I am putting together an approach for my organization to follow to apply various levels of governance for visualizations on our internal Tableau Server while still allowing the flexibility that comes with Tableau. This is early development of the process so any input you have on the approach is appreciated.
First, some details on our Project structure on Tableau Server. The project folders are assigned to content areas, and within each content area we have three folders: "Development"; "Production - Unvalidated"; and "Production - Validated". The "Development" folders are highly ungoverned and are intended to be an area where publishers can share their work with minimal restrictions. The "Production - Unvalidated" folders are more controlled and are intended to be an area to share polished visualizations that are frequently used. There are fewer publishers with rights to the "Production - Unvalidated" folders and they must execute a production release checklist prior to publication. The "Production - Validated" folders are highly governed and are intended to be an area to share visualizations that have successfully completed validation testing and the production release checklist.
Next, some details on the Production Release Checklist. This is a short checklist that guides the publishers though common pitfalls encountered when publishing workbooks. Items include verifying unnecessary data is not included for publication, filters are shown only where needed, tooltips show helpful information, colors are consistently applied, titles are shown, instructions are provided, calculations are documented, intended use is documented, intended audience is documented, etc. Anything that we want to ensure is consistent across the content area will be included in the checklist. There may be several checklists, one for each content area, but we'll probably wait to see how the checklist develops before breaking one out for each content area. The checklist is intended to be executed by someone other than the workbook author.
And finally, details on validation testing. For workbooks that require formal validation testing our internal validation methodology will be applied. This includes defining intended use and requirements, identifying areas of risk, documenting the workbook design details, and writing/executing test scripts to prove that the workbook fulfills the requirements and satisfies the intended use statement.
In summary, I'm trying to build a system and process that enables data discovery with rapid exploration and sharing of discoveries while at that same time providing the framework to maintain control of a validated state where deemed necessary.
I'd appreciate any feedback from the community on this approach, suggestions for improvement, or areas where you think "that just won't work". Thank you in advance for your replies!