Data Value in Decison Process:
Survey on Decision Support System
in Small and Medium Enterprises
Maurizio Pighin(*) and Anna Marzona (**)
(*) Department of Mathematics and Computer Science University of Udine (Italy)
(**) LiberaMente srl – Udine (Italy)
miproBIS - Business Intelligence Systems - 2012 Opatjia , Slovenia
Agenda
• The economical context on analysis
• Survey targets and methodology • Survey Results
Economical context
• The province of Udine with its 4,905 sq km is about
62% of the territory of Friuli Venezia Giulia • It is the largest province in the region also for:
Concentration of population, with 529,000 inhabitants,
representing 44% of regional total
Number of employees, with 228,000 employees,
representing 44% of regional total
Number of businesses, with 49,500 businesses, 48% of
the regional total.
• High rate of entrepreneurship: one production company every 9.4 inhabitants.
• The companies are mainly small ones (such as considering up to 49 people).
Economical context
• Production specialization,
metal-mechanical
production with 1,600
units (26% of total manufacturing)
woodworking and furniture production
with 2,020 units (33% of total manufacturing).
• Strong propensity to
export
.
European Union with 60% of export value
America (especially Northern) (11% of
exports)
Agenda
• The economical context on analysis • Survey targets and methodology
• Survey Results • Conclusions
Survey targets
• Survey on mechanical companies
represent the trends of the entire territory
heterogeneous in size, incoming, type of
products and processes
• We inquiry
how many companies use data warehouse
systems
what is their profile
Survey targets
• In general we expect that
greater use of data warehousing systems on medium-large size companies
small businesses are less interested in these systems
• not so important amount of data • less computerized processes • low proneness to investment
• low attention to new technologies and innovative practices
companies with data warehouses are those
• with higher technology
Agenda
• Introduction
• The economical context on analysis • Survey targets and methodology
• Survey Results
Profiling companies - Dimension
• Group A: more than 100 employees • Group B: 51 - 100 employees
• Group C: 21 - 50 employees • 38% of companies’ income is
between 5 and 15 million euros
C ompanie s In sample
Group A 16
Group B 14
Group C 15
Profiling companies - Age
C ompany Ave rage start ye ar Ave rage ye ars of activity St. De v. Group A 1961 45 24 Group B 1977 29 13 Group C 1975 31 13 Total 1970 36 18• The average age of companies is about 36 years of activity
companies of group B and C, significantly more recent (29 and 31years)
companies of group A on the market for about 45 years.
Profiling compaines – Export-Quality
• High level of export and foreign relations
• High percentage (75.5%) with quality certification
94% of group A 53% of group B 75,5% of group C
C ompany
% avg.
Export
Group A
65%
Group B
41%
Group C
25%
Total
43%
Profiling IS – Specific function
• group A and B have a function dedicated to the Information System • Group C in 74%: Information System is kept by executives or top managementC ompany % spe cific function for I.S. Avg. numbe r of I.S. staff Group A 100% 2,4 Group B 93% 1,4 Group C 26% 2 Total 73% 2
Profiling IS – Computerized areas
• The areas mainly computerized are Administration, Sales, Purchase, Logistics and Production
• The percentage drops down in the areas of Quality and Control, while still not widely used are the CRM
subsystems.
Are a Group A Group B Group C Ave rage
Accounting 100% 100% 100% 100% Logistics 94% 93% 67% 84% Sales 94% 100% 93% 96% Purchase 100% 100% 87% 96% Production 100% 86% 67% 84% Quality assur. 87% 57% 60% 69% CRM 44% 21% 13% 27%
Profiling DW – Data analysis areas
• The areas most involved in the data analysis are Administration, Sales and Purchase
• Logistics, despite having a high percentage of computerization, is less often the subject of data analysis.
Are a Group A Group B Group C Me dia
Accounting 100% 100% 87% 95% Logistics 69% 43% 33% 49% Sales 88% 93% 100% 93% Purchase 88% 86% 80% 84% Production 88% 71% 73% 78% Quality assur. 75% 36% 47% 53% CRM 31% 21% 6% 20% Control 88% 36% 60% 62%
Profiling DW - Knowledge
• group A: 94% know the existence of DW
• the percentage drops to 50% of companies of group B and 47% of group C C ompany % Knowle dge Group A 94% Group B 50% Group C 47% Total 64%
Profiling DW - Usage
• 24% use DW systems for data analysis
• among the companies that still do not have this tool, 26% will adopt one in the future, and 11% in the short term.
• 20% in group C
orientation of small organizations into decision support systems.
• introduction of DW was fairly new
except some rare cases, DW systems were introduced in the last 3-5 years.
