You are on page 1of 4

The 6th International Conference on Cyber and IT Service Management (CITSM 2018)

Inna Parapat Hotel – Medan, August 7-9, 2018

Data Quality Management Maturity Model: A Case


Study in BPS-Statistics of Kaur Regency, Bengkulu
Province, 2017
Rela Sabtiana Satrio Baskoro Yudhoatmojo Achmad Nizar Hidayanto
Faculty of Computer Science Faculty of Computer Science Faculty of Computer Science
Kampus UI, Depok, 16411 Kampus UI, Depok, 16411 Kampus UI, Depok, 16411
Universitas Indonesia Universitas Indonesia Universitas Indonesia
Jawa Barat, Indonesia Jawa Barat, Indonesia Jawa Barat, Indonesia
rela.sabtiana71@ui.ac.id satrio.baskoro@cs.ui.ac.id nizar@cs.ui.ac.id

offices under BPS-Statistics of the Republic of Indonesia that


Abstract- Data are widely used in an organization not only for has the core duty to conduct statistical researches in the
operations but also for strategic level uses. Poor data quality can regency area by collecting data directly to the respondents.
have negative impact for an organization such as poor decision [5] stated that one of the mission of BPS-Statistics is
making and planning. Therefore, data quality management
“Providing quality statistical data through statistical activities
becomes an issue growing today not only to the academic but also
professional communities. Based on this issue, this paper presents that are integrated and national or international standards.”
and analyzes a case study developed in a governmental agency, Apparently, in the effort to realize the mission, BPS-Statistics
BPS-Statistics of Kaur Regency. For analysis, a data quality of Kaur Regency identified some problems both internal and
maturity model is used to measure the implementation of data external. One of the problem that stated in the document of [6]
quality management in the organization. The results show that was the quality of data generated by BPS-Statistics of Kaur
for the dimension of ‘Data quality expectations’ is at a maturity Regency was not yet optimal. In [7] was stated that the target
of 4.25. ‘Data quality protocol’ is at a maturity of 3.50. ‘Policies’ of the percentage of customers who were satisfied with the data
reaches a maturity of 3.67. ‘Data quality protocol’ and ‘Data quality was 84 percent but [8] stated that the realization until
standard’ are at a maturity of 4.42. ‘Data governance’ is at a
the end of the year was 79 percent. It shows that there is a gap
maturity of 3.00. ‘Technology’ is at a maturity 3.17.
‘Performance management’ is at a maturity of 3.33. However, of six percent between the target and realization. It also gives a
this also implies that implementing these particular dimensions meaning that there is a problem concern about data quality.
will lead to a direct increase in overall maturity. [9] identified four clusters for critical success factors of data
quality: (1) data quality management; (2) people and
Keywords—data quality; data quality management; maturity assessment; (3) environmental and personnel; and (3)
model; data quality maturity model organizational factors. Based on the issue above, in this
research we focus on one of the clusters namely data quality
I. INTRODUCTION management. A maturity model for data quality management is
Data quality management (DQM) becomes an issue to the measured which aims to create evaluation for the organization.
academic and professional communities. It is related to the The model creates points of discussion for organization and
poor data quality that means inaccurate information which may guide organization in making plans for improving data
could harm the organization and waste resources. Mainly it quality management. As we know that data quality maturity
could affect the relationship with its customer [1]. Data quality model has never been measured before. Therefore, from this
is synonymous with information quality since poor data quality study we expect to give a contribution for the organization
results in inaccurate information and poor business specifically to evaluate and improve its data quality
performance [2]. Furthermore, [3] stated two most common management. Generally, we expect this study could be a base
problems caused by poor data quality. The first one is about for other BPS-Statistics offices in Indonesia to evaluate their
extra time required to reconcile data. The second one is about data quality management. Finally, we expect this study could
loss of reliability in the system or application because of make contribution for another research in data quality
existing errors. management.
BPS-Statistics is a governmental agency that has duty to In this research, the following central research question is
treated:
provide data by conducting censuses and surveys [4]. BPS-
Statistics conducts statistical researches up to smaller area.
BPS-Statistics of Kaur Regency is one of the representative
The 6th International Conference on Cyber and IT Service Management (CITSM 2018)
Inna Parapat Hotel – Medan, August 7-9, 2018

