Adopting Self-Service BI with Tableau - Notes from the field

(originally this article was created and posted by me on March 7, 2016 at datasciencecentral.com, now I am transferring it here)

I have spent many hours planning and executing in-company self-service BI implementation. This enabled me to gain several insights. Now that the ideas became mature enough and field-proven, I believe they are worth sharing. No matter how far you are in toying with potential approaches (possibly you are already in the thick of it!), I hope my attempt of describing feasible scenarios would provide a decent foundation.
 
All scenarios presume that IT plays its main role by owning the infrastructure, managing scalability, data security, and governance.

Scenario 1. Tableau Desktop + departmental/cross-functional data schemas.

This scenario involves gaining insights by data analysts on a daily basis. They might be either independent individuals or a team. Business users’ interaction with published workbooks is applicable, but limited to simple filtering.

User categories: professional data analysts;
Technical skills: intermediate/advanced SQL, intermediate/advanced Tableau;
Tableau training: 2-3 days full time (preferably) or continuous self-learning from scratch;
Licenses: Tableau Desktop.
Pros:

• Pure self-service BI approach with no IT involved in data analysis;
• Vast range of data available for analysis with almost no limits;
• Fast response for complex ad-hoc business problems.

Cons:

• Requires highly skilled data analysts;
• Most likely involves Tableau training on query performance optimisation on a particular data source (e.g. Vertica).

Advice:

• Create a “sandbox” that allows data analysts to query and collaborate on their own and without supervision. Further promotion of workbooks to production is welcome.

Scenario 2. Tableau Desktop + custom data marts.



In this scenario, business users are fully in charge of data analysis. IT provides custom data marts.

User categories: business users, line-managers;
Technical skills: basic SQL, basic/intermediate Tableau;
Tableau training: two or three 2-3h sessions + ad-hoc support on daily basis;
Licenses: Tableau Desktop + Server Interactors.

Pros:

• Easy access to data for ad-hoc analysis;
• Self-answering critical business questions;
• Self-publishing for further ad-hoc access across multiple devices.

Cons:

• Adding any data involves IT support;
• Requires elaborated data dictionaries.

Advice:

• Make requirements gathering a collaborative and iterative process with regular communication. That would ensure well-timed data delivery and quality;
• Deliver training in 2-3 wisely structured sections with 2-3 week breaks for business users to have time for playing with software, along with generating needs for the new skills.
• Focus on reach visualisations, not tables.

Scenario 3. Tableau Server Web Edit + workbook templates



This scenario fully relies on data models published by data analysts and powerful Web Edit features of Tableau Server.
User categories: line-managers, top managers;
Technical skills: Tableau basics;
Tableau training: one 30 min demo session + ad-hoc support;
License: Server Interactor.

Pros:

• No special training;
• Fast Tableau adoption with basic, but powerful Self-service BI capabilities (Web Edit);
• Thin client access via any Desktop Web Browser;
• Could serve as a foundation for self-service BI adoption among C-Suite.

Cons:

• High level of accuracy for data preparation and template development;
• Any changes in the data model require development and republishing of a template.

Advice:

• Try to select the most proactive and “data hungry” line manager or executive, who could help to spread the word;
• Investigate analytical needs, ensure availability of a subject matter expert;
• Start with simple visualisations, but be ready to increase complexity;
• Provide as much ad-hoc assistance as you can.