C ompany % Usage % Future usage % Future usage in short te rm Group A 50% 31% 6% Group B 21% 14% 7% Group C 0% 33% 20% Total 24% 26% 11%
Profiling DW – Correlation with export
• The companies that
use DW systems
have the
high
percentage of
export
need to keep under
control the remote
activities
• The initial assumption
is reflected by the
C ompany % Export
Using DW 66%
Profiling DW – Correlation with market
•
High-tech
companies
tend to
adopt innovative tools
• The initial assumption
is reflected by the
survey
Product marke t
% DW usage Ele ctronic and automation 66%
Tools 66%
C ompone nts and subsupply 25%
Mechanic workshop 20% Machinary production 20% Carpentry and assembly 16% Installations -T hird party work -Metal furniture
-Profiling DW – Architecture - source
• The data that flow into the data warehouse comes from
ERP sources (in 100% of cases)
other external sources (73%)
other internal sources (63%) • DW as instrument of data reconciliation Archite cture % C ompany 1 level 80% 2 levels 10% 3 levels 10%
Profiling DW - Supplier
• 80% - DW built by the supplier of the ERP system
• 20% - DW designed by other suppliers or consultants
A single known partner who already knows the company’s information system (better
comprehension of its dynamics and needs) • 88% - one-level architecture in DW built by ERP
supplier
• 50% - two-levels architecture in DW built by specific
consultants
ERP vendors offer solutions for Business
Intelligence, but usually of a lower profile compared
Profiling DW – Kind of tools
• OLAP tools
↔
“drill-down” or “roll-up” features
• Data Mining
↔
simple data analysis package,
like “classification and prediction” or
“association” analysis.
Tool Group A Group B Group C Ave rage
Reporting 100% 100% - 100% OLAP 75% 67% - 73% Data Mining 13% 33% - 18%
Profiling DW – Internal use and investment
• The general trend
global monthly analysis
investigate some small data on a daily basis
• In 90% of cases data is updated daily and automatically
• Budget spent by companies to acquire data warehousing systems is on average between 10,000 € and 20,000 €
• Annual budget for planned maintenance or for any developments of the system is less than 10,000 €
Role of use rs
Group A Group B Group C
Ave rage
Area managers 88% 67% - 82%
Staff 63% 67% - 64%
Profiling DW – Simplicity and Usefulness
• The
simplicity of the analysis
tools used, in a
scale of 0 to 10, has an average answer value
of about
6.5
with a variance quite low (1.25).
This shows a certain uniformity of opinion,
considering fairly simple the analysis tools
available.
• The
usefulness
of these tools found positive
answer with an average value of about
8
on a
Profiling DW - Activation
• The
activation process
of a data warehousing
system
The process is not very simple: the mean value is 5
on a scale of 0 to 10
• Exploring the reasons for this difficulty through
the use of open questions, we found
Determining what information to require
The lack of internal knowledge
• the design is almost exclusively dependent on external consultants or on the same suppliers of ERP
Profiling DW - Motivation
• Almost 60% of companies say they have
been pushed to invest in this direction
to be
more competitive
on the market
the need to have a
single tool
to conduct
analysis and obtaining clear and usable
information.
• Barriers to investment
lack of knowledge
Survey results
Agenda
• The economical context on analysis • Survey targets and methodology
• Survey Results • Conclusions
Conclusions
• Desire to use methods and tools of business intelligence:
amount of data that modern transaction systems
generate
more competitive on the market, taking quick and
appropriate strategic decisions based on fast and complete information
synthetic indicators that allow to monitor corporate
performance and to have crossed and parametric analysis on raw data provided by operational
Conclusions
• Knowledge
the
theoretical foundations
that underlie the
formation of these indicators are fairly
consolidated,
much less are foundational aspects and
engineering skills with which
to build
business intelligence systems
the instruments used are
not always
Conclusions
• In most cases data warehousing systems are made by the ERP vendors,
relationship of trust
• Software companies often push to solve the informational question through their ERP
develop reporting or interactive investigations as
customized ERP functions
use of simple OLAP navigation instruments that read directly the operational database
(one-level-architecture)
• poor knowledge of tools and methodologies of business
intelligence
• attention to operational core business, the ERP system
Conclusions
• Producers of business intelligence tools are
very
oriented to architectural and
technological aspects
, much less to
application and organization
the solutions they propose oversimplify the
collection, cleaning and physical organization of data. • Poor ETL instruments
• One-level-architecture → vertical decay of performance, complexity of user views.
Conclusions
•
Unrealistic
vision of the procedures necessary
for effective DW construction
this kind of solutions relative new
• Innovative methodologies requires
years of
gestation
proposed in formal terms
perceived by the market as a whole
tuned
successfully transposed to the end user (especially the SMEs)
Conclusions
• We can state a profile of the companies that makes use
of data warehousing systems:
mostly medium to large companies
in the market since long time
correlation between the use of the DW and the percentage of
export
• the need for control over foreign operations and the usefulness of a centralized data warehouse is high;
nature of the products may be related to the use of DW
• high-tech companies are more likely (from the cultural
Conclusions
• The
usefulness
of data warehousing tools is
still
not fully understood
in companies
difficulties
to quantify (not only in terms of
money) the ROI
lack
of a specialized figure within the
company
• The
adoption
of these tools is going to
increase
• This evolution must go hand in hand with the
transformation
of corporate culture that must