Challenges of using DEA-Solver for data envelopment analysis

Data Development Analysis (DEA) was developed in 1978, as a benchmarking technique for evaluating the efficiency or performance of a group of entities (Cooper, Seiford, & Tone, 2007). DEA analysis has emerged as a powerful tool of efficiency and benchmarking-based analyses, as it uses linear programming methodology. DEA analysis remains possible over various software. However, DEA Solver is the most recent and popular software for DEA analysis. Data Envelopment Analysis using DEA Solver, which is basically an MS Excel plugin and downloadable freely from DEA solver 8.0 includes 28 clusters of DEA models (“User ’ s Guide to DEA-Solver-Learning Version,” 2015). However, there are two versions of DEA solver, free and paid versions and can compute the results for maximum 50 decision-making units (DMUs).

The basic process of the data envelopment analysis using DEA-Solver

The process of incorporating the DEA solver into the MS Excel spreadsheet requires a few steps as mentioned in the following steps;

  1. Install and enable the excel spreadsheet to choose the program referred to as add in.
  2. MS Excel>Excel options>Add-ins>Excel Addins>Go>Solver>OK.
  3. After setting up the DEA model, involve the solver by using the solver option indicated under Data tools.
  4. To run the solver model, set the input and output parameters in the solver and indicate the type of DEA analysis.
The basic process of the Data Envelopment Analysis using DEA-Solver

The basic process of the Data Envelopment Analysis using DEA-Solver

Challenges faced while conducting data envelopment analysis using DEA-Solver

From the process of installation of DEA-Solver to the computation of the analysis, the DEA process undergoes a various set of challenges. The first challenge incurs at the time of the installation itself, as it requires to be manually included in MS Excel. However, the free version allows analysis on a certain set of models. This free version of DEA-Solver allows the analysis only on Slack-based model and the Malmquist productivity summary. The output and input-oriented models such as CCR-O, CCR-I, BCC-O, BCC-I, and other non-discretionary variable models can’t be analyzed through this version. A paid version of the DEA-Solver allows all types of DEA model analysis.

Another challenge includes a longer process of entering the data and defining variables in the free version of DEA-Solver. Every time, the researcher needs to characterize the data into input and output units. Further, the free version of DEA-Solver software does not provide the option to save the results or data file. However, there are a certain set of solutions that can be used by the researcher to avoid these challenges that are explained in the following section.

Missing data as a challenge of DEA-Solver

DEA analysis using DEA-Solver does not help analyze incomplete data for inputs and outputs. There are many situations under which one may face the problem of missing data, especially from secondary data. The most common solution for this problem is to eliminate the DMUs, inputs, and output variables that have missing data. However, eliminating the DMU’s may cause inefficiency in developing the model. Another alternative comprises using previous year value to fill the missing data.

Another alternative approach is to use the Fuzzy mathematics approach (Qian, 2009). This technique involves significant effort to find the missing values. Through this approach, the missing values are forecasted based on the value of mean and the series of variations. Therefore, the best solution is to fill the gap of the missing data using previous year values or use the predicted values calculated using the series of variations.

Time-consuming as a challenge of DEA-Solver

Importing data and defining the input and output variables requires multiple insertions. This is because the DEA-Solver does not provide any option to save the data and result file. So, after analyzing the results of a particular model for each year, the results have to be copied in a separate excel sheet. This is in the case of the free version of DEA-Solver. One needs to import the data file multiple times and define the input and output variables for further analysis. This process continues for each assessment and is time-consuming.

Negative values as a challenge of DEA-Solver

Many a time, the data set entered can have negative values. DEA analysis using DEA-Solver can’t execute the analysis with negative numbers. Again this is a limitation of the free DEA-Solver version. For this purpose, it is necessary to ensure that all negative values are removed from the data before importing it. The solution includes the reduction of the magnitude of negative values by defining the values based on a constant value. Another method includes the rejection of data that have negative values and choosing all positive values for the DEA analysis.

Alternative models as a major challenge of DEA-Solver

The free version of DEA-solver does not provide the output for CCR and BCC models. The software provides the output only for Slack-based model (SBM) and the components of Malmquist Index Productivity. In order to analyze the score of CCR technical efficiency and BCC models, the paid version of DEA frontier software should be used. One can also use a variety of user interfaces and advanced modelling options like Max DEA software and DEA Frontier.

Recommendations of conducting DEA analysis using DEA-Solver

Following recommendations help to obtain accurate results for DEA analysis.

  • The number of DMU’s selected for the analysis should be chosen carefully. A large set of the population causes homogeneity of data and in result cause exogenous impacts on the data.
  • Use Max DEA 7 software in combination to DEA-Solver, for calculating CCR technical efficiency scores and output of BCC models which is not allowed under the free version of DEA-Solver (Qian, 2009).
  • Avoid imbalances in the data set and comprise of a similar magnitude.
  • Select the DMU’s with sufficient data available for inputs and output variables are available. In addition to this, ensure that all the values in the data set are non-negative.


  • Cooper, W. W., Seiford, L. M., & Tone, K. (2007). Data envelopment analysis: A comprehensive text with models, applications, references and DEA-solver software: Second edition. Data Envelopment Analysis: A Comprehensive Text with Models, Applications, References and DEA-Solver Software: Second Edition, (June), 1–490.
  • Qian, C. and. (2009). Data Envelopment Analysis : Methods and MaxDEA Software.
  • User ’ s Guide to DEA-Solver-Learning Version. (2015), 15(1), 1–17.
Divya Narang

Divya Narang

Research analyst at Project Guru
Divya is a Masters in economics with a specialization in econometrics. She has worked on various research works in the field of economics and finance. Apart from this, she has developed keen interest in business management, business policy, international marketing and strategic management. Her methodological work focuses on analysis on panel data using statistical software like SPSS and STATA. She loves to spend her spare time reading novels and playing badminton.
Divya Narang

Related articles

  • How to develop questionnaire for correlation and regression test? There different instruments to collect primary data and the most widely used is the questionnaire in a survey method. Correlation and regression tests are two of the basic statistical tools that are widely applied to analyze data.
  • Structural equation model (SEM) Structural equation model is a statistical modeling technique. Structural equation model (SEM) tests estimate or establish relationships between variables. It is a multivariate statistical data analysis technique. SEM analyzes the structural relationships or to establish causal […]
  • Importing data and defining the variables in DEAP In the previous article, "Using Data Envelopment Analysis (DEA) program for measuring efficiency" the process to download Data Envelopment Analysis program (DEAP) was explained and the case problem in hand was defined. The first step is to understand how to import data to DEAP.
  • Hamlet II is a tool for quantitative textual analysis Quantitative analysis of textual data has been long used in the field of social sciences research. This includes data collected from open ended surveys and interview questions. Analysis of text-based data helps in assessment of trends in social sciences.
  • Advantages of using R statistical software for predictive modelling Predictive modelling is a data driven, induction based modelling that is continuously used by big sized companies to gain useful insights into trends and risks budding in the future. The modelling on the basis of data extraction, cleansing and analysis helps in predicting the value of a […]


We are looking for candidates who have completed their master's degree or Ph.D. Click here to know more about our vacancies.