Standing up any new analytics tool in an organization is complex, and early on, new adopters of Tableau often struggle to include all the complexities in their plan. This post proposes a mental model in the form of a story of how Tableau might have rolled out in one hypothetical installation to uncover common challenges for new adopters.
Tableau’s marketing lends one to imagine that introducing Tableau is easy: “Fast Analytics”, “Ease of Use”, “Big Data, Any Data” and so on. (here, 3/31/2017). Tableau’s position in Gartner’s Magic Quadrant (referenced on the same page) attests to the huge upside for organizations that successfully deploy Tableau, which I’ve been lucky enough to witness firsthand.
New adopters need to think across five dimensions when planning their Tableau rollout. Each dimension represents an important planning perspective for those envisioning development and maintenance of a suite of Tableau reports:
- One analyst, one customer
- One analyst, many customers
- Adding Tableau Server for viz distribution
- Expanding to a data analysis team
- Many Tableau analysis teams
One Analyst, One Customer
This is the most basic level of Tableau use; for a new adopter it may be the pilot project where a single data analyst installs Tableau desktop to try it out. Success in this step depends on the ability of the analyst. Can the analyst come up to speed with the tool and effectively apply Tableau to business questions? Can the analyst get sufficient time with the internal customer to understand the need and deliver effectively? Can the analyst find and apply solutions from the Tableau Community without getting totally stuck?
One Analyst, Many Customers
Success in the first dimension often brings us into the next as peers of the business customer begin to request reports, too. The analyst faces new challenges. Here are just a few:
- As the audience becomes more diverse, the analyst starts to think about a standard look and feel, and might begin to write down some usability standards
- Different users request different but related analyses, additional filters, more detail or different summaries, and so on. The analyst must choose between a lowest common denominator approach, delivering the analyses all have in common, or an “all of the above” approach delivering everything to everyone, or something in between.
- The analyst seeks a business sponsor to referee conflicting priorities
Finally, the work outgrows the analyst’s desktop copy as some business customers want to manipulate the analyses themselves rather than work with images and crosstab downloads delivered from the analyst.
Adding Tableau Server for viz distribution
So the organization chooses Tableau Online or brings in Tableau Server to distribute analyses via web browsers rather than extracts and images. This step raises a ton of questions with different possible answers depending on the organization:
- Who gets security access to the visualizations, especially if they contain non-public information?
- How to organize projects (also known as folders in some installations) on the server? Is this an opportunity to separate production from dev/test?
- Who administers the Tableau server? Is this an IT responsibility or does it fall to the analyst?
Clearly, the analyst is getting very busy and might be straying into areas beyond his or her expertise.
Expanding to a data analysis team
With the analyst becoming overloaded, the organization decides to bring in others to help. At this point, the questions typical of a data analysis team come to fore. Some of the questions raised along the way suddenly become more critical. For example:
- Development and usability standards keep the work of the team consistent, and peer review among team members ensures that the standards are applied.
- How to make sure analysts don’t work on the same thing, and overwrite each other’s work. Versioning was introduced as a Tableau feature in 2016. Will that be a sufficient fallback or is stronger code management functionality required?
At this point in the organization’s Tableau journey traditional project management concerns emerge in regard to how tasks and responsibilities divide among team members, how the team reports progress, who works with the sponsor and the growing user community, and how tasks initiate and flow through the team.
Many Tableau Analysis Teams
And finally, in a large organization success of the first team results in other teams being initiated and funded across the organization. Here the concerns center on how, and to what extent, the teams should follow similar controls and practices. For example:
- Will all teams use the same server, or have their own, or share across servers?
- Will there be a process for ensuring that all teams use the same server and desktop versions, and apply the same configuration and security practices?
- Who will support the server(s) and provide help to desktop users?
- Will the same security processes that applied to the single team work across the organization?
- …and so on…
Hopefully this post helps provide a mental picture that helps new adopters of Tableau plan for what’s ahead. Having been along this journey a few times with a few different analytics tools, here’s what I’d offer as the takeaways:
- For an organization, there are many dimensions involved in Tableau success beyond getting a viz right for a user.
- The Tableau Community is amazing and essential to success with Tableau.
- Growth to a large community of users and teams of analysts demands robust processes for quality, communication, and coordination.
Good luck on your Tableau journey!