Data Analysis Retail Case

Management

The five management programs have the same common dependent variables. These are the average turnover, the weekly profit and the monthly staff time cost. The independent variable for this experiment is the management system that is used. There are five different management systems that are being used at the company, and they differ in their methods. The data presented show the impact of the different management systems on the different output measures (dependent variables).

The wild card is the type of store data. The company investigated this using three store categories, and presented its findings, but they were not presented with statistical analysis. As such, they should not be considered to be an independent variable.

Outcome variables are the dependent variables.. Ultimately, for this company the variables should reflect a wider variety of output measures for each store. . The output variables should be related to the success measures. First, it is important to measure the different levels of profit, not just net profit. This will provide more refined data and will allow management to be sure that the net profit figures are related to the management system.

Turnover is already a dependent variable, and are staff time costs, but if turnover is higher that implies that there are more new employees who are working. They are less efficient, and having higher training costs. I would like to capture the full set of costs for this turnover. There may be a point where the decision is made weighing turnover versus sales, and management will need to know what a basis point of turnover is worth. So capturing things like recruiting and training costs is important for the company, to help make that sort of decision.

The weekly profit is something that has to be compared to a baseline. Thus, I would want to capture the change in weekly profit. Right now, the company reports the weekly profit, but it is also known that there are differences in store composition for the different management types. Thus, it is important to know whether the high performing stores are the ones that were always the high performing stores. The change in profit before and after the new management system is actually more important than just knowing the amount of profit, because the experiment needs to compare current results with prior results.

Other measures related to performance are also worth measuring. For example, in retail the sales per square foot is a critical output measure (Investopedia, 2015), and sales per employee is another. These both can be captured with the data that the company has available. By capturing this data, differences between the high profit stores and low profit stores can be smoothed out.

I would also want to know what stores were performing the best in terms of the high margin goods. There should be a way to operationalize this variable, though it may well be captured with an average ticket metric. The average ticket metric is another common retail metric -- knowing how much each customer buys, on average, helps to information management about whether it is capturing higher margin goods, and in particular the impulse goods that increase the average ticket. If the average profit/ticket can be determined, that might also shed light as to which stores are doing a better job on moving high margin items.

Inventory turnover metrics are also worth examining, and refined data on product categories would be important. Basically, once the successful stores have been identified, it is important for management to learn why those stores have become successful. The company is at the first stage -- trying to identify the successful stores -- but gathering more data, with more variables, will allow the company to apply the lessons that these stores are learning across stores. The reason that this is important is because right now there are many stores that are holding meetings and at each of these meetings they are brainstorming, then they test the ideas, and then at the end of all of this they find out what works. But each store is doing this independently. If head office starts to gather this information about what each store is doing right, and then sharing that information, then that will allow the organization to move more quickly to implement the best ideas system-wide. Otherwise, each store will have to learn each good idea on its own, which is haphazard, and does not leverage the benefit of being part of a large chain of stores. So to capture this data, management can do something like find out what its highest margin products are, and test what stores saw the biggest changes in turnover for those products. This may be one of the important keys to improving profitability.

An input measure has to be the economic growth rate in the region where the store is located. The stores using the different management systems tend to be located in different types of areas. For example, System I and II stores tend to perform poorly on the profit measures, but they are also in areas that are economically depressed. Their performances might actually be quite good relative to the poor economies in those areas, whereas stores that are in high growth areas might only see profit growth as a consequence of growing populations and incomes. In other words, management needs to understand whether the relative success or failure of a store is related to where it is located, rather than related to the management system that was put in place. The overall economy has to be controlled as an input variable, in order that the other input variable (the management system) is isolated.

The structure of the inquiry basically starts at the current level, but with more variables, especially outcome variables measured, the information becomes more refined so that not only is it understood which management system is better, but why it is better. The current level of inquiry is basically looking at correlations, but for management practice the above variables will help to identify causal factors.

3. There are two measures for each dependent variable -- the mean and the standard deviation. To determine which is best requires an understanding of which is more important, because there are trade-offs between these two. It is believed that a lower standard deviation is important because the company is working with tight margins. The standard deviation is important in that respect because companies with tight margins require a fair amount of cost and revenue certainty to ensure that they are consistently profitable. This is especially true in retail, where the company will typically have high fixed costs associated with operating a store.

For the average turnover, System V is the best. It has a mean of 20%, where System IV has a mean of 17%, but the standard deviation for System V is 12%, versus 20% for System IV. What this means is that while the mean is slightly higher, that figure is more reliable. If the company is considering whether or not a store can be reliably profitable, then it will need to have a higher level of cost certainty. As a manager looking at a go/no-go decision, the certainty afforded by System V would be beneficial. That said, if the store is assuming that all stores are going to remain open, then the lower average is probably the more important variable, which would mean that System IV is better.

It is worth considering that System IV and System V share some common elements. They both have brainstorming, and it is likely that since they are both among the best systems for having a lower turnover that the brainstorming is correlated with a generally low average turnover. The relationship between sharing information and turnover is one that is less clear, but a reasonable hypothesis is that sharing information is going to result in higher turnover, as such a system may not be popular with some employees.

