Measuring DEA efficiency in Internet companies
Introduction
The short story of the dot com business sector is full of highs and lows. In the early years of the Internet, their main objective was to be well known; i.e., gaining market share. During those early years, there was little financial information available for specific companies, something that resulted in an extensive use of non-financial information.
Using non-financial information when financial information is not available is not new. Studies on this subject analyze how tangible and intangible assets are associated with share prices and non-market-based value estimates [5]; concentrate on R&D and Capital Markets [26]; focus on assets and stocks prices [23]; focus on value-relevance of non-financial information in the wireless communication industry [4]; and analyze the value-relevance of intangible assets in software capitalization [1].
In the specific case of the dot com business, non-financial information normally takes the form of web metrics. These are obtained from Internet traffic measurement reports. The impact of a dot com firm can be assessed, for example, by looking at the number of “unique visitors” who access the web pages, or by the number of “page hits”. Some empirical studies have concentrated on the relevance of web metrics on share prices of U.S. companies and web traffic [20]; have asked the question of whether the valuation of dot com firms is in any way different from the valuation of firms in any other emerging industry [11]; have constructed a non-parametric production function that contains page views as an output and web pages, scripts, and certain other web constructs as inputs [3]; and have studied the income of chief executives in relation to performance indicators, some of which are based on web traffic [12].
Achieving a high market share is of paramount importance in any emerging industry. Such considerations as cost control, profitability, income, productivity, or cut-off points take second place. Internet firms have been no exception. In fact, in the search for web impact, some firms paid little attention to expenditure; but, even in emerging industries, there must be a limit to how much expenditure a firm can stand in order to establish itself in the market place. Unable to convert this setting up expenditure into actual income, and not being a viable business proposition, many such firms failed starting in the year 2000. The Internet bubble and the way it bust has been extensively studied [6], [13], [14], [21], [34], [35]. Web metrics have been used in managerial practice as a tool for valuing dot com firms [3]. The rational is that one of the objectives of dot com firms is to have an impact in the web. “Unique visitors” is one of the ways in which market share (reach) can be measured. However, the number of times that a web page is visited is not sufficient as an indicator. This has to be seen in the context of the effort made to obtain such visits. This is what the paper tries to do.
This paper concentrates on the study of efficiency in dot com firms. An efficient firm is defined here as one that, which with few resources—personnel, expenditure, infrastructures—obtains high levels of output—to be well known, to obtain high income, and so on. Financial information and web metrics are used as inputs and outputs in the assessment of efficiency. Efficiency can be measured in many ways. For example, taking an engineering point of view, it would be necessary to know the best performance that can theoretically be achieved with the given inputs. In business, there is no theoretical maximum that can be used as a benchmark: efficiency is a comparative concept. The paper estimates efficiency using Data Envelopment Analysis (DEA), a non-parametric approach to the empirical estimation of production functions. For an introduction to DEA, see, for example, [28], [25], [33].
An important decision in DEA modeling is the selection of inputs and outputs that are included in the specification, as different inputs/outputs combinations will produce different efficiency rankings of firms. A particular DMU may or may not be efficient depending on the selection of inputs and outputs. Decision makers may be reluctant to use a technique that is so sensitive to decisions taken at the modeling stage. The paper suggests a new approach to the problem of deciding which inputs and outputs the model should contain. A series of DEA specifications are contemplated, and the resulting efficiency scores are analyzed using multivariate statistical techniques. There is a further problem with DEA: efficiency is a mere score between 0 and 1 (or between 0% and 100%). Two different DMUs may achieve the same DEA score while, at the same time, being very different: they just take a different route to the achievement of efficiency. The approach proposed here makes it possible to go behind the score and explain what is special about the DMU, revealing the strengths and weaknesses associated with each DMU. A further methodological innovation relates to the ranking of DMUs; we propose a new approach based on principal components scores.
As far as model selection is concerned, we just list a series of inputs and outputs that could contribute to efficiency. Clearly, one has to be careful not to ignore anything relevant in the modeling process, but we do not want redundant information either. The methodology proposed here consists of estimating a wide range of possible models, and visualizing the results by means of Principal Components Analysis (PCA). PCA has the advantage of displaying what is similar and what is different about the various models; it is a data reduction technique, which reveals the dimensionality of the data; and it helps to relate models and DMUs in a very powerful way since we can identify similarities, differences and discordant behavior. This methodology is easy to implement in the form of a DSS, as the approach reveals that model selection is a management decision and not a technical aspect of the analysis.
Not all Internet firms are similar [7]. Some make a living out of advertising; others sell products; and others put the customer in contact with services for a share in revenues. The variety of Internet firms is reflected in the different models of e-business that coexist, such as search and portals, communities, and e-tailers. For any given dot com firm, the various DEA specifications produce different efficiency scores. The assessment of the specifications that show a given firm in the best possible light, and how this is related to the type of e-business, is another objective of this paper. The hypothesis that the type of e-business matters is tested in the paper using discriminant analysis.
Two outputs of dot com firms are revenues and unique visitors. When taken separately as outputs, how are efficiencies related? Can it be assumed that a firm that is efficient at obtaining visitors is also efficient at obtaining revenues? How are efficiency in obtaining revenues and efficiency in obtaining visitors related to profitability? The paper tests the hypothesis that they are related.
The paper unfolds as follows. Section 2 concentrates on the analysis of efficiency of dot com firms under various definitions of efficiency. This is followed by an analysis of efficiency scores using multivariate statistical techniques. The relationship between efficiency and profitability, before and after the year 2000 crash, is explored. A concluding section summarizes the findings.
