Major changes are under way at Statistics New Zealand that will ultimately integrate data to provide a much wider picture for users than provided by the traditional statistical model. Over the past two years, the department has been developing a range of strategies that include statistical directions, a generic business model, and information strategy. CIO Graeme Osborne, who joined Statistics three and a half years ago, has had his title broadened to group manager, Information and Business Services. He is, in essence, across all of the changes. He has five business units reporting to him: integrated data collection; product development and publishing; information customer cervices; information management systems; and business solutions.
In the past, users have had to rely on statistical methodologies that sometimes took two and three years to produce results. “When I joined the department, I was told why everyone was different and why we needed these systems,” he says. “There was a considerable cost maintaining separate statistical collections built up over the years. We had created too many unique solutions that were too expensive to maintain.”
For example, there are 22 different surveys for business, and 11 different systems to manage them. This raised performance issues around size for technology.
Osborne says statisticians liked to own their own systems and were able to individually get the funding to improve those systems. He undertook some high-process modelling. “It didn’t take me long to identify that the organisation was in a whole lot of silos.” Protection of patches by individuals was also an issue.
“Our chief executive, Brian Pink, backed us becoming a world-leading statistical organisation with an overall view of information life-cycle management. I introduced the concept of business process modelling, which was well accepted. It gives us an over-arching view and drills down so we can dismantle the silos.”
He took a generic approach to the process modeling, rather than using a branded methodology. “I chose a generic approach because it could be made to fit this organisation. It’s good to show how we can transform data and is a very clear way of understanding how to create a data set.”
Such changes don’t come without considerable cost. It was, therefore, a timely coincidence that the government had set aside appropriate funding over three years for Statistics New Zealand to introduce six new strategic programmes.
X 2006 census, which will be electronic.
X National accounts re-engineering.
X Business performance.
X Social statistics.
X LEED (linked employer/employee dataset), which will integrate agency data to understand the dynamics of the labour market.
X Growth and innovation process to understand growth in ICT, biotechnology and design.
“These were in gestation while I was trying to think of the future of this organisation,” Osborne says. “In October, 2003, our senior executives decided we should run a new programme by using the six government requirements to create a new platform. We have designed one platform for the six; otherwise, we would have had six platforms.”
The initial approach was to write up a charter road-map that described the benefits and how to know if they were successful. Financial outcomes over three years of the business transformation strategy include:
X A reduction in the operating cost to produce a statistical output (currently operating on a separate subject matter system) by between 10-20%.
X A reduction of 50% in the investment (of time and money) required to implement the end to end processes and systems required for a new statistical output.
Over the past three years, the numbers of surveys of small to medium businesses has been reduced by 30% by using tax data. Statistics has signed up to Inland Revenue rules so that there is no breach of privacy – all addresses and names are stripped out.)
Osborne says he thinks that all data should be managed in the same way. “Every time we try to add value, we transform the data. If we understand all the steps that are applied, we can re-apply them if we need to.” He says this approach has not been hard to sell internally at a conceptual level but that the statisticians have taken time to decide if it is acceptable.
“We are in the thick of design and build now. We’re using a business architecture team led by a statistician with international experience. The team produces white papers and oversees what is created.”
Because of the six major government programmes, a project management office was created nine months ago. It had 30 projects to deal with from day one. Statistics is effectively going through a rebirth. Several statisticians have been brought on board from other parts of the government sector. There has also been reorganisation of management.
“Statisticians are generally not the strongest leaders of people,” Osborne says. “The organisation has focused on recruiting stronger leaders and allowing specialist statisticians to focus on their strengths.”
Treasury had asked Statistics the cost of introducing the new programmes. “We’ve steered away from giving them a number because we wanted them to focus on the outcomes. We expect business transformation services to give productivity savings over the six years following the three-year project to introduce the new programmes.
The Information and Business Services group was formed by merging several separate service units over the past year. Its role is to provide internal and external clients with end to end information services and transparent and effective statistical processes and systems, from collection through to publishing.
It has begun to introduce a life-cycle approach to the way it manages people, processes, information and technology systems. During this first year of a three-year strategy, it is creating a new business model based around the successful completion of the six major statistical programmes.
Designs for the core components of the new platform, with the first versions of components developed for implementation, are scheduled to be completed in the 2005-2006 year. The information architecture will change by replacing a range of data stores by subject area (silos) with summary input and output data stores, allowing the integration of data. The end to end platform, with associated metadata, is scheduled to be delivered over the three years in tandem with the six new statistical programmes.
