Update business case document

[brainwasher]

Related to https://github.com/SDMX-SWG/Guidelines/blob/master/Business_Case_for_SDMX_Dec_2019%20post%20SWG%20meeting.docx

[dbarr99]

Submit this to the secretariat after removing the comments/accept changes

[dbarr99]

I sent this to the SDMX Sec on 21/2/2020 with a 1 month deadline

[dbarr99]

Feedback from World Bank:

Overall looks good. A few suggestions: Use case and scenario of SDMX the first three bullet points on push pull and hub focusses on IT architecture and are not connected to the remaining bullet points on use cases. It will be good to flip the order and move the IT architecture below the other benefits as the audience I believe is primarily non-technical. In challenges of SDMX we could highlight initial high capital investment for future efficiency and cost gain. This in my opinion is what causes a lot of reluctance in adopting SDMX as there are no immediate tangible gains to be seen. How do we plan to communicate this other than in conferences?

[dbarr99]

Feedback from the ECB: Thank you for the very clear guideline. We have only few remarks attached for SWG’s consideration. Not included in attached guideline but it could be clarified which database the document refers to (is this a generic, theoretical database, or a relational database, or another, or all database types) perhaps to mention some examples in a footnote/annex if possible (since the document will be potentially be read by solution designers/project managers which are looking to various solutions). We added the use case of mappings of data structures (to be applied when mappings are required and needed), as another possible use case (since even if mappings should be reduced by harmonisation, sometimes mappings are needed and SDMX support this use case). We suggest to add a link to the SDMX Tools in the references area https://sdmx.org/?page_id=4500 (since they are mentioned in the IT tools) to provide an idea to guideline user of which kind of tools are encompassed in the toolkit.

[dbarr99]

Feedback from IMF: As a general feedback, we suggest that the document be expanded in scope to include the perspective of developing countries where SDMX implementation is still a challenge. More specific comments and edits are included in the document. More feedback in document comments: https://github.com/SDMX-SWG/Guidelines/blob/dev/Business_Case_for_SDMX_IMF_Comments.docx

[delcada]

Clicking on [https://github.com/SDMX-SWG/Guidelines/blob/dev/Business_Case_for_SDMX_IMF_Comments.docx] generates an error message:

No results matched your search. You could search all of GitHub or try an advanced search.

[dbarr99]

Please find below Eurostat’s comments to the document entitled “The Business Case for SDMX” drafted by the SWG. Due to the nature of these comments and extensive work that would require implementing them, they are to be seen as ideas for further improvements as time and resources allow.

• Premise: The premise for my comments is that I believe this document should mostly be oriented at middle and senior management in organisations that might be looking at the possibility of adopting SDMX. The document currently contains the following sections: o A description of typical use cases o A description of the six benefits of SDMX o A couple of paragraphs on the “How” of SDMX implementation (management buy-in and challenges) o An Annex showcasing one implementation / use case

Below you will find comments on each of these sections.

• Use cases: o I currently find two main comments on this section. The first one is that the list of use cases is very long and partially overlapping. The second one is that there is an imbalance in the list. The first use case (data and metadata reporting) is much longer than the other ones and is the only section that contains details about the IT architecture. o Given the target audience, I think it would be more beneficial to provide a shorter list of broader use cases. A suitable list could be “Data and Metadata Reporting”, “Data validation and processing”, “Data Warehousing” and “Data Dissemination and Discovery”. It would be extremely useful for each of these use cases to provide a concrete implementation example / case study. This is done in Annex I for the “Data discovery and visualisation” use case, but it should be done for all of them. In this way we could hopefully also reuse the work done by the BIS in collecting and drafting case studies for SDMX implementation across different organisations.

• Six Benefits of SDMX o I have once again two main comments on this section. The first comment is that the way the benefits are described is not consistent. Benefit 1 is described with a series of short paragraphs, whereas benefits 2, 4 and 5 follow more a bullet point approach, with little emphasis given to each bullet point. The second comment is that two of the benefits (“promoting standardisation” and “SDMX supports modernisation”) are not benefits in and of themselves, but they are rather means to an end. o I would propose therefore to:  Make sure the benefits are described in a consistent manner. I definitely prefer the approach of benefit 1: a few short paragraphs rather than a “laundry list” of minor points. This document does not need to be exhaustive, it simply needs to be convincing.  Make sure all the benefits relate to key outputs senior management cares about. Regulation 223/2009 outlines a set of statistical principles (professional independence, impartiality, objectivity, reliability, confidentiality and cost effectiveness) and quality dimensions (relevance, accuracy, timeliness, punctuality, accessibility, coherence and comparability) for European statistics. While the regulation is for European Statistics, the principles and quality dimensions are generic, and it would be advantageous to structure the benefits by showing how SDMX contributes to realising these principles / implementing these quality dimensions. The list of benefits could therefore look like the following: • SDMX inspires trust (current benefit 1) • SDMX improves coherence and comparability (current benefit 2) • SDMX improved timeliness and accessibility (current benefit 4) • SDMX improves cost effectiveness (combining current benefits 3, 5 and 6)

• Management buy-in and Challenges: I am not sure these sections belong in this document, as the document should rather address the question “Why SDMX” rather than “How SDMX”. It amy makes more sense to write a separate document on the How, reusing existing material (e.g. the SDMX Started Kit for Statistical Agencies, which could definitely be improved), rather than mentioning these two very important points “en passant” in this document.

• Overall structure and format: o In terms of structure, it would make sense to me to list the key benefits first, then the use cases (with a Case Study for each use case to illustrate it and to show how the benefits were realised). o I would hope that, rather than keep this as a Word document, we could implement the document as a set of pages on sdmx.org, possibly in a new section called “Why SDMX”. But this may require some reshuffling of the current content of the website.

[Fernandomorente]

Feedback from UNSD: Dear all,

Thank you for the very clear guideline. We have only few remarks attached for SWG’s consideration. Not included in attached guideline but it could be clarified which database the document refers to (is this a generic, theoretical database, or a relational database, or another, or all database types) perhaps to mention some examples in a footnote/annex if possible (since the document will be potentially be read by solution designers/project managers which are looking to various solutions). We added the use case of mappings of data structures (to be applied when mappings are required and needed), as another possible use case (since even if mappings should be reduced by harmonisation, sometimes mappings are needed and SDMX support this use case). We suggest to add a link to the SDMX Tools in the references area https://sdmx.org/?page_id=4500 (since they are mentioned in the IT tools) to provide an idea to guideline user of which kind of tools are encompassed in the toolkit.

[dbarr99]

Feedback from SDMX TWG: The business case for SDMX: I don’t understand what is the purpose of the document and who is the target audience. Depending on the target audience, it may discourage rather than encourage the adoption of SDMX. I would have expected something easier to read with simply to understand use cases like: o A statistical institute is looking for a way of making more accessible its data to users, including making it more machine readable -> SDMX for dissemination: Webservices, tools like .STAT for visualisations … o A statistical organisation aims to collect data from other organisations, there is a need for standard inputs -> SDMX for data collection: push, converters, DSDs.. o A statistical organisation aims to collect data already existing and disseminated by others -> SDMX for data collection: pull, webservices… o Statistical organisations want to reduce the burden when there is shared reporting obligations -> SDMX for data collection: global DSDs … I would try to explain these scenarios in a plain an easy to read language rather than with the technical terms and explanations provided in the document (maybe I didn’t get right who is the target audience for this document).