By: Kirk Olson and Eric Parker
Kirk and Eric co-founded OneNumber in 2016. They’ve helped many individuals and organizations answer the question “What does success look like?” and help them build systems to measure, analyze and improve relevant data to help them accomplish their goals.
Great! You have now hosted a successful Design Sprint session and mocked up some whiteboards everyone feels good about building. What are the steps you can now take to ensure your groundwork results in Tableau dashboards that are used by your audience?
We have several steps we follow when facilitating projects that give them the best chance of succeeding.
1. Show stakeholders an intermediate mock-up with “fake” data.
Sometimes just seeing a “dashboard” in a graphical interface as opposed to a whiteboard sketch helps aid decisions about visualization and usability. Building a decent looking mock-up may only take a couple hours but it could save you days of time down the road. PowerPoint or KeyNote are great tools for throwing something together quickly.
2. Build an alpha version in Tableau.
After generating a mockup and receiving feedback, building an alpha version of your dashboards in Tableau gives people a chance to see what it would look like in the product. You can also now quickly show people different options (e.g. Here is what this would look like as a scatter plot and here is how it would look as a diverging bar chart).
On many projects, stakeholders aren’t seeing a version of the whiteboard concepts in Tableau until weeks or maybe months later. Don’t let that happen to you! A gap that long can seriously inhibit momentum on the project. Ideally you should be looking over a mockup with your stakeholders no later than two-weeks after the initial design sprint meeting.
Lots of people tell us, “There is no way we could have a mockup ready in Tableau within two weeks. Our data is too messy.” Our usual response is, “No problem, fake it!”. We don’t mean to lie and say you are using real data when you aren’t, but don’t let the data be a roadblock. Maybe you can use some exported files from your database that are loosely stitched together. Or one of our favorite approaches is to generate realistic looking, fake data. Our go to tool for that is mockaroo.com. Your biggest challenge if you use mockup data is to ensure your audience knows it’s fake and doesn’t get lost in the weeds as you are presenting this alpha version of your new product.
3. Get the data right.
Now that you’ve got buy-in on some mockups, it’s time to prepare the data. Don’t be surprised if this is the most time-consuming step of the entire project. There are a number of elements that might make this challenging; establishing connections to databases, data elements scattered across many platforms, data is in the wrong shape for Tableau, values are unclean.
This is one stage you don’t want to rush. If the data isn’t right, your dashboard’s credibility resets to zero. Humans are naturally resistant to change. If someone is already hesitant about moving away from their manual Excel reports and they can spot issues with the data in the new product, you can bