Comparing Retail Management Systems: Variables and Analysis

Dependent and Independent Variables

The five management programs share the same common dependent variables: average turnover, weekly profit, and monthly staff time cost. The independent variable in this experiment is the management system in use. There are five different management systems being used at the company, each differing in its methods. The data presented show the impact of these different management systems on the various 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 those findings were not accompanied by statistical analysis. As such, store type should not be considered 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, and those output variables should be related to the relevant success measures.

Recommended Additional Outcome Measures

First, it is important to measure different levels of profit, not just net profit. This will provide more refined data and allow management to confirm that net profit figures are actually related to the management system rather than to other factors.

Turnover and staff time costs are already captured as dependent variables, but if turnover is higher it implies that more new employees are working—employees who are less efficient and who generate higher training costs. Capturing the full set of costs associated with turnover would be valuable. There may come a point where management must weigh turnover against sales, and to make that decision they will need to know what a basis point of turnover is worth. Capturing costs such as recruiting and training is therefore important.

Weekly profit must be compared against a baseline. The company currently reports weekly profit, but it is also known that there are differences in store composition across the different management types. It is therefore important to know whether the high-performing stores today were always the high-performing stores. The change in profit before and after the introduction of a new management system is actually more important than the absolute profit figure, because the experiment needs to compare current results with prior results.

Other performance-related measures are also worth capturing. In retail, for example, sales per square foot is a critical output measure, and sales per employee is another. Both can be derived from data the company already has. Capturing these metrics allows differences between high-profit and low-profit stores to be smoothed out.

It would also be useful to know which stores are performing best on high-margin goods. One way to operationalize this is through an average ticket metric—knowing how much each customer spends, on average, informs management about whether the store is successfully moving high-margin and impulse goods. If average profit per ticket can be determined, that may shed light on which stores are doing a better job of selling higher-margin items.

Inventory turnover metrics are also worth examining, and refined data on product categories would be important. Once the successful stores have been identified, management needs to understand why those stores have become successful. The company is currently at the first stage—identifying successful stores—but gathering more variables will allow the company to apply lessons learned across all stores. Right now, individual stores are holding meetings, brainstorming, and testing ideas independently. If head office begins collecting this information and sharing it, the organization can implement the best ideas system-wide far more quickly. Otherwise, each store must rediscover every good idea on its own, which is haphazard and fails to leverage the benefit of belonging to a large chain. One concrete step: management could identify its highest-margin products and measure which stores saw the greatest changes in turnover for those products. This may be one of the important keys to improving profitability across the chain.

An important input measure is the economic growth rate in the region where each store is located. The stores using different management systems tend to be located in different types of areas. System I and System II stores, for example, tend to perform poorly on profit measures, but they are also located in economically depressed areas. Their performance might actually be quite strong relative to the poor local economies, whereas stores in high-growth areas may be seeing profit growth simply as a consequence of rising populations and incomes. Management needs to determine whether a store's relative success or failure is related to its location rather than to the management system in place. The overall economy must be controlled as an input variable so that the effect of the management system can be properly isolated.

The structure of the inquiry currently operates at a broad level, but as more outcome variables are measured, the information becomes more refined. Ultimately, this allows management to understand not only which management system performs better, but also why it performs better. The current level of inquiry is essentially correlational; the additional variables recommended above will help to identify causal factors.

Evaluating Management Systems by Mean and Standard Deviation

There are two measures for each dependent variable: the mean and the standard deviation. Determining which is more important requires acknowledging the trade-offs between them. A lower standard deviation is considered particularly important because the company operates on tight margins. Cost and revenue certainty are essential for consistent profitability—especially in retail, where high fixed costs associated with running a store leave little room for variability.

For average turnover, System V is the best performer. It has a mean of 20%, compared to System IV's mean of 17%; however, the standard deviation for System V is 12%, versus 20% for System IV. While the mean is slightly higher, the figure is considerably more reliable. For a manager facing a go/no-go decision on store viability, the certainty afforded by System V is beneficial. That said, if the assumption is that all stores will remain open, then the lower average becomes the more important figure, which would favor System IV.

It is worth noting that System IV and System V share some common elements—both include brainstorming sessions—and it is likely that this shared feature is correlated with generally lower average turnover across both systems. The relationship between information-sharing and turnover is less clear; a reasonable hypothesis is that such a system may not be popular with some employees, potentially driving higher turnover in certain contexts.

The least effective system for controlling average turnover is System III, which has both the highest average turnover and the highest standard deviation. Its results are poor and, in some cases, exceptionally poor. Management oriented toward consistent profitability cannot accept the risk of turnover rates implied to reach 57% under System III. However, it is important to note that System III is most likely implemented in urban stores with younger workforces, who are naturally prone to higher turnover. This is why capturing the change in turnover, rather than the absolute rate, was recommended above. System III may have actually produced the greatest reduction in turnover of any system—but the company cannot know this from the data as currently presented.

With respect to profits, System V is again the best performer. It has the highest average profits and a lower standard deviation than System IV stores, which had the second-highest profits. There is no trade-off here: System V stores have both the highest average profit and the second-lowest standard deviation. Any decision based purely on profits would unequivocally identify System V as the best.

The worst-performing system in terms of profits is System I, which delivers both the lowest average profit and the lowest standard deviation. These stores are not only poor performers—they are consistently poor. System II, which also has a low average profit, at least contains some better performers within the group. System I represents the weakest stores in the company; stores operating under System IV or System V would almost never perform as poorly as the best System I stores.

The most significant limitation of the profit figures is that they do not reflect the change in profit that may have occurred under each system. Since all stores are currently profitable, it is unlikely that the company faces imminent store closures. If all stores are to remain open, the most important data would be those that reveal which system is best correlated with improving profits over time. System I and System II stores tend to be older stores in economically depressed areas, predisposing them to lower profits. The company needs to isolate the broader economic circumstances of each area in order to understand what role the management system has actually played in determining outcomes.

References

Baldwin, S. (2006). Organisational justice. Institute for Employment Studies.

Hannan, M., & Freeman, J. (1984). Structural inertia and organizational change. American Sociological Review, 49(2), 149–164.

Investopedia. (2015). Sales per square foot.

Taylor, C. (2015). What is a simple random sample?