I work on the data visualisation team at Just Eat—an online food order and delivery service based in the U.K., operating in 13 countries. We have been using Tableau for around five years. I’ve been managing the Tableau Server for over two and a half years, and over that time, I have helped deliver an increase in users from 600 per month to now over 1500.
I recently spoke at one of the Tableau Cinema Tours about Just Eat’s journey to self-service analytics. My presentation told the story about how we, as a business intelligence team, were not able to deliver data and insights to our organization in the expected time frames. Our “core” data centers around transactions—orders placed with restaurants by our customers—and as you can imagine, there was a lot of it. The business wanted to look into their own data, but our systems weren’t set up in a way that supported a self-service model. We expanded our use of Tableau as part of a widespread self-service analytics strategy—all to generate more value and get data in the hands of the business—fast.
After an effort to curate our data, our next step was to train business users how to understand the data they were looking at. The training program I launched was branded “DataChefs.” In this blog, I’ll focus on how I launched this program and some ideas for building a curriculum of your own.
Considerations for building an internal training program
Starting and running a training program was a bit of an unknown for me. I had previously shown people how to use Tableau in 1-2-1 situations, but I had never taken on the full responsibility for onboarding new users.
I am often asked about the effort required for this style of program. My answer—and also Andy Cotgreave’s favorite response—is “it depends.” If you’re looking to do something similar, hopefully your answers are along the lines of “yes—empower people with data” and “yes-bring on the data revolution.” If so, then my answer is...put in as much effort as you can possibly put into it.
Before I planned anything, I reached out to my Tableau account management team. Lots of other customers had already done similar things, so the Tableau team had some good tips. Overall, it took me about six weeks to plan and build materials, but this timeline can differ depending on your company and your goals.
What did I do?
1. Establish your goals: This is an important one. For me, I wanted to introduce beginners to data and give them the skills to be able to understand the data and use Tableau proficiently to answer a variety of business questions, with minimal support. Remember that you’ll need sign-off from leaders and possibly an executive sponsor to carve out the time and space for people to learn this new skill.
2. Build your curriculum and gather feedback: This part is the most time consuming. Start by preparing a checklist for logistically hosting and conducting training (you have no idea how hard it is to find meeting rooms for two-hour time slots). Next, identify your curriculum and build your materials—what topics do you want to train on? For me this was a course outline document, slides for training, and a Tableau workbook with examples and exercises. Part of this process is reviewing and testing. Check for errors, flow, and progression and gather feedback early to make sure you’re teaching at the right level. Skip to more information about our curriculum.
3. Find your cadence and invite participants: You’ll want enthusiastic people looking to learn and add value to your business. Open up the program and communicate the value of the time investment to your audience. Next, choose your cadence. I decided to offer the training classes over the course of six weeks, once a week for two-hour sessions. For me it felt like the topics I wanted to cover could be delivered in that time (6 weeks). I wanted to ensure I kept people engaged (2-hour session) but not annoy or impact the business units by taking their staff away for days on end.
Establish an audience and find your participants
My goal when building out the curriculum was to take general users of Tableau Server and people who tended to interact with our analysts frequently and enable them to limit these interactions and answer their own questions via the data published in Tableau Server. Generally these were people who had never interacted with Tableau Desktop before.
For this particular program, we had identified (after trial and error) what an ideal participant represented. These three points helped me recruit enthusiastic participants who were eager to develop some new skills and support their teams:
● Familiar with data in some format—Data sources like Microsoft Excel, Google Sheets, or Salesforce.
● Not too junior and not too senior—Positions in the company that gave them time to dedicate to developing a new skill, but also the experience and authority to use their time effectively and not just follow instructions from a manager
● Employed by the business for at least 9–12 months—This was so that the participants had the knowledge and understanding of how the business works, how their team fits in, and what the challenges and questions were.
Based on the beginner level experience with Tableau, I set up the sessions to progress from almost an introductory level all the way through to a fairly proficient Tableau Desktop user.
