I am coming to xxxxxxxxx with a research proposal that could quite possibly revitalize the spirit in which it was founded. Since xxxxxx first began in 1978, quality workmanship has been one of the principles around which its foundation was constructed. It is that very same principle that established a clientele which has supported xxx over the years with repeat business and word of mouth recommendations.

After having the privilege of working for xxx this past summer I was able to gain some insight on a problem that could potentially crack that foundation. After years of high quality performance it appears to me that, like many other opportunistic companies, xxx has let targets and numbers get in the way of the thing that made them the largest residential painting company in the world. That thing is quality.

It is my assertion that the production target driven structure which xxx has grown to embrace is the culprit in an ongoing decrease in overall quality. If given the chance I am capable of doing pertinent business research that can determine the existence or lack there of of an association between production targets and quality ratings.

This research will not only allow management to understand the connection between these targets and quality but also enable them to pick any desired level of overall quality by choosing the corresponding production target.

Such understanding will inherently lead to and increase in overall quality. However, there are countless indirect results of such an understanding. To name a few: increased profits, increased demand, lowered stress, less turnover and on and on.

The majority of this proposal details specifically how I would go about obtaining and analyzing this data as well as what it could do for you. Thank you for your time and consideration.

NOTE: If the reader is not familiar with xxx please read the brief company description in Appendix A.

xxxx has three main principles by which it stands. These principles are intended to guide administrators in the decision making process and lead managers on a daily basis as they work with customers and painters. As I understand it, these three statements are the commandments of the xxx business. Do as they say and you can do no wrong.

In everything you do quality should be the motivating factor.

three principles and today this commandment is under scrutiny. Recently I had the privilege of working for The Franchise Company(founder of xxx), so I have some insight as to how the company operates. One of the things that I noticed was a discrepancy between quality and production targets(aka Designated Target DT).

Larger production targets do not inherently imply that quality will go down. If an emphasis is placed on quality above all things then a well run business can maintain the same level of quality no matter what the DT. However, I believe that one can find some drastic inconsistencies between what xxx preaches and what xxx teaches.

From personal experience, I can attest to the fact that hitting your DT is the number one priority. As I would expect, xxx administration maintains that quality(one of the founding principles) is their number on priority no matter what the DT. The problem here is, through no fault of their own, xxx is unaware of the impact that higher DTs have on young managers interest in quality.

In an insert from a xxx field manual called “Manager Success Model,” the contradiction is blatantly obvious. On one side of the sheet points are given for quality, profit and volume.(see Appendix C) The points awarded at the “star” level are twice as much(twice as significant) as both profit and volume which are equal in points. However, on the other side of the sheet the exact wording is “xxx and your General Manager will deem your summer a success provided that you hit your sales target, make between $6,000 and $10,000 and deliver quality service to your customers…” In my opinion the order of the latter is more indicative of the true culture.

There are three factors that undermine the quality commandment.

1) When a manager hits his DT he pays less royalty on each job produced thereafter.

2) The general managers who set the DTs for the managers receive a bonus when each manager hits the designated DT.

3) The company as a whole will naturally make more money as more DTs are met.

It is this structure that undermines the culture and specifically the quality that xxx’s three principles try and instill. It is not necessarily intentional but rather unavoidable.

In my opinion, xxx has the right idea in trying to emphasize quality. Quality is the cornerstone of the most powerful means of getting more contracts and that is word of mouth. It is the number one goal of most companies to provide a service or product at such a level that new business is generated by word of mouth.

According to Gitlow the benefits of improving quality are similar. He puts them in this order:

5) Cost per good unit or service is lowered

6) Worker morale goes up”(Gitlow, 1990)

Poor quality of work damages the business more than large DTs help it. It is hard to calculate the effects of low quality on a business because the repercussions are intangible.

Here are three of the main outcomes from a low level of quality. For one, there are an inflated number of complaints that have to be dealt with on a daily basis by employees whose sole job it is to listen to complaints. These are employees that could be used elsewhere or not at all.

Secondly, every year there are several huge lawsuits that are the direct result of neglect. The majority of these settlement fees and lawyer fees could be avoided with a consistent emphasis on quality.

Finally, one of the toughest things for xxx to do year in and year out is bring in qualified managers and train them. This requires a huge amount of resources. Quality is one of the main factors that ultimately drives managers away which leads to xxx having to rehire and retrain. When quality is high there are fewer job-site problems and an overall lowered level of stress for the manager. This means a happy manager at the end of the season and usually a returning manager (CPPs most valuable asset).

By continuing to emphasize quality and restructuring certain aspects of the DT program, xxx may initially show a drop in revenue. On the other hand, the ultimate benefits of such an adjustment bode increased profits, smoother operation, and long term stability. As Brown puts it “The goal of research is not to measure past performance, but to drive future behavior.”(Brown, 1991)

With the right data and the proper research I feel that I can provide xxx with the information necessary to create a marriage between production targets(DTs) and the quality of the work. Such a marriage would allow for a maximum level of production at a desired level of quality.

