It is quite difficult to measure the size of the contents of KMS as opposed to mea- sure the size of e.g., relational data base systems. In the case of relational data base systems size is quite easily measured as the number of rows of a table times the number of bytes in every row. The sum total of all tables is the total size of a data base system. Alternatively, the size of the data base file managed by a DBMS or a set of files in the case of distributed data base systems gives the size of the data
Stt.010.Mssv.BKD002ac.email.ninhddtt@edu.gmail.com.vn.
base system. However, a “knowledge base” in most cases consists of a large num- ber of semi-structured files that are dispersed over a number of servers and often also client PCs which not only contain files that are part of the KMS, but also more traditional documents which might also be managed with the help of a KMS. Thus, measuring the size of a KMS is a non-trivial problem and hardly any respondent or interviewee could answer the questions easily, if at all.
Moreover, terminology is still in its infancy, so the simple (sounding) question
“How many knowledge elements does knowledge management handle?” was answered by just nine out of 73 respondents (12.3%). The reported numbers ranged between 0 and 120 knowledge elements with eight organizations reporting up to 10 knowledge elements. These answers show that the question was clearly misunder- stood. Thus, the data obtained in this question cannot be interpreted. The closest proxies of knowledge elements in traditional information systems are documents, files or parts of a hypertext document. Certainly, organizations of the size like the ones in this sample were expected to have many more knowledge elements and telephone calls to several respondents showed that this was the case.
The interviewees also had difficulties to estimate the size of their KMS and par- ticularly the number of knowledge elements and were not sure what types of elec- tronic data107 they should count as knowledge elements. Estimates, though, ranged between tens of thousands up to several millions of knowledge elements. This shows once again that KM is a decentral approach where the contents of KMS are not “managed” in the sense that a central unit is in control of all the contents.
The second measure was storage capacity used by knowledge elements. Respon- dents (and interviewees) in most cases again had problems to estimate this figure, though it seemed that they found it easier to answer this question than the question about the number of knowledge elements. As mentioned above, the difficulties are due to the fact that knowledge elements are spread over a number of servers and even client PCs so that a central unit cannot easily obtain the total size of knowl- edge elements “woven” into a knowledge net. However, fourteen organizations (19.2%) answered the question.
Table C-44 shows the distribution of responding organizations according to the size of contents of KMS. The amount of content managed by a KM unit can be quite considerable. The figures vary widely with a minimum of 25 MB and a max- imum indicated by one organization at 1.5 Terabyte of data managed. The variation in the size of the contents was not correlated with the size of the organization.
Correlation analysis is statistically not possible neither with the number of knowledge elements nor with the size of contents. This is due to the misinterpreta- tion of the question in the case of the number of knowledge elements and due to the small number of organizations that could answer the questions in the case of the size of the contents. Not surprisingly, the correlations with the existence of a sys-
107. In the end, every knowledge element stored electronically is data. The interviewees were uncertain about to count data sources towards KMS, e.g., customer relationship data, contributions in newsgroups, data in data warehouses, experience, lessons learned, best practice data bases, CBT modules, email messages.
Stt.010.Mssv.BKD002ac.email.ninhddtt@edu.gmail.com.vn.
tematic KM initiative were insignificant. Hypothesis 15: ’Organizations with sys- tematic KM handle a larger knowledge base than organizations without such an ini- tiative’ was not supported.
Figure C-24 gives an overview of the types of media used in the responding organizations’ KMS. The numbers in the figure represent the average share of the types of media.
FIGURE C-24. Types of media used in knowledge management systems108
Documents are stored in varying formats109. In most organizations documents were the most important type of format used in KMS. More than four in five orga- nizations (31 out of 38 organizations, 81.6%) had KMS with more than 50% docu-
TABLE C-44. Size of contents of KMS
x = storage capacity used (in MB) frequency percent
x < 1,000 4 28.57
1,000 d x < 10,000 3 21.43 10,000 d x < 100,000 4 28.57
x t100,000 3 21.43
total 14 100.00
108. Means and ranges show percentages, 37 d n d 38.
109. See section 7.2.3 - “Size and media used” on page 296 for a brief description of docu- ment formats as well as the formats discussed in the following categories and for links to the literature.
Mean Std. dev.
74.65 12.46
6.27 3.96 3.08
0 20 40 60 80 100
documents data base elements multimedia contents
contributions in newsgroups
other media
Stt.010.Mssv.BKD002ac.email.ninhddtt@edu.gmail.com.vn.
ments and in about two thirds of the organizations (63.2%) documents made up 80 or more percent of their KMS with four organizations (10.5%) having KMS made up of 100% documents. Four organizations (10.5%) had 30 or less percent of docu- ments in their KMS.
Data base elements, are stored in conventional data base systems (e.g., hierar- chical, relational, object-oriented DBMS) and data warehouses. About a third of the organizations did not consider data base elements as part of their KMS (12 out of 37 organizations, 32.4%). Seven organizations (= 18.9%) had less than 10 per- cent, five organizations (= 13.5%) had between 10 and 20% and 13 organizations (= 35.1%) had a share of data base elements in their KMS of 20 or more percent with a maximum of 70%. The APQC reported that 77.3% of the innovative organi- zations questioned in 1996 used data bases for “institutional knowledge” (APQC 1996, 55) which is higher than the 67.6% found here.
Multimedia contents could be audio files, video files, vector graphs or pictures.
All these multimedia contents could also be part of hypertext documents. Almost half of the organizations had no multimedia contents in their KMS at all (16 out of 37 organizations, 43.2%). 15 organizations (40.5%) had up to 10 percent and six organizations (16.2%) had more than 10 percent of multimedia contents in their KMS with a maximum of 30 percent.
Almost two thirds of the responding organizations did not handle any contribu- tions to newsgroups within their KMS (24 out of 37 organizations, 64.9%). Ten organizations had 10 or less percent, two had 20% and one organization reported a share of contributions to newsgroups of 60%.
Five out of 37 organizations (= 13.5%) reported the use of other media. The per- centages ranged between 4 and 64.8%. Only one organization specified the type of other media. This organization additionally handled email messages and other internal electronic communication in their KMS.
It seems that most organizations still do not pay a lot of attention to the more interactive side of knowledge management. Contents of KMS are mostly focused on the explicit, descriptive side with documents representing the lion’s share and data base elements ranking second. Multimedia elements and contributions to newsgroups are less commonly found in KMS.
Organizations with a systematic KM initiative differ, though, from organiza- tions without such an initiative with respect to the types of media used in their knowledge bases. Having a KM initiative, organizations on average have a signifi- cantly higher share of data base elements, multimedia elements and contributions to newsgroups combined (Spearman’s rho: -0.413, significance: 0.011, n=37110) and consequently a significantly lower share of documents (Spearman’s rho: 0.321, significance: 0.049, n=38) than organizations without such an initiative. Thus, Hypothesis 16: ’Organizations with systematic KM handle a higher share of multi-
110. The negative sign means there is a positive correlation between having a KM initiative and the share of the type of media tested.
Stt.010.Mssv.BKD002ac.email.ninhddtt@edu.gmail.com.vn.
media elements, contributions to newsgroups and data base elements in their KMS than organizations without such an initiative’ was supported.
A detailed analysis reveals that the relationship is strong with respect to data base elements and insignificant with respect to contributions to newsgroups and multimedia elements. The latter even shows a slightly lower share for organiza- tions with a KM initiative. Table C-45 compares the means of organizations with and without a systematic KM initiative with respect to the types of media used.