Data envelopment analysis
Data Envelopment Analysis (DEA) is a mathematical approach used to assess the efficiency of decision-making units (DMUs) within an organization. DMUs, such as franchises or branches of a business, have some autonomy but operate within a larger controlling framework. DEA compares these units by creating ratios of outputs (resources produced) to inputs (resources invested), allowing organizations to identify which units are underperforming. The method was first introduced in 1978 by researchers Abraham Charnes, William W. Cooper, and E. Rhodes, and it has since been utilized in various sectors to improve performance.
DEA works by measuring relative efficiency, which is visualized through charts or graphs for easier interpretation. Accurate data collection is crucial; any inaccuracies can lead to misleading efficiency results. Once efficiency ratios are generated, organizations can make informed decisions to enhance performance, either by increasing outputs or reducing inputs. For instance, a poorly performing restaurant might improve efficiency by boosting sales through marketing efforts or by reducing workforce numbers. Regular application of DEA can ensure that organizational changes effectively enhance operational efficiency over time.
On this Page
Subject Terms
Data envelopment analysis
Data envelopment analysis (DEA) is a mathematical system designed to measure the efficiency of decision-making units (DMUs) within a larger structure. DMUs have limited autonomy, which means that they are allowed to make some of their own decisions but are still controlled by a larger organization. Franchised businesses, such as stores or restaurants, are a common example of DMUs.
DEA measures the relative efficiency of DMUs within a structure. It creates a ratio of outputs and inputs for each DMU to measure efficiency. In this context, outputs refer to the resources created by the DMU, while inputs refer to the resources invested by the organization. Once efficiency ratios are collected, they are often displayed in a convenient chart or graph format. The organization can then identify which DMUs are underperforming and take corrective action. In many cases, corrective action includes assisting the DMU in increasing outputs. However, it may also involve reducing or eliminating any unnecessary inputs.
Background
Data envelopment analysis (DEA), sometimes called frontier analysis, is a metric designed to measure and improve performance. It was first proposed by Abraham Charnes, William W. Cooper, and E. Rhodes in 1978, after which it was quickly put into practice. DEA is designed to measure the relative efficiency of decision-making units (DMUs) within a business, an organization, or other structure. It accomplishes this by comparing such units against one another, giving the person analyzing the organization a concrete set of data to work with for planning future changes. Should DEA work as intended, the individual planning to make changes to the organization can then continue to employ DEA, checking if these changes have increased or decreased the relative efficiency of the DMUs.
Frontier analysis works best when the DMUs of an organization have limited autonomy. They need to be capable of making some decisions for themselves and thus may need to undergo some self-prompted changes to increase their efficiency as measured by DEA. However, in order for DEA to function properly, the DMUs cannot be completely autonomous and must still be part of a larger organization. Without this framework in place, it is difficult for any central authority to prompt the changes required to increase the efficiency of any DMUs.
Because of these requirements, DEA tends to fit best within the context of a large business or organization. DMUs can be branches, franchises, or worksites. In many cases, these locations are granted limited autonomy in their day-to-day operations but are still under the authority of a powerful central organization.
Overview
DEA uses visual data representations to show the relative efficiency of DMUs. This allows a central organization to accurately compare the effectiveness of DMUs within an organization or network, leading to further analysis and an eventual increase in efficiency. In order for DEA to function, the person conducting the analysis must first choose a metric to measure. For example, a franchised restaurant might choose the number of meals sold per day. The organizer must then conduct surveys, gathering the relevant data from each of its branches. It is important that the organizer use sound methods when collecting this data. If the information collected is inaccurate, the results displayed by DEA will also be inaccurate.
In this instance, each restaurant is a DMU. It is allowed to make many decisions on its own, including its staffing choices, hiring practices, and shift assignments. However, the menu, décor, and pricing are set by a central organization. The number of meals sold per day, the output measure, is accurately reported by each DMU. However, that is not enough information to accurately gage the efficiency of each restaurant. Input measures, meaning the amount of resources the company spends to create the output measure, must also be measured. In this instance, a simple input measure is the number of people employed by each restaurant.
To gauge the efficiency of each restaurant, DEA employs ratios. One of the most common ratios used in DEA is the output measure divided by the input measure. In this example, dividing the number of meals sold by the number of people employed at a restaurant roughly shows the efficiency of the employees at each restaurant. DMUs with higher ratios are considered more efficient, while DMUs with lower ratios are considered less efficient. Assuming that all employees are paid similarly, the DMU with the highest ratio is the one contributing the most to the overall organization.
Once the data collected has been checked for accuracy and efficiency ratios have been generated, the data must be displayed in a way that makes it easy to communicate. In many cases, interpreting a long list of ratios is not intuitive for readers. Instead, the organization conducting DEA arranges the results in a chart or graph.
If carried out properly, DEA shows which DMUs are performing well and which are performing poorly. Because DEA accounts for a ratio between inputs and outputs, organizations seeking to improve poorly performing DMUs have two options: increase the output number or decrease the input number. In the case of a poorly performing restaurant, the organization may seek to help the location drive up sales with a new series of advertisements. This would drive up the output numbers, increasing the ratio and improving performance. However, the organization could also choose to reduce the number of employees at the location. Eliminating unnecessary employees reduces the input number, which also improves the efficiency ratio. After a course of action has been decided upon, DEA should be carried out periodically to ensure that any changes are continuing to increase efficiency.
Bibliography
Beasley, J. “Data Envelopment Analysis,” people.brunel.ac.uk/~mastjjb/jeb/or/dea.html. 26 Jun. 2019.
Feroz, E. H. “Financial Statement Analysis: A Data Envelopment Analysis Approach,” Journal of the Operational Research Society, 2003, hwww.tandfonline.com/doi/abs/10.1057/palgrave.jors.2601475. 26 Jun. 2019.
“How Data Envelopment Analysis Works,” Frontier Analyst, 2018, banxia.com/frontier/resources/how-data-envelopment-analysis-works/. 26 Jun. 2019.
Khezrimotlagh, Dariush, et al. “Data Envelopment Analysis and Big Data,” European Journal of Operational Research, May 2019, www.sciencedirect.com/science/article/abs/pii/S0377221718309123. 26 Jun. 2019.
Kumbhakar, Subal C. et al. “Stochastic Frontier Analysis: Foundations and Advances,” University of Miami, 2017, www.bus.miami.edu/‗assets/files/repec/WP2017-10.pdf. 26 Jun. 2019.
Mcclure, Ben. “Measuring Company Efficiency to Maximize Profits,” Investopedia, 14 Apr. 2019, www.investopedia.com/investing/measuring-company-efficiency/. 26 Jun. 2019.
Shero, Jeffrey A., et al. "Data Envelopment Analysis (DEA) in the Educational Sciences." Journal of Experimental Education, 9 Apr. 2021, doi: 10.1080/00220973.2021.1906198. Accessed 22 Nov. 2024.
Wober, Karl W. “Data Envelopment Analysis,” Journal of Travel & Tournism Marketing, 2007, www.tandfonline.com/doi/abs/10.1300/J073v21n04‗07/ 26 Jun. 2019.
Yong-bae, Ji. “Data Envelopment Analysis in SATA,” The Sata Journal, www.cgdev.org/sites/default/files/archive/doc/stata/MO/DEA/dea‗in‗stata.pdf. 26 Jun. 2019.