ABS Business Architecture - Statistical Production Activity Model (SPAM) Version 0.3 1. Introduction This paper is intended as a covering paper for the Phase 1 Statistical Production component of an ABS Business and Capability Architecture project. The paper aims to articulate business activities in scope of this line of business, in the form of a more detailed model, namely the Statistical Production Activity Model (SPAM). Complementary Phase 1 work on developing a Statistical Capability (previously called Statistical Support) Activity Model (SCAM) is also being undertaken in parallel with this model. Once the first cuts of the models have been discussed with key stakeholder groups and further refined, they will form key inputs to the Phase 2 capability identification and specification work. 2 The primary objective of the project is to input to preparation of an ABS New Policy Proposal (NPP) seeking government funding for the ABS2017 change program. The business case for the second pass process needs to be completed by end August 2013, and the outputs from the project are expected to provide a framework that will support planning and prioritizing investments for the development of capabilities needed to sustain the ABS into the future. 3 There is an expectation that a 'draft' set of lower level business activities will be of sufficient quality for the Phase 2 capability identification and assessment work needed to support the NPP. This means that in the short term, the project will not be aiming to deliver a more robust model that could be readily implemented across the ABS, and which could better support the international collaboration projects. As such, outputs from the project should only be used as reference material to support: refinement of the ABS's Business Architecture (potentially replacing the Production component down the track) ABS2017 business re-engineering activity the international Statistical Network project on Business Architecture the international HLG-BAS project developing a Common Statistical Production Architecture the international review of the GSBPM (expected during 2013). 4 Two drafts of the Model (v0.1 and v0.2) have been discussed, initially with an expert group and later with a reconstituted EA Reference Group and a few interested business areas. However, the small Enterprise Architecture (EA) team is not resourced to further consult and refine the model, as it must move quickly to deliver on the phases of the project addressing the NPP capability need. The model therefore won't have been through sufficient consultation and refinement for widespread implementation and 'production' use, until further funding is made available. 5 Section 2 of the paper provides an overview of the Business Architecture signed off by the ABS2017 Board in October 2012. Section 3 provides some context for the ABS Statistical Production Activity Model. Sections 4 and 5 respectively outline a few related, internationally agreed models, and some of the main structural changes between the newly proposed Model and related models. Sections 6 and 7 contain links to a diagram and descriptions of level 1, 2, and 3 activities in the Model. Section 8 mentions possible uses of the Model in the ABS, and Section 9 broadly outlines next steps for the Statistical Production elements of the project. Australian Bureau of Statistics May 2013 2. Background 6 In October 2012, the ABS's Enterprise Architecture (EA) project produced a set of high level artefacts which were endorsed by a number of ABS2017 governance fora (see October 2012 Board presentation). These artefacts include an articulation of the ABS's 'To Be' World, EA Decision and Design Principles, a high level EA Conceptual Model (the cube), and an ABS Business Model. 7 As well as advancing the EA high level artefacts, a description of the Business Architecture domain was developed, covering the main activities undertaken in the ABS. These activities were divided into four discrete lines of business, as shown in the front facing part of the cube artefact: Statistical Leadership provides an overview of the high level strategic, externally focussed activities undertaken by the ABS (e.g. Statistical Coordination, Statistical Programme Management, International Collaboration) Statistical Support (to be renamed Statistical Capability) provides an overview of the activities supporting the successful development and management of capabilities underpinning statistical production processes (e.g. Manage Statistical Capability Library, Design Methods, Standards and Frameworks) Statistical Production provides an overview of the activities covering statistical design, acquisition, compilation and dissemination of statistical content Corporate Support (to be renamed Corporate Capability) provides an overview of the crosscutting ABS support activities that help the ABS to effectively deliver on its work program (e.g. HR, Finance, Administrative Management, ICT Management). 