“How good is the implementation of a maturity model for B. Data Quality Maturity Model
determining the state of data quality management in the Some different types of maturity models are introduced by
organization?” many researchers right after the introduction of Capability
The remainder of this paper is organized as follows: Section Maturity Model (CMM) proposed by [10], The CMM is about
II presents related works. Section III presents a literature prioritizing improvement effort for organization and how to
review, which is focused on the data quality and data quality continue improvement. The CMM has five different maturity
maturity model. In Section IV, methodology is briefly levels, as it is known that each level determines the
introduced. Section V shows the results and discussions. accomplishment of a different component in a software
Section VI presents conclusions and future works. process. The result is the increasing of organization process
capability.
II. RELATED WORKS Based on the CMM, [11] developed a maturity model. This
maturity model focuses on data quality management. Next,
From the literatures, we found limited studies that [19] developed a maturity model especially for Information
concentrate on the measuring the data quality management Quality Maturity Capability Maturity Model (ICM-CMM)
maturity model. We identified some different way used by the based on [10]. [13] also developed a maturity model patterned
researchers to measure the data quality maturity especially in after CMM but it was for data quality management.
the level of maturity and framework. [10] developed a basis Furthermore, [15] had created a checklist of design
for maturity model. From his study, [11] adopted it became principles which can be used by researchers who are involved
four maturity level for their study of data quality management. in the design of a maturity model. For example, it was found
[12] conducted a research of data quality management for that [14] used the checklist in the development of the maturity
enterprise. [13] developed a framework of data quality model.
management. It consists of five maturity level adopted from Hence, maturity model cannot be separated from maturity
[10]. [14] developed a maturity model adopted from [15] level. Maturity level is about the level reached in the
design principles that consist of five maturity levels for small organization. Therefore, skipping a maturity level is not
firm. After completed the literature studies, we used advised, because each maturity level lays the foundation for
framework by [13] in our research. The reason for using this is achieving the next and is therefore counterproductive [14].
that the framework focuses on data quality management for From the literatures we got a comparison of existing maturity
practitioner. Compare to the other framework, as it has been levels in data quality management developed by researchers as
described above, it is known that [12]’s framework is not shown in Table 1.
relevant to use since it is purposed to evaluate an enterprise
and [14]’s framework is used to evaluate small firm. TABLE I.
A COMPARISON OF EXISTING DATA QUALITY MATURITY LEVEL
III. LITERATURE REVIEW Maturity Paulk et Ryu et al., Loshin,
Kirikoglu, 2017
Level al., 1993 2006 2013
A. Data Quality 1 Initial Initial Initial
Person dependent
Some experts and researchers have defined the concept of and basic
data quality. Therefore, there are many definitions of data Policies, standards
2 Repeatable Defined Repeatable
and procedures
quality. For example, [16] defined the data quality based on
two related factors. The first factor is how well it meets the 3 Defined Managed Defined Defined and stable
expectations of data consumers (how well it is able to serve the 4 Managed Optimized Managed
Managed and
purposes of its intended use or uses) and the second factor is standardized
Continues
how well it represents the objects, events, and concepts it is 5 Optimized - Optimized
improvement
created to represent.
The judgment of data quality depends on the data In this paper, we use the framework developed by [13]. This
consumers was proposed by [17]. They also mentioned some framework consists of eight dimensions to measure the
data quality challenges. First, there are many types of data implementation of data quality maturity model namely (1) data
sources that bring different data types and complex structures, quality expectation, (2) data quality dimensions, (3) policies,
this causes difficulty with data integration between different (4) data quality protocols, (5) governance, (6) data standards,
systems. The second, the amount of data within systems has a (7) technology, and (8) performance management.
lot of volume, therefore it is difficult to judge the quality in a
given time. The third, data within firms change very fast and IV. RESEARCH METHODS
thus the ‘timeliness’ of data is very short. This may cause
outdated or invalid information. A. Questionnaire
In addition, some literatures [2][18] also mention some In order to respond to our research question, an unsupervised
dimensions that could provide an overview of the data quality survey was conducted, which was generated to ask respondents
characteristics such accuracy, completeness, consistency, to give their opinion of data quality issues and maturity model.
currency etc.
The 6th International Conference on Cyber and IT Service Management (CITSM 2018)
Inna Parapat Hotel – Medan, August 7-9, 2018