Hypothesis: An increase in production targets will lead to an overall decrease in quality ratings.

1) I will attempt to demonstrate to you that there is a relationship between production targets and the quality ratings from customers.

2) I will dissever the nature of the relationship determined in (1).

3) I will provide you with the necessary information for determining the level to adjust your DTs to in order to achieve a desired average quality rating.(ie you want and average quality rating of x out of y—I will tell you how to set a DT to achieve that level of success)

Plainly put, the goal of the research which I would conduct is to provide xxx with the know-how to dictate their own quality level. The information from this research will allow me to demonstrate to you how xxx can maintain a desired level of excellence(as per its principles) while maximizing production.

In order to achieve the former objectives we must go through all the steps that are involved in a research project. The following is a practical order of events from start to finish.

 Identify the problem/opportunity (this has already been done-see above)

 Design a questionnaire/survey for the target population

 Determine method of sampling

 Distribute questionnaires to sample

 Decide which method of displaying data makes analysis most efficient

 Figure out which test will best suit the objectives

 Work with management to determine appropriate levels of certainty

 Present results in layman’s terms to management

To do this research for xxx I would need information first on quality ratings and then information on the corresponding DTs. (ie customer x gave a rating of 3 and he was the customer of a manager who had DT y)

Because xxx does not have existing data on customer satisfaction as it pertains to quality, I would be designing the questionnaires from scratch. When designing a questionnaire there are a few things to keep in mind.

– questionnaire should be relatively short (about 1 page)

– questions should move from impersonal to more personal

– specifically, this questionnaire should cover all aspects of quality(idea of quality may vary from person to person)

(personal communication, Dickson, BPA 402, 1999)

See Figure A in appendix(rough draft of a questionnaire)

This particular questionnaire asks about quality of a xxx job. Possible answers are provided on an ordinal scale which provides for distinct differences between responses. Specifically, it was designed around the Likert scale model. In this model a statement is made about quality and respondents are asked to respond based on how much or how little they agree with the statement. (ie the quality of work was good: 1)strongly disagree 2)disagree 3)neutral 4)agree 5)strongly agree) By making a statement and making the end points of responses polar opposites, I can determine the degree to which the customer feels one way or another about quality. In this case, all statements about quality are positive so that a high score(5) represents high quality and vice versa.

These types of questions are also known as closed questions. “Closed questions have become the mainstay of survey researchers.”(Patricia Labaw, 1980) However there are several drawbacks to the closed question, the most notable being “that researchers may not know what the answers really meant to the respondent.”(Ibid) In this case, I feel like the closed question will be easier to analyze and more efficient.

In this case we want our population to consist of people who have already had their houses painted by a xxx manager. In order to minimize variables that might affect quality other than DTs, we also should limit the population to customers in the state of Washington and Oregon from the 1999 season.

Sample size represents the number of people from the target population that are going to receive the questionnaire. Sample sizes will vary depending on what we want to do with the data and how confident management wants to be that the information wasn’t tainted by sampling error.

The sample for our purposes will be a probability sample. This is generally more time consuming and costly, but it will allow us to project our results onto the entire population and make a more confident conclusion since it is a random sample. (each member has an equal chance of being selected) The probability of being selected is equal to 1/n where n is the number of people in the population.

I would choose the stratified sampling method which would allow me to split the population into two mutually exclusive groups.(ie income or geographic location) Next, samples are taken from each group at random. This can help to minimize random sampling error. For instance, we don’t want people from only one level of income. In this case, I would split my population by geographic location. In the xxx business it is common knowledge that customers vary drastically from area to area. Another appropriate reason for using the stratified method is that is requires smaller sample sizes. In this situation, the total population is not that big. Because the total sample size should not excede 5% of the total population a smaller sample is adequate.

One of the strengths and weaknesses inherent in the stratified sampling technique is the relatively smaller size of the sample. This means less time and cost but potentially greater random error. Another weakness is that the information necessary to properly stratify the sample is not always available.(Gates, McDaniel; 1999) In our situation this is not the case. In addition, stratification can be time consuming because of the time required to obtain the necessary information. Again this doesn’t affect our research because the segmenting information is internal and readily available.

Once the means of sampling has been established you can determine the size of the sample by working with management and the designated objectives.

(See Appendix Ba for equation and calculation)

What this example data is telling us is with a sample size of 89 we can be 95% confident that the true mean of the quality ratings will be within 5% of the true population mean.

The standard deviation of this equation will have to be estimated. Usually we would use the results of a pasts study but there is no such study. Another option would be to conduct a pilot study and calculate the standard deviation for those results. In this case we can be extremely positive that any answer will be within 6 standard deviations of the answer. Thus, 6/5=1.25 is the equivalent of our standard deviation.

Once the sample size is determined we are ready to send the questionnaires out to the customers. Ultimately we want to find average quality ratings for different DTs. This means that we need to separate or distinguish the questionnaires that come back by their respective DTs. One way to do this is to make a mark on the questionnaire itself. (ie DT=$60,000 gets a red dot; DT=$100,000 gets a yellow dot; DT=$120,000 gets a black dot) This will allow the person receiving the returned questionnaires to separate them by DT and average them separately.