8 Each of these lines of business was then broken down into further layers of discrete business activities (see artefact 5). For the Statistical Production line of business, these activities were made identical to those described in the Generic Statistical Business Process Model (GSBPM) version 4.0. Australian Bureau of Statistics May 2013 9 At the same time, the ABS 2017 Program Board endorsed further development work on the Business and Capability architecture layers of the EA (see Appendix 1 for an overview of how the different Architectures link together). The ABS 2017 Program Board acknowledged the practical and strategic value of the EA work, noting it was useful for organising what the ABS does, supporting discussion of priorities, assessing investments against principles, and choosing where to deploy our people. A project proposal was then put forward, and endorsed by the Strategic Design Committee (SDC) in December 2012, to advance development of the Business and Capability architecture layers within the EA framework. The ABS's Senior Management Group agreed in principle that Resource Centres would provide input to the development work, but noted that it would be difficult to free up staff at short notice, and consultation would need to be kept to a minimum. To minimise the impact on Resource Centres, the EA team are developing the first cut of the activity models without an initial round of consultation, choosing instead to consult when there is a 'concrete' proposal on the table for people to react to. 3. Context for the Model 10 This section of the paper provides an overview of the types of activities in scope of the SPAM, some of the criteria used in designing the Model, and the structure of the Model. Applicability 11 The Model is generic enough to apply to all activities undertaken by the producers of official statistics within the international Statistical Network in their Statistical Production line of business, which result in data outputs. The Model focuses on the activities expected to be undertaken by statistical organizations in the future ('to-be' environment), and is expected be stable for at least the next 5 years, rather than be changed regularly. There are a wide range of statistical projects underway in the ABS, and the SPAM is expected to apply the great majority of these projects. Note that some of these projects produce outputs which are themselves inputs to other projects (e.g. macro level sample survey and administrative outputs being input into a National Accounts compilation project). Applicable types of projects include: Producing statistical outputs by combining data from a range of sources (e.g. administrative records, sample surveys, censuses, and other non-statistical or mixed sources) Acquiring and validating administrative or transactional data records for a range of other projects Producing sample survey outputs Compiling accounts and price indices Compiling other special statistical outputs (e.g. Measuring Australia's Progress) Producing ongoing or regular statistical outputs, with any frequency (e.g. hourly, daily, weekly, monthly, quarterly, annually, adhoc, preliminary vs final, etc.) Developing new statistical outputs (to meet emerging needs) Small or large information consultancy or client servicing jobs producing statistics (e.g. a NIRS information request) Large scale testing for statistical collections (e.g. for the Population Census) Australian Bureau of Statistics May 2013 Pilot testing a new statistical capability in a particular production project (e.g. introduction of a new selective editing methodology). 12 Like the GSBPM, some projects would utilize all high level business activities in the Model, while others would only utilize a subset of high level activities or sub-activities. Some activities may be more relevant for one type of project than another, which may be influenced by the types of data sources used or the outputs to be produced. The Model should therefore be applied and interpreted flexibly. It is not intended to be a rigid framework in which all steps must be followed in a strict order, but rather a Model that identifies the activities or steps in a statistical business process. Although the presentation follows the logical sequence of steps in most statistical business processes, when mapping processes, the activities in the Model may be used in different orders in different circumstances, and be iterative in nature. In this way the Model aims to be sufficiently generic to be widely applicable, and to encourage a standard view of the statistical business process, without becoming either too restrictive or too abstract and theoretical. 13 A range of statistical products are expected to be disseminated internally and/or externally from statistical production projects. Examples of these products are data cubes, relational data tables, datasets, data visualizations (e.g. graphs, maps, tables), Confidentialised Unit Record Files (CURFs), analytical/review reports, analytical articles or stories, publications, explanatory and technical notes, media releases, blogs, vodcasts, quality metrics, marketing material, and metadata accompanying data products. Note that dissemination tools and services would be shared corporate resources, and therefore delivered via other lines of business (ICT component from Corporate Capability and metadata standards component from Statistical Capability). Design rationale 14 The model or framework of business activities in the Statistical Production line of business was developed with the following aims in mind: activity names, scope and descriptions come across as intuitive to business people include all of the activities required to undertake a specific iteration of an end to end statistical process or statistical project should describe all activities by which