Section snippets
Efficiency in internet firms
The objective of this section is to analyze the efficiency of dot com firms. In Economics, as in Engineering, the efficiency concept is associated with the ability of producing outputs with given quantities of inputs. For a discussion on this topic, see Ref. [17]. This is the subject of the analysis of Production Functions. Production function estimation can be performed using either parametric or non-parametric methods [18], [16]. The non-parametric alternative to the estimation of production
Multivariate analysis of DEA efficiency scores
The contents of Table 2 are visualized using multivariate statistical analysis techniques in order to reveal the full features of the data. Similarities and differences that exist between firms in terms of the 21 DEA models of efficiency will be revealed using Principal Components Analysis (PCA). The reasons for such similarities and differences will be explored using regression analysis. The section will be completed with some hypothesis tests to assess the relationship between the type of
Conclusions
E-commerce has been accused by financial analysts of being a mad world. This is due, at least in part, to its emerging industry nature. Many firms have accepted initial losses in order to position themselves in the Internet market. In fact, before the year 2000 crisis, profitability and revenues were found negatively correlated. This means that non-financial information, as well as financial information, has to be examined in order to judge their performance. In this paper, we have studied one
Acknowledgements
The participation of Cecilio Mar-Molinero in this project was possible thanks to a Ramon y Cajal grant from the Spanish Ministerio de Ciencia y Tecnologı́a. The authors are grateful to an anonymous referee whose comments resulted in many improvements to the original text.
Carlos Serrano-Cinca is a Lecturer in Accounting and Finance at the University of Zaragoza (Spain). He is “Telefonica Professor of Quality in New Networks and Telecommunication Services” at the University of Zaragoza. He coordinates the Centre for Research in E-business at Walqa Technology Park (http://www.ptwalqa.com). He has been Visiting Scholar at the Department of Management at the University of Southampton (United Kingdom). His research interests include: E-business; multivariate
References (35)
- et al.
Value-relevance of nonfinancial information: the wireless communication industry
Journal of Accounting and Economics
(1996) - et al.
Efficiency of financial institutions: international survey and directions for future research
European Journal of Operational Research
(1997) - et al.
Measuring the efficiency of decision making units
European Journal of Operational Research
(1978) - et al.
A survey of frontier production functions and of their relationship to efficiency measurement
Journal of Econometrics
(1980) - et al.
An application procedure for DEA
Omega
(1989) - et al.
A multivariate statistical approach to reducing the number of variables in data envelopment analysis
European Journal of Operational Research
(2003) - et al.
The value-relevance of intangibles: the case of software capitalization
Journal of Accounting Research
(1998) - et al.
Review of ranking methods in the data envelopment analysis context
European Journal of Operational Research
(2002) - et al.
Measuring the efficiency of web site traffic generation
International Journal of Electronic Commerce
(2001) - et al.
Revalued Financial, tangible and intangible assets: associations with share prices and non-market-based value estimates
Journal of Accounting Research
(1998)
Valuation of internet stocks—an IPO perspective
Journal of Accounting Research
The Dark Side of Valuation: Valuing Old Tech., New Tech., and New Economy Companies
The value relevance of revenue for internet firms: does reporting grossed-up or barter revenue make a difference?
Journal of Accounting Research
A rude awakening: internet shakeout in 2000
Review of Accounting Studies
Cited by (129)
Influence of sequential fireballs on thermal safety distance estimations for organic peroxide drums
2022, Journal of Loss Prevention in the Process IndustriesExploring the reasons for efficiency in Spanish airports
2020, Transport PolicySewage treatment process refinement and intensification using multi-criteria decision making approach: A case study
2020, Journal of Water Process EngineeringThe Match Between Structural Attributes and Content-Based Orientation of Managerial Cognition: An Exploratory fsQCA Study of 'Hidden Champions'
2022, Management and Organization Review
Carlos Serrano-Cinca is a Lecturer in Accounting and Finance at the University of Zaragoza (Spain). He is “Telefonica Professor of Quality in New Networks and Telecommunication Services” at the University of Zaragoza. He coordinates the Centre for Research in E-business at Walqa Technology Park (http://www.ptwalqa.com). He has been Visiting Scholar at the Department of Management at the University of Southampton (United Kingdom). His research interests include: E-business; multivariate mathematical models; Intellectual Capital, and Information Technologies in Accounting and Finance. Dr. Serrano-Cinca has published articles in Journals such as The Journal of Forecasting, Decision Support Systems, Omega, The European Journal of Finance, The Journal of the Royal Statistical Society (D), The Journal of Intellectual Capital, etc. His personal web page is http://www.5campus.org/charles.htm.
Yolanda Fuertes-Callén holds a Ph.D and BSc in Business Management from the University of Zaragoza (Spain). She is currently Assistant Lecturer at the Department of Accounting and Finance of the University of Zaragoza. Her Ph.D was on the subject “Usefulness of financial and non-financial information in the analysis of Internet firms”. She has been a research visitor at the Universities of Sheffield (United Kingdom) and Stern at New York (USA). Her personal web page is http://ciberconta.unizar.es/cv/yolandafuertes.htm.
Cecilio Mar-Molinero is a Reader in Operational Research at the University of Southampton (UK) currently on leave as a researcher at the Instituto de Investigación y Control, Polytechnic University of Catalonia. His main interests are: Multidimensional Scaling, Data Envelopment Analysis, Quantitative Models in Accounting and Finance, OR in Education, and Community OR. He has published widely in journals such as the Journal of the Operational Research Society, the European Journal of Operational Research, the European Journal of Finance, Omega, the Journal of the Royal Statistical Society (D), and others.