The questionnaires were divided into two parts. The first The results of our case study are elaborated in this section.
questionnaire was made up of a total of nine questions related From the analysis, we got that the total percentage of
to the data quality from the respondent perspective to identify implemented capabilities versus not yet implemented ones is
the issue in the organization. We used open questions. And the approximately 75% already. From the 111 defined capabilities,
last one, which was a maturity model aspect, we used a BPS-statistics implements 83 (74.8%). Finally, Table 4 shows
questionnaire adopted from [13] to collect data. The an overall tendency to be seen from more implemented
questionnaire was made up of a total of 111 questions related capabilities towards not yet implemented ones throughout the
to data quality to identify the maturity model in the rise of maturity levels.
organization. The questionnaire was divided into eight parts as
TABLE IV.
shown in the Table 2 below. IMPLEMENTED VERSUS MISSING DQM CAPABILITIES PER MATURITY LEVEL
Maturity per level Total Implemented Not yet
TABLE II. implemented
QUESTIONNAIRE CATEGORIES AND NUMBER OF QUESTIONS
Part Questionnaire categories Number of questions 1 20 (100%) 17 (85.0%) 3 (15.0%)
1 Data quality expectations 16 2 22 (100%) 21 (95.5%) 1 (4.5%)
2 Data quality dimensions 12 3 23 (100%) 15 (65.2%) 8 (34.8%)
3 Policies 13 4 28 (100%) 16 (57.1%) 12 (42.9%)
4 Data quality protocols 16 5 18 (100%) 14 (77.8%) 4 (22.2%)
5 Data Governance 14 Total 111 (100%) 83 (74.8%) 28 (25.2%)
6 Data Standards 15
7 Technology 12 Summarizing the results, the single maturity levels are as
shown in Table 5. For the dimension of ‘Data quality
8 Performance management 13
expectations’ is at a maturity of 4.25. ‘Data quality protocol’ is
Total 111 at a maturity of 3.50. ‘Policies’ reaches a maturity of 3.67.
‘Data quality protocol’ and ‘Data standard’ are at a maturity of
We used closed questions. From first to eighth part, the scale 4.42. ‘Data governance’ is at a maturity of 3.00. ‘Technology’
of 0 to 1 was used to answer the questions. The scale was as is at a maturity 3.17. ‘Performance management’ is at a
follows in Table 3. maturity of 3.33. However, this also implies that implementing
these particular dimensions will lead to a direct increase in
TABLE III. overall maturity.
THE SCALE OF ANSWER AND DESCRIPTION
Scale Description TABLE V.
SUMMARY OF MATURITY LEVEL
0 Not yet implemented
Dimensions Maturity score Description
1 Implemented
Data quality expectations 4.25 Managed but Optimized
Data quality dimensions 3.50 Defined but Managed
B. Data Collection Procedure
The questionnaire was given to a heterogeneous subject Policies 3.67 Defined but Managed
matter exist in the organization. There was a total of five Data quality protocols 4.42 Managed but Optimized
subject matters to be the samples of this research, one subject Data governance 3.00 Defined
matter related to data processing and four technical subject
Data standards 4.42 Managed but Optimized
matters. The survey was sent out by e-mail and/or online form.
In both cases, respondents were asked to read the instructions Technology 3.17 Defined but Managed
slowly and carefully. From the questionnaires sent out that Performance management 3.33 Defined but Managed
were answered, a total three questionnaires collected for data
quality issues and one questionnaires collected for maturity
The ‘Data quality protocols’, ‘Data governance’, and
model. A total of one questionnaire was therefore processed
‘Performance management’ value are low because some
and analyzed for measuring maturity model.
capabilities are not implemented yet at BPS-Statistics of Kaur
C. Tool Regency. However, each dimension has its capabilities. The
The data collected was processed and analyzed using result of reality and target from eight dimensions of data
Microsoft Excel spreadsheet. quality maturity model in BPS-Statistics of Kaur Regency is
shown in Fig. 1 below.
V. RESULTS AND DISCUSSIONS
The 6th International Conference on Cyber and IT Service Management (CITSM 2018)
Inna Parapat Hotel – Medan, August 7-9, 2018

financial support with grant number


1891/UN2.R3.1/HKP.05.00/2018.