the ABS produces statistical data outputs (i.e. should not be limited to sample survey based production) activities are independent of the type of data source, data collection channel (e.g. internet, mail, phone, physical presence), or mode of collection (e.g. self-administered, person/phone interview, web queries, file transfer, etc.) activities are independent of organisational structure, as roles and responsibilities can change over time, and may be different in other organisations at the higher levels, activities are designed to be discrete and non-overlapping with any other activities in the Business Architecture, unlike the capabilities which can map to multiple activities, or be split across activities activities are designed to align with the ABS's Enterprise Architecture design principles (see artefact 3), particularly those relating to re-use and assembly, and being described by metadata cover only those activities which are project specific as any activities relating to the development, implementation and maintenance of shared resources are covered in the activity models for Statistical Capability (e.g. planning, development and management of registers, frames, standards, frameworks, etc.), and Corporate Capability (e.g. building and maintenance Australian Bureau of Statistics May 2013 of ICT systems and services, etc.). Only the use of these shared resources is in scope of a statistical production related project (e.g. specifying metadata in a user interface to extract a survey frame from the common frame, and generating the survey frame, feeding back frame and sample updates, etc.). 15 In designing the Model, a wide range of material was referenced (see Appendix 2). In particular, ABS outputs from the Business Collections, Accounts, Household Collections and Census 'as-is' business analyses, and business process mapping work from the Acquire, PSIP, Economic Accounts, and Admin Data Capabilities projects, were used extensively when thinking through, designing, and describing the activities the ABS needs to support in the future. Structure 16 The SPAM has been designed to support many levels, however this draft version stops at level 3. These levels are shown in the diagram below. 17 Further levels of detail are expected to be appropriate for certain broad activities, but only where these activities are sufficiently generic to be included in the Model. The intention is to stop just before the activity name and description become specific to certain types of data, outputs, products, or projects. A diagram showing three levels of business activities within the SPAM is included in Section 6. These activities are described in more detail in Section 7. Complementary ABS activity models 18 In addition to the Statistical Production Activity Model (SPAM), there will be activity models for the other three lines of business in the ABS's Business Architecture outlined in paragraph 7. In the short term, there is no capacity to develop the Statistical Leadership and Corporate Capability models any further than the existing two levels. Statistical Leadership Activity Model (SLAM) Statistical Capability Activity Model (SCAM) Corporate Capability Activity Model (CCAM). Australian Bureau of Statistics May 2013 19 Projects responsible for planning, building and maintaining generic ABS statistical capabilities are covered by the SCAM. The SPAM can, however, be used as a generic basis for developing these shared capability elements (e.g. methods, standards & frameworks, ICT systems). 4. Related international models 20 The modernization of statistical production is needed for statistical organizations to remain relevant and flexible in a dynamic and competitive information environment, particularly one with growing demands for data in our increasingly digital societies. Across the world, statistical organizations undertake similar activities, albeit with variations in the processes each uses. Each of these activities use and produce similar information. In recent years, the United Nations Economic Commission for Europe (UNECE) have released a number of frameworks, two of which are the GSIM and the GSBPM, and both are aimed at modernizing statistical production and adopting a common language across the international statistical community. Both of these frameworks sit within the Information Architecture domain of the ABS's EA (as do all standards and frameworks). GSIM 21 GSIM v1.0, released in December 2012, is the first internationally endorsed reference framework for statistical information. GSIM defines and describes the pieces of information about the real world (called information objects) that are important to statistical organizations. Examples include data and metadata (such as classifications) as well as the rules and parameters needed for production processes to run (for example, editing rules). It also gives users the relationship between information objects. GSIM has been designed to work in conjunction with the GSBPM. It supports GSBPM and covers the whole statistical process. GSBPM 22 The current version of the GSBPM (v4.0) dates back to 2009. Since that time, international statistical organizations have made efforts to put the GSBPM into practice as much as possible, particularly those organizations undergoing major change programs. The GSBPM is currently a subset of an organization's Business Architecture, and can be tailored