A Fortune 5 technology company asked Lab Zero to explore Tableau; a business intelligence tool for connecting to myriad data sources and creating advanced data visualizations. We built a reporting infrastructure for an organization with disparate data sources, and gained some insights on Tableau along the way. 

Users wanted a no-code option to access and visualize data. Our partner selected Tableau because it's used by other departments and it's cleared by security. Licenses are available, and there’s extensive training material available. We put one of our best full-stack developers on it, and we were off to the races.

What We Learned

The Good

Tableau is a powerful tool when a user's need is to visualize a clean data set. It really flexes its muscles when a user does not have the skill set in analytics needed to perform certain work. For example, Tableau easily connects to databases. Instead of needing to learn SQL or have a developer build custom tools, Tableau can connect directly to databases. Our user could then access and explore the data leveraging a nice user interface.

There are also extensive options for visualizations which aren’t available in Excel. Geographical data is a large component of reporting for our stakeholders. Instead of using a bar chart in Excel, they were able to create a variety of maps in just a few minutes. 

The Unexpected

Our task initially seemed easy. The organization had amassed considerable data on employee health and wellness. Users wanted to create reports by directly accessing the database, bypassing the need to ask for development work. 

Our team previously built applications for the client, so we were familiar with the data. Tableau also advertises a fairly seamless experience out of the box. Kicking off the project, our team was confident this would be a seamless integration with little friction.

Configuration Limitations

Even with loads of institutional data knowledge, we immediately ran into roadblocks. Upon loading the data into Tableau, we were stymied by using the Tableau Prep Builder. It could not handle many-to-many joins or joins involving more than one condition. We eventually had to switch to “Custom SQL” in Tableau Prep Builder as a way to work around this. 

Cleaning Data in Tableau

Once we had successfully set up the custom queries to surface data, we began formatting the employee health data in sensible ways. Immediately we hit another roadblock. When comparing the Tableau reports to their Excel equivalents, we found discrepancies in the data. As it turned out, an essential part of this report was poorly structured. After exporting from the source, extensive Excel wrangling was being done to prepare the data for reporting. Although this custom cleanup work was easily done in Excel, it’s much harder to do in Tableau. 

In general, institutional knowledge like this needs to be captured in the way an organization’s data is structured, or else: danger zone. Tableau may be a powerful reporting tool, but it is not a panacea for organizations that have done a poor job with their information architecture. There’s no escaping GIGO (Garbage In, Garbage Out) with Tableau.

User Challenges

Tableau promises pretty visuals with an intuitive user experience. We found it takes more of an investment than advertised to get there though. If the goal is to make a simple chart using clean data, it is an easy user experience. But when an unfamiliar user jumps into an already built-out workbook, or needs more advanced functionality, it can get frustrating. 

Our users were also accustomed to having complete control manipulating the raw data in Excel which would then be used in reports. They would perform lookups, clean the data, make formatting changes, tweak text here and there, etc. This was difficult in Tableau, especially never having used the tool. The good news is that there are lots of training materials available online. We also found a company wide Slack group which was a great resource.

What Lab Zero Can Do For You

Tableau is a powerful tool which presents a seamless facade on the surface. The actual plumbing underneath takes time to 'make real' though. It’s required to know what the desired value is before beginning a project. And it takes work to build out a custom Tableau experience. Once built, though, it can be a powerful tool to open up data to the entire organization.

Lab Zero has tackled these challenges, and knows what it takes to leverage Tableau to create organizational value. Let us know how we can help you get there, too.