The platform is based on 10 components: input data store; output data store; metadata store; analytical environment; information portal; transformations; respondent management; customer management; reference data stores; and workflow.
Osborne says the model is a very important communication “translator” for statisticians, business people and IT professionals to understand what is being created. He also stresses the importance of the link between a generic business model and the underlying information life-cycle, enabled by the processes and system.
Statistics has a long horizon. Further out, at four to six years, new work and migration of current statistical programmes is planned, and more of the same at seven to nine years.
Commercially, Statistics New Zealand generates around $5 million of income a year. Most of the buyers are New Zealand-based (though they may be banks owned in Australia.} The information provided is “standard” with some bespoke projects undertaken, which involves additional work.
The department also has a responsibility to provide data to international organisations such as the United Nations and the Organisation for Economic Cooperation and Development to enable international comparisons.
Osborne says he wants the commercial side of the business to grow slowly. “We compare ourselves against Commonwealth and OECD nations (statistical agencies),” he says. Partly, this is done by sharing white papers for feedback.
“Because we are smaller, with fewer resources, we are arguably stepping into areas others aren’t asked to get into. Some can’t get the buy-in to make the sorts of changes we are making. Research shows us to be in a group of five or six investing in simpler statistical models, though some are doing it in different ways.
At a technology level, Statistics New Zealand is one of the biggest Lotus Notes users in New Zealand. Notes will remain as its collaborative tool. “In the 1990s, we went client-server, running Centura on Sybase. Our architecture is moving more closely to Microsoft, and we are using .Net in our new projects,” Osborne says. The glue and the framework of the new platform will be written in .Net. “We’re moving down the workflow track, managing data as an asset in a data warehouse This is the biggest information factory I’ve been involved with. We collect 60 million items of information a year across 200 different subject areas. “We publish important statistics every second day at 10.45am.”
20 years in IT
Graeme Osborne, aged 40, has been in the IT industry for 20 years. After university, where he specialised in pure mathematics and marketing, he worked in the marketing department at National Mutual as an actuarial statistician, under Theresa Gattung, now Telecom chief executive.
He had stints at the old Government Computing Service, then QED Software, after which he joined Cigna Insurance. Regular travel to Asia became a burden on the family and he joined Southern Cross as CIO before moving to Statistics New Zealand.
He has been chairman of the Telecommunications Association of New Zealand (Tuanz) for the past two years, and sits on the government electronic standards management committee.
On the move
At the end of the year, Statistics New Zealand will relocate to a purpose-built building on the Wellington waterfront, opposite the Westpac stadium.
The lease is up on its current premises, in Molesworth Street, where it has been for 25 years. These will be turned into apartments.
Role of Statistics
The Statistics Act 1975 sets out the Government Statistician's role in, and responsibilities for, all official statistics whether produced by Statistics New Zealand or by other government departments.
The functions of Statistics New Zealand can be summarised as follows:
X To collect, compile, analyse, abstract, and publish with or without comment, statistics on New Zealand economic, financial, production, environmental, and social matters, making or constructing such estimates, forecasts, projections, and statistical models as the Government Statistician may consider necessary
X To advise the Minister of Statistics on statistical policy matters and to keep the minister informed of the nature, scope, purpose, and usefulness of existing or proposed statistical projects of all government departments
X To define, lay down, and promote standard concepts, procedures, definitions, and classifications for use in official statistics, and to advise other government departments on the suitability of statistical projects initiated or carried on by them
X To examine proposals (other than those covered by ministerial waiver) by government departments to commence or commission any new statistical survey or to make a substantial alteration to an existing survey, then, in consultation with the department concerned, to prepare a submission to the Minister of Statistics seeking approval for the new or revised survey to be undertaken
X To review from time to time the collection, compilation, analysis, abstraction and publication of official statistics prepared by the department and other government departments
X To convene a meeting of users of statistics at least once every five years, in order to review user needs
X To carry out investigations and provide certificates as may be required by any other Act.
X To maintain general liaison with international organisations such as the various United Nations agencies on statistical matters, and to provide, where appropriate, technical advice and assistance and statistical training facilities to developing nations, particularly in the Pacific area.
Join the CIO New Zealand group on LinkedIn. The group is open to CIOs, IT Directors, COOs, CTOs and senior IT managers.