methods section is used to analyze the relative efficiency of firms among the survey sample. 

Efficiency in this context is defined as the measured satisfaction on a five point Likert scale, 

ranking e commerce impact from  not at all  to  a great deal  and based on the individual set of 

IT and e commerce implementations. A high satisfaction rate based on a certain set of IT 

infrastructure can therefore be defined as  efficient.  An aggregated result of the DEA model is 

provided in Table 4, where the efficient establishments are categorized by their appropriate 

industry and firm size. While one third of all establishments can be regarded as  efficient  in the 

sample, a detailed analysis reveals more SMEs as efficient in the retail/wholesale and 

banking/insurance sectors in comparison to large establishments in these sectors. Even more 

interesting, the retail sector gains more efficiency from e commerce usage than the 

manufacturing or the banking sectors.  

TABLE 4  Percentage of Relative Efficient E Commerce Users by Size and Industry 

 Manufacturing 

Retail/ 

Banking/ 

Total 

Wholesale 

Insurance 

SME 

24.2 47.1 31.4 34.3 

Large  37.1 40.6 24.2 33.0 

Total 

30.9 42.4 27.9 33.7 

Source:CRITO Global E Commerce Survey, 2002;  unweighted sample.

Looking at the differences between  efficient  and  inefficient  establishments per sector reveals 

that  efficient  establishments have not only put into operation an equivalent or higher 

percentage of easy to implement e commerce solutions such as online advertising or online 

sales, but also a significantly higher percentage of more complex solutions such as EDI or 

Internet based supply chain management. The data in Table 4 to Table 7 are the detailed results 

of the conducted DEA analysis. DEA uses a linear programming approach to differentiate the 

efficient from the inefficient establishments which DEA investigated as significant. The 

proposition that the degree of complexity increases (see the directionality reflected by the arrows 

in the following tables below) is based on the assumption that the complexity of technical 

integration together with the necessity of external coordination causes difficulties to increase. On 

the other hand, a real application to application integration reduces manual interruptions and 

media flow interruptions and increases the benefits by automation. This thesis is also true for 

online sales or online procurement. While online sales are, from a technical integration point of 

view, not really difficult to establish   one  only  needs to link a firm's internal system with a 

web server   the online procurement procedure is much more difficult. Online procurement for 

large enterprises and for SMEs are based on automatic replenishment systems generating order 

messages automatically and submitting these by using EDI over the Internet.  

Considering the differences in the usage of e commerce technologies, efficient establishments 

often use more complex solutions, as Tables 5 to 7 indicate. The efficient establishments are 

marked by the  +  sign. 

12