The method that is the less effective at lowering the average turnover is System III. This method has the highest average turnover and the highest standard deviation. As such, its results are poor and in some cases they are exceptionally poor. Management oriented towards consistent profitability cannot take the risk of having a store with turnover rates of 57%, which is what is implied to be a part of System III. As such, System III delivers unacceptably poor results in terms of average turnover. At issue here is the fact that System III is most likely to be implemented at urban stores where the workers are younger. This would hint at the likelihood that turnover was always higher in those stores. This is why it was recommended above to capture the change in turnover, rather than just the turnover rates. Stores with younger workers are much more likely to have high turnover rates than stores with older workers. Indeed, System III may still have had the greatest reduction in turnover rates of any store -- but the company does not know this, given the data that have been presented. It is easy to pick out System III as having the worst turnover rates, but with the youngest workforce this group of stores probably always had the worst turnover rates.

With respect to profits, System V has the best profits. It has the highest profits, and a lower standard deviation than the System IV stores, which had the second-highest profits. There is no trade-off here as the System V stores have both the highest average profit and the second-lowest standard deviation of profits. They are the best performing group of stores on both measures. Any decision based on profits would unequivocally result in the conclusion that System V is the best.

The worst-performing system in terms of profits is System I. This system delivers the lowest average and the lowest standard deviation for the profits. This means that these stores are not only poor, but consistently poor. At least with System II, which also have a low average profit, there are some better performers within that grouping. System I is comprised of the worst-performing stores in the company -- System IV and System V stores would almost never be as bad as the best of the System I stores.

The biggest issue with the profit figures is that they do not reflect the change in profit that might have occurred. The change is important. First, all stores are profitable, so it is unlikely that the company is facing a decision to close any stores. If all stores are to remain open, then the most important data would be that which reflects which system is best-correlated with improve profits. Remember that System I and System II stores tend to be of one type while the System III stores, and IV-V stores are of other types. System II stores are better than System I stores in terms of average profit, but these types are only really comparable with each other. They are older stores in economically depressed areas, so they are likely to be among the worst performers. This means that they are predisposed to lower profits -- the company needs to isolate the overall economic circumstances of the area in order to understand what role the management system has played with the dependent variables.

4.a The comparison of the numbers does influence my decision, but only partially. The numbers are illustrative in vague, broadly indicating that System Vis the best management system. However, there are some factors that need to be considered. First, the economies of the different areas and the composition of the workforces may be influencing the outcomes of the stores -- perhaps as much or more than the choice of management system. The data also do not illustrate the change in the dependent variables, which is more important than the absolute number. So while these numbers help to paint a picture, that picture is not detailed enough to genuinely be useful for making a decision about the best of the management systems.

4b. Yes, the fact that managers are choosing what system they are using, and the fact that managers of specific types of stores are oriented towards specific management systems does influence the analysis of the data. What manager choice means is that the data collected are not randomized. There are inherent biases in the choices that the managers are making, and that bias is reflected in the fact that one store type usually opted for I or II, another store type for III, and a final store type for IV and V. At best, a choice can be made for each store type (so choosing between I and II), but the data here are not comparable across the entire company because it is not randomized, and the biases of the managers of certain types of stores skews the data. This is why I would want to know, for example, the change in the profits under each system -- the stores that are most profitable today may have been the most profitable stores before as well.

4c. The advantages of randomization are that the data can be applied across the entire population. Right now, the samples are skewed by managerial bias. There are basically three different populations, based on types of stores. The conclusions about older stores in depressed areas cannot be applied to urban stores with younger workers, because these store types are not similar enough . Randomization of the management systems would provide more robust data that would reflect the entire population. The current method of sampling does not provide data that can be used across the entire population (Taylor, 2015).

5. Both the nature of the workforce and the economic data for each area will affect the decision-making. Right now, each type of store is its own population and the data that have been gathered can only apply to the three different types of stores. The data cannot be applied across all stores, because of the differences in the economic situation of the store areas, and the differences in the compositions of the workforces.

The participation of an experienced workforce does not help or hurt here. The company is trying to determine a management system that it can apply across the entire company. Experienced workers are a part of the population, so they must be included in the sample. You cannot have a sample that only features experienced workers, but as part of the population they must be included.

A program might work better in an economy that is not doing poorly, but it might not. The current outcome measure of profits can change entirely based on the economy. The most profitable stores are in the best areas economically, but that might just be a function of the economy and have nothing to do with the program. Management has to find a way to separate out the success of a store based on the management system and the success of a store based on the strength of the area's economy.

6. Clausen has done relatively poorly. There are a few reasons for reaching this conclusion. First, Clausen has not randomized the study, so the findings cannot be applied across the entire company, which was one of the objectives of the study. Second, there is no way for the company to distinguish whether the dependent variables are changing because of the independent variables, or because of some other factor. This is also related to the sampling flaw. Thus, Clausen at this point does not know enough to make a decision based on this information.

The other thing about Clausen's design here is about the system does not actually measure motivation or goal-setting. Those variables have been poorly operationalized. A good way to explain this is to look at System III. That system has the highest turnover, which looks like a bad system for motivation. But that system was implemented in stores with mostly young employees, who are more naturally prone to turnover. So turnover, if not controlled for factors like age, is not a good measure of motivation or goal-setting. This is not to say that motivation and goal-setting should have been the objectives of the study. This is a business, so profits are the most important outcome measure. The study captures that data. Motivation and goal-setting are only important to the extent that they contribute to profits. The fact that this study does not capture motivation and goal-setting effectively is a negative mark on Clausen's work, but the fact that it captures something about profits is a positive attribute that the company can build on going forward.