REFERENCES
[1] A. Lucas, “Corporate Data Quality Management: from Theory to
Practice,” 5th Iber. Conf. Inf. Syst. Technol., 2010.
[2] DAMA International, The DAMA Guide to The Data Management
Body of Knowledge (DAMA-DMBOK Guide). 2009.
[3] W. W. Eckerson, “Data Quality and the Bottom Line,” Data Wareh.
Inst., pp. 1–32, 2002.
[4] BPS-Statistics, “Act of Republic of Indonesia Number 16 Year 1997
about Statistics,” 1997.
Fig. 1. Radar Chart
[5] BPS-Statistics, “Strategic Planning of BPS-Statistics of Kaur
Regency,” 2015.
The three biggest gaps are in ‘Performance management’ [6] BPS-Statistics, “Review of Strategic Planning of Kaur Regency,”
‘Technology’ and ‘Data governance’ dimensions. And the 2017.
smallest gap is in ‘Data quality protocol’ and ‘Data standard’ [7] BPS-Statistics, “Performance Agreement of BPS-Statistics of Kaur
Regency,” 2017.
dimensions. It can be implied that the implementation of both [8] BPS-Statistics, “Performance Report of BPS-Statistics of Kaur
dimensions is already well enough. But it is still lack in data Regency,” 2018.
governance, technology, and performance management. [9] H. Xu, “Data Quality Issues for Accounting Information Systems’
The figure above show that a gap occurs in each dimension. Implementation: Systems, Stakeholders, and Organizational
It means that BPS-Statistics needs to increase all maturity Factors,” J. Technol. Res., pp. 1–11, 2010.
[10] M. C. Paulk, B. Curtis, M. B. Chrissis, and C. V Weber, “Capability
dimension capabilities value to enhance the level of maturity. Maturity Model for Software , Version 1 . 1 Software Engineering
From these results, it can be discussed that the organization Institute,” no. February, 1993.
focused more on data standards, data quality protocols, and [11] K. Ryu, J. Park, and J. Park, “A Data Quality Management Maturity
data quality expectations. However, the organization is still Model,” vol. 28, no. 2, pp. 191–204, 2006.
[12] M. Ofner, B. Otto, and H. Österle, “A Maturity Model for
lack in many areas such as there is no service level agreement Enterprise Data Quality Management,” no. December, 2013.
of data quality, monitoring of data quality validity, [13] D. Loshin, The Practitioner Guide to Data Quality Improvement.
standardized procedures for using data quality tools for data 2011.
quality assessment and improvement in place, reward to the [14] O. Kirikoglu, “A Maturity Model for Improving Data Quality
Management,” 2017.
employees, framework to analyze the business effects, audit [15] J. Pöppelbuß and M. Röglinger, “What Makes a Useful Maturity
related to data quality dimensions, etc. The important thing Model? A Framework of General Design Principles for Maturity
from this result is those lack areas are able to be a base for Models and Its Demonstration in Business Process Management,”
making data quality strategy in order to enhance the data Ecis, vol. 4801, p. Paper28, 2011.
[16] L. Sebastian-coleman, Measuring Data Quality for Ongoing
quality for the organization. Improvement : A Data Quality Assessment Framework Appendix E :
Quality Improvement and Data Quality, no. c. 2013.
VI. CONCLUSIONS AND FUTURE WORKS [17] L. Cai and Y. Zhu, “The Challenges of Data Quality and Data
Quality Assessment in the Big Data Era,” Data Sci. J., vol. 14, pp.
A. Conclusion 1–10, 2015.
The conclusion that can be drawn from this research is the [18] I. Kirchen, D. Schutz, J. Folmer, and B. Vogel-Heuser, “Metrics for
the Evaluation of Data Quality of Signal Data in Industrial
level of maturity of data quality in the BPS-Statistics of Kaur Processes,” 2017 IEEE 15th Int. Conf. Ind. Informatics, pp. 819–
Regency is at level 3.00 to 4.42 for single maturity dimension. 826, 2017.
In order to increase the maturity level, the organization needs [19] Saša Baškarada, “Information Quality Management Capability
to improve all areas but more specifically in Performance Maturity Model,” 2009.
Management, Technology, and Data governance.

B. Limitations and future works


This research had some limitations. First, this research was
only done in one office in one regency. Second, the collecting
data method used questionnaire without direct interview and
observation. Future researchers might explore wider coverage
all over regencies in Indonesia and more respondents. Also,
this kind of research can be conducted in another
governmental institution.

ACKNOWLEDGMENT

We gratefully thank the Universitas Indonesia for the


International Publication Grants for Student Thesis for

You might also like