to fit within this Architecture. In applying the GSBPM within the ABS, a number of issues around its use have been identified: the perception that the GSBPM does not cover the ‘entire statistical programme of work’, losing many of the strategic elements of statistical production with its strong focus on sample surveys friendly wording similarly, it’s hard for some areas to clearly picture all of their statistical compilation activities, including processes based on administrative data, account compilation and censuses the distinction between a number of the phases and sub-processes is not sufficiently clear the GSBPM doesn't adequately support the ABS2017 vision for an assemble to order, Service Oriented Architecture environment, both at the high level and in the level of detail required to support the environment in practice. 23 It is expected that during 2013 there will be an international review process for GSBPM based on implementation experiences with GSBPM to date, the need to more clearly support "non survey based" statistical production (including harnessing new sources of data), and the need to align with the more recently developed Generic Statistical Information Model (GSIM) in terms of Australian Bureau of Statistics May 2013 concepts and terms related to information flowing through statistical production processes. 2013 is therefore expected to present the first major opportunity since 2009 for strong, practically oriented feedback and suggestions from organizations to shape the terms and concepts used in the next evolution of GSBPM as an internationally agreed framework. 24 The ABS Transitional Metadata Model (ATMM) is a collection of metadata and business objects that represent core components of statistical activities within the ABS. Each ATMM object is a blueprint that defines what the object is and what information is used to register and understand the object. This allows the object to be discoverable and reusable in the Metadata Registry and Repository. The ATMM maintains its alignment with international standards through its relationship with GSIM, and is, to an extent, the ABS's operational version of GSIM. 25 The SPAM is intended to eventually be the ABS's operational version of the GSBPM, similar to the role the ATMM is expected to play with the internationally agreed GSIM. While SPAM's development strives to be consistent with the GSBPM as much as possible, it needs to deviate somewhat for it to adequately reflect the ABS's 'to-be' future operational environment, and to address the shortcomings listed in paragraph 22. That said, mappings between SThe upcoming GSBPM review is not expected to radically change the model, as it must continue to support processes in developing countries, as well as those countries at the cutting edge of change (e.g. countries part of the Statistical Network). What is really needed internationally are common Business Architecture models, rather than a model just for statistical production activities, which only partially cover the operations of statistical agencies. 5. Key structural changes from earlier models 26 Two almost identical reference models have been recently used in the ABS as frameworks for statistical production activities, the older GSBPM, and the high level ABS Business Architecture developed in 2012. The table below summarises the level 1 name and scope changes between these earlier models, and the SPAM being proposed in this paper. The nine high level phases and two overarching statistical processes in the GSBPM have been reduced to eight high level (level 1 activities), and one overarching Project Management activity in the revised ABS model. See Section 7 for more detail around changes to lower level activities. Australian Bureau of Statistics May 2013 Earlier models (level 1) 1 Scope SPAM v0.2 (level 1) 1 Specify Needs 2 Design 2 Design 3 Build 3 Assemble 4 Collect 4 Collect 5 Process 5 Compile 6 Analyse 6 Validate Australian Bureau of Statistics Changes A key change in the scope of this activity is the moving of business case preparation work to the overarching 9 Manage production project activity. This change allows more flexibility to secure funding and negotiate formal agreements for data provision after some of the more detailed design work has been done. Although there is still some specification of needs early in this activity, its main purpose is to do enough scoping work to prepare a business case for a project to address these needs. The name of the activity has therefore been changed to 'Scope' to reflect this purpose. The focus of the Design activity has been extended slightly to include the building, testing and finalizing of data collection instruments (previously in Build 3.1). This activity is closely integrated with the specification of instrument metadata, and the development of the data collection strategy. It also no longer fits within the new scope of the Assemble activity, which is only to assemble production processes from pre-existing capabilities. The name of this high level activity has been changed from Build to Assemble to reflect the significantly reduced scope of this activity, and future aspiration to assembling system workflows to order, potentially in a 'plug and play' way, from a library of shared interchangeable services, rather than building new silo'ed systems for each business area. In the ABS's Business Architecture