Customers will receive the questionnaires on quality via ground mail. Although personal interviews would be preferable, they would not be cost effective or efficient. Instead, the respondents will receive a questionnaire including an envelope that is self-addressed and stamped so that it is easy to return. 89 questionnaires will go out to customers of managers with a DT equal to $60,000. 89 questionnaires will go out to customers of managers with a DT equal to $100,000 and the same for the $120,000 DT level.

One of the things that mail surveys always have to deal with is non-response bias. In a smaller sample, it could drastically affect our results if people do not respond for one reason or another. To minimize this possibility, we will hand stamp and hand address envelopes in order to personalize the exchange. If this doesn’t have the desired affect and people don’t respond then we will send out a second wave of questionnaires.

Initially, we want to test the results of the data we receive for a direct relationship. In other words, we want to make sure that the differences that are apparent between what is observed and what is expected are not due to chance. But rather, we want to be able to show that average quality ratings vary significantly enough with changes in DTs to warrant our continued analysis. One positive aspect of initially determining whether or not DTs affect quality is that if DTs do not affect quality, we can stop the research and save money and time.

This is not as complicated as it sounds, but rather one of the simpler ways to conduct research analysis. The normal way to set up a crosstabulation table is to make a table where the rows(horizontal) consist of categories that influence the data in the columns(vertical).(ie the production targets influence the quality ratings) See Crosstabulation Table in Appendix.

Once the table has all the data filled in, percentages can be easily calculated on the basis of row totals. The results lend themselves to easy comparison of the degree of correlation between DTs and quality ratings.

In addition, we will be able to look at a comparison between the observed(O) value and the expected(E) value in each cell. The expected value is representative of the value that would be observed if there was no difference between the variables.(then O=E) The observed value is taken from the data off the questionnaires. The comparison of the observed and the expected is done using a test called Chi Squared or X2.

“The X2 test enables a research analyst to determine whether an observed pattern of frequencies corresponds to or fits and “expected” pattern…many marketing research studies, possibly most, go no further than crosstabulation in terms of analysis.”(Gates, McDaniel; 1999)

When conducting a Chi Squared test the first step is to establish the null and alternative hypothesis. In the case of X^2, the null hypothesis (Ho) is always an association of no relationship between the two variables(DT and quality). The alternative hypothesis (Ha) is a significant relationship between the two variables. By looking at the answer to the equation, we can compare it to the Chi Square table. From that we determine whether to accept or reject a difference between the two variables.

From the crosstabulation table we can easily apply our data to the X2 equation. See Figure Bb for equation and calculations.

We know the observed value from our data collection. The expected value can be calculated using row sums, column sums and totals. See Figure Bd. And K is equal to the number of categories. This is all the necessary information to calculate Chi Squared. See Figure Bc.

Note: the data that was used was completely fictitious and is used for demonstration purposes only.

To calculate the significance of this information two more things are needed. The first is degrees of freedom and the second is confidence level. With these two pieces of information the results from the Chi Squared test are used to determine a corresponding value from the Chi Squared Table that will relate back to the null hypothesis(Ho). From this information it can be determined if a significant relationship truly does exist between production targets and quality.

Degrees of freedom can be calculated from the crosstabulation table by simply subtracting one from both the number of rows and columns and then multiplying the two.

(R-1)(C-1) = (3-1)(5-1) = 8 degrees of freedom

After that, the researcher would work with management to determine how confident management would want to be in the answer in order to move forward and take action. Because it is better to error on the side of confidence I chose a %95 confidence level for this example.(industry standard)

At these levels the Null Hypothesis exists at 15.5073

Consequently we can reject the null hypothesis at every level DT because our calculated value is much higher than Ho.(54.5 vs 15.5073) From this we conclude that there is a significant difference between quality and DTs.

Further analysis of the data will allow us to determine a very close estimate of the exact relationship between DT and quality. To do this type of extended analysis I would recommend bivariate regression analysis. Although seemingly daunting, this sort of analysis will clearly show the data and the nature of the relationship in such a way as to permit the management to determine where to set production targets in order to achieve desired levels of quality.

We use bivariate regression analysis to determine the strength of the linear relationship between two variables when one is considered dependent(y) and the other independent(x).(Gates, McDaniel; 1999) In this research, the DT is the independent(x) variable while the quality is the dependent(y) variable.

Interpreting the graph of the bivariate regression analysis is not difficult. There is an obvious inverse or negative relationship. This means that as one variable increases the other decreases. The benefit of such an analysis to the management is our ability to extrapolate that data. By taking a straight line graph and making a best fit line through the data points you can look at a DT and immediately get a sense of what sort of quality to expect from your managers.

Finally, I feel I have demonstrated that the objectives laid before you in this proposal are reachable and actionable. The insight that this sort of research could provide xxx management would be unparalleled. It could provide tremendous leverage for maintaining and strengthening your customer base into the next millenium.

The overall quality of the work was high.