Curate your sessions: Crafting the DataChefs training program
As I mentioned earlier, our program is made up of six sessions. Note that your program might take a similar path, or it could be very different, depending on your audience and goals. This is the curriculum that worked for our needs:
Session 1: Why we visualise data—In the first session, I set the stage for what DataChefs is all about and define the basics of data visualisation for the group. For this piece, I took cues from Andy Kirk from his data visualisation training course along with exercises from Ryan Sleeper’s blog, Playfair Data. I also like to inspire the class with the infamous TED talk by the late Hans Rosling from 2006. Then we are ready to open up Tableau Desktop. We cover the basics—things like connecting to data, going over connection types, and understanding published data sources. Finally we create a few basic charts, mainly bar charts and line charts using date dimensions and use the marks card to alter their appearance.
Session 2: Chart types, attributes and cognitive load—This session is all about creating charts. We begin by understanding what a chart is and how they are constructed. There is a lot of theory incorporated into the training, especially around themes like cognitive load and psychological schemas. I explain marks and attributes and how they are used to represent the information in the form of a chart. We go over the different classification of charts and then we start “speed charting,” making 11 different charts in the session. I try to make it a bit of fun and light hearted with the time pressure to get all 11 completed, as well as getting participants to share what chart they thought was worthy of a second date.
Session 3: Calculations and data types—One for the finance people. Lots of calculations both table calculations and creating calculated fields. We create examples of calculations for a range of functions: aggregate, number, string, date and logic. I try to make this as realistic as possible and try to frame exercises in which participants are asked a question which requires a calculated field to be created to get the answer.
Session 4: Interactivity (filtering, sets, parameters etc.)—This session focuses on how we can use the functions in Tableau to really engage an audience, partly to try and answer secondary questions and avoiding re-work. We also cover why interactivity is useful along with several examples of interactivity—and then we start creating a working example together.
Session 5: Design and formatting—This is one of my favorite sessions. It is really hard to compress all the information and content out there about designing visualizations in a two-hour time frame. I have tried to solve this by spending time on colors, fonts, layout, and other functions (like layout containers). For the last part of the session I invite a guest speaker—one of my lead analysts in the team who talks about the process of designing. We then end with a bit of inspiration to show what is possible when you practice and iterate.
Session 6: Data Storytelling and dashboards—The last technical hands-on session is used to talk about dashboarding or stories in Tableau—including the different ways you can bring charts together and build a story with titles, text, and annotation. Also, ensuring that you answer business questions and that end users can interpret the information and take action. Finally, we cover publishing and adding all the housekeeping to the final dashboard, establishing principles and best practices.
The Showcase—The showcase is the culmination of our six weeks of training. After week two, I give a task to all participants to find a business question from their area. The goal is that by the end of the course, they can demonstrate an understanding of the theoretical concepts of data visualisation as well as the technical elements of Tableau Desktop. I also want to make sure that the participants are comfortable talking about data and Tableau in front of a small audience.
So week seven of the DataChef program, we all gather to see the final projects from each person on the course. They have 10 minutes to present their business question and demonstrate the dashboard they created—and then we discuss ideas, give feedback, and offer any other comments about it. I also invite special guests and line managers to give them a real view of the achievement from the group.
Phew. Seven sessions of training over nine weeks. Along with preparing all the training content and organizing all the sessions, there are lots of 1–2–1 meetings with the participants to discuss data requirements for showcase projects.
That being said, I wouldn't change it for the world. For me the best part is when each person has that lightbulb moment...the instant they realise what is possible for them in the context of their role and team and the value they can add. It is priceless. This was a quote from one of my trainees: “It made me a lot more interested in data in general and how to use it. Tableau is now second nature to me and I think about how to use it in all my projects.”
Over the course of the program, I have made tweaks and iterated based on knowledge sharing from the Tableau community. I learned that these programs should not be set in stone—they should be flexible enough to change with the needs of your users and your organization.