Modernisation of Statistics
Production
Stockholm November 2009
Summary and Conclusions
New York 24 February 2010
Mats Wadman
Deputy Director General
Statistics Sweden
Reasons for organizing a conference
on Modernisation of Statistics
Production
• We all face an increased demand for new accurate
and timely data.
• At the same time pressure to reduce costs.
• There is a broad understanding that we have to focus
our efforts to enhance efficiency, cost effectiveness,
and quality.
• That a key aspect of this is to successfully
standardise the production process.
• Need for more improved collaboration between
agencies.
• Striving for a uniform perception of quality.
The topics of the sessions
1.
2.
3.
4.
5.
The role of the customer in statistics production.
Quality assurance systems.
Service-oriented architectures.
Management of processes and resources.
Collection and use of paradata in statistics
production.
6. Innovations in data collection and design.
7. Optimal use of registers and administrative data.
8. Generic business models and standardised
processes.
Main findings from
conference
Standardisation
of processes
Learning
and
sharing
Organizational
issues
Leadership
and culture
Customer
orientation
Quality
Organizational issues
• Knowing what users need.
• Striving for excellence - Identify world-class
approaches.
• Adopting the process view. Henry Ford : “We will not
put into our establishment anything that is useless.”
• Developing standards and operating procedures.
• Creating a culture of continuous improvement.
• Check process variation by means of paradata and
control charts.
Customer orientation (1)
Our statistics must be based on the needs of the
customers.
• Customers demand more and more tailor-made
statistics and quick information about upcoming
trends that require a large degree of adaptability
from the NSIs.
• If NSIs are unable or slow in reacting to new
demands, someone else will react instead.
Competition in the information market is quickly
increasing.
• NSI:s must be able to respond to new needs of
customers, but still handle continuity.
• NSIs must focus on the customer needs without
letting the image of the NSI be damaged.
• Maintaining the scientific approach.
Customer orientation (2)
• Not everything is possible to produce, NSI must be
able to explain the restrictions without losing the
confidence of the customers.
• Build trust in statistics.
• NSIs must also consider to provide more
accompanying analysis to increase the value of the
data. But maintaining the objective and impartial
perspective.
• We need to reduce response burden e.g. by re-use
survey data, use administrative data.
Customer orientation (3)
• Seek partners that can help make statistics
available using new media.
• I.e. Intermediaries such as Google or
Bloomberg can help in meeting customers
needs, using NSI data and making them
available in new forms.
• Developing reports with users.
Customer orientation (4)
• Metadata needs to be available for customers to
judge the quality of statistics and the available
outputs.
• Metadata is important in a language and format that
is understood by customers. NSIs have not yet come
far in describing “fitness for purpose” and this is an
area where more research is needed.
• Find out how users work with information on quality.
The importance of using quality
frameworks
• To get good product quality we need to consider the
way we do thing i.e., process quality.
• To achieve good product quality we need to consider
organisational issues such as leadership and
strategic planning i.e., organizational quality.
• We need to consider the way we carry out our
production process since it will affect product quality.
– quality frameworks such as ISO 9001 or EFQM
can help us work in a systematic way with
improvements.
Leadership and culture
• Strong leadership and strong commitment in
leadership is important.
– targets must be measurable.
– Management must also engage in the follow-up of
implementation.
• A general agreement that a step by step approach is
preferable to balance change against the ongoing
activities of the organisation. NO BIG BANG.
• Standards and guidelines is absolutely essential or
much resources are wasted.
Leadership and culture (2)
• It seems that culture in statistical agencies often is an
obstacle to change.
• Involvement of the staff is absolutely essential.
• Top management must be able to explain changes in
a language that is understood across the
organisation in order for change to gain momentum.
• Good examples of implementation of new methods
and approaches in statistics production to justify
further approaches. Picking the right projects to start
with is a strategic decision.
• Follow-up and evaluate.
Learning and sharing (1)
NSIs can and must learn from each other
• Technology is developing fast. One NSI can't cover
all aspects.
• We all face an increased demand for new data. At
the same time pressure to reduce costs.
• Agencies are not competitors and Most problems are
shared.
• Cooperation and sharing work is the only way that
NSIs will be able to keep up with this development
and make efficient use of the new possibilities. Can
we do our IT-systems more homogeneous?
• Benchmarking is a good alternative to cooperation.
• Work in smaller constellations on urgent matters that
people care about and want to engage in.
Learning and sharing (2)
• Nothing can be successfully compared unless the
information, e.g. paradata, is standardised (much can
be improved)
• There are too many areas of cooperation, focusing is
key.
• One important area of cooperating is training and
development of competence.
• The use of administrative data especially between
countries necessitates a large degree of
standardisation to ensure comparability.
Learning and sharing (3)
• A lack of methodology for use of
administrative data.
• What the are the roles of NSIs and
universities.
• NSIs have to take the lead in developing
methodology and effective use of
administrative data.
Processes
• Develop standardized processes.
• Define key quality indicators.
• Check process variation by means of
paradata and control charts.
• Improve processes when needed e.g.,
selective editing.
• Build paradata data banks.
To sum up
There are strong links between
standardisation of processes, learning
and sharing, organisational culture and
leadership, customer orientation, and
quality.
Use a step by step approach. NO big bang.
Thank you