model, the building, testing and release of new/enhanced system components is captured by the Statistical Capability and Corporate Capability lines of business. Collection set up activities which should be part of an integrated data collection strategy in the Design activity (preparing a collection strategy), or which fit better in the Project Management activity (training collection staff and ensuring collection resources are available), have been removed. In the future, the focus of the remaining activities is to collect, load and validate (or quality assure) the integrity of any type of externally provided data (micro or macro level), via any type of data channel. The Collect activity has therefore been re-structured along these lines. In the previous version, it wasn't clear whether some of the validation of administrative datasets took place in this activity or in the Process phase, but users wanted to put it here as they often needed to do some integrity checks and fix any issues, both before and after loading these datasets. The main scope change for this activity is the inclusion of all output compilation activity, regardless of whether it was compiled using microdata or macrodata. The end point of this phase is now a set of statistical data and metadata outputs which have been loaded to stores, ready for packaging into products and services for release. The new structure, which brings in activity 6.1 Prepare draft outputs, aims to reduce confusion around where to draw the line between the previous processing related phases. As data files are not finalized until they've been cleared in the Validate activity, there is no longer a need for 5.8 Finalize data files at the end of this activity. In line with the scope change and desire to better reflect the future world of compiling outputs using a range of different datasets, the activity has been changed from Process to Compile. Output preparation or compilation activities previously in 6.1 Prepare draft outputs have been moved to the Compile activity, May 2013 7 Disseminate 7 Disseminate 8 Archive out of scope 9 Evaluate 8 Evaluate Quality management (overarching) Metadata management (overarching) Client & provider issues management (overarching) ABS BA only 9 Manage Australian Bureau of Statistics since this is where all the other compilation related activities sit. The scope and name of this activity have therefore been changed to reflect the main focus of this activity, which is to validate statistical outputs, ensuring they are 'fit for purpose' for dissemination. Analysis activities most definitely do occur here, but they also occur in many parts of the production cycle, and according to dictionary definitions would only cover the examination of outputs, and not the treatments that need to be applied to make outputs valid (or 'fit for purpose'). In addition, to facilitate the scope change for the Disseminate activity, all content preparation for dissemination has been moved into the Validate activity. The main change of scope in the Disseminate activity is the removal of content preparation, to focus this activity around the dynamic assembly and release of products and services from dissemination stores. Another change is the moving of 7.1 Update output systems to the previous Compile and Validate, and part of the Design activities. If data and metadata are formatted and loaded to output databases earlier they can also be used for validation activities, and not just dissemination activities. The release of dissemination services have been excluded where they are shared capabilities (e.g. a new tool), as they are planned, developed, released and maintained in the Capability lines of business (e.g. ICT component from Corporate Capability and metadata standards component from Statistical Capability). The Archive activity has been removed as an explicit high level activity. Although there is still an activity 2.7.3 Design archiving strategy for the particular project, the act of archiving should largely be done automatically from implementing a range of activities, not just at the end of the project (i.e. part of metadata management). An activity 7.2.7 Preserve disseminated products & services has been retained to ensure they are automated and quality assured for 'forever' support by another area. Archiving rules should be defined, and data and metadata repositories managed in an organisation wide way in the future. In the ABS's Business Architecture, this archiving activity is captured by the principle 'Managed as a National statistical asset', and the activity itself within the Statistical Capability line of business. No change. Continues to cover the end of project type evaluations, rather than the evaluation or statistical quality management activity that occurs throughout the life of a project. Ongoing quality management for individual projects will be embedded throughout the phases, and during Project Management activity, and quality reports will be produced from many activities (especially automated ones) and checked. Corporate initiatives in this area are also covered in the Statistical Capability line of business. Metadata management occurs throughout the model, so any breakdown into lower level activities would overlap many other activities. Corporate initiatives in this area are also covered in the Statistical Capability line of business. Client and provider issues management occurs throughout the model, so any breakdown into lower level activities would overlap many other activities. Corporate initiatives in this area are also covered in the Statistical Capability line of business. Management of a production project has been added as an May 2013 production project (overarching) 6. overarching activity, as it is one of the key risk mitigation activities used in individual statistical production projects to increase the likelihood of work being done on time, on budget, and to the required level of quality. As such, there are some activities in the GSBPM which have been moved under this activity (e.g. training of collection staff, timetabling, ensuring collection resources are available). In addition, preparation of business cases have been moved from the previous Specify Needs activity as it is really a project management activity. As funding and formal agreements for data provision may not be secured until some of the more detailed design work has been done, it was considered better to place this activity under managing the project, rather than tie it directly to either the Scope or Design activities. Presentation of the Model 27 The attached diagram provides a pictorial overview of the hierarchy of levels 1, 2, and 3 in version 0.3 of the SPAM. Link to the latest version is in Notes Link. 7. Activity descriptions 28 The attached spreadsheet provides descriptions of each level 1, 2 and 3 business activity in the Model. For some activities, there may also be a few ideas around level 4 activities. The spreadsheet also provides a mapping to GSBPM phases/sub-processes, examples, and an articulation of the state of the output expected at the end of each activity (similar to a 'post condition' in business process analysis models). 29 At this stage, not all of this information has been drafted, so there are still lots of blank cells. Our intention was to draft enough information to allow stakeholders to provide useful feedback. Too little information would potentially waste a lot of their time, as they'd have to think a lot harder about where certain activities have been covered, and providing the complete set of information would take too long to prepare. Unfortunately, given time constraints and pressures, we've had to circulate an incomplete draft of some of the information. It should also be noted that our TSD colleagues would like a lot more detail included in the spreadsheet , including inputs to each activity, and making sure the boundaries of each activity are very clear. A new project would need to be formed to take the work further to the point where it could be considered suitable for use in operational work. Link to the latest version is in Notes Link. Australian Bureau of Statistics May 2013 8. Using the Model in the ABS 30 Besides being needed to input the ABS2017 Program NPP, there are a number of practical uses in the ABS for the business activity models currently being developed. The uses for this framework include: 9. developing shared capabilities business analysis and re-engineering activities, both now and for the ABS2017 program for which funding is being sought categorization in a library of shared resources (Business, Information, Applications, and Technology) - for example, level 3 or 4 activities might align with the reusable (business) services supporting assemble to order, Service Oriented Architecture, and technical implementation supporting a corporate approach to development and management of data and metadata promoting a common language that more ABS business people will be happy with (some are not happy at trying to force their activities into the limited and outdated GSBPM) managing knowledge in production projects underpinning effort recording and understanding where effort is expended over time (note only need revised high level model for this) aligning with other frameworks being developed in the ABS (e.g. ABS People Capability Framework). Next steps 37 As version 0.3 of SPAM is considered fit for purpose for the ABS2017 NPP business case, there will be no further consultation and refinement of the Model until additional funding and resources can be found. Strategic Design Committee members will, however be individually briefed on this version of the Model, and asked to endorse the interim Model at an upcoming meeting (probably June 2013). It is also possible that changes identified from work on later phases of the ABS Business and Capability Architecture project may be put through, resulting in a new version of the Model in July/August 2013. 38 The EA project team has commenced Phase 2 and 3 work to identify a hierarchy of capabilities, and assess the maturity of the high level capabilities to enable business activities. The EA Reference Group and selected Capability projects have been asked to provide input on the 'as-is' and future changes needed for our methods, systems, standards & frameworks and other elements of capabilities, in a format which maps back to the current version of SPAM. Once this feedback has been consolidated, reviewed and gaps addressed, it will input the maturity assessment of capabilities the ABS will need in the future. Lynne Bismire & Eden Brinkley Enterprise Architecture & Business Solutions Development Australian Bureau of Statistics 24 May 2013 Australian Bureau of Statistics May 2013 Appendix 1 Australian Bureau of Statistics May 2013