefully assess and measure the costs you are spending to manage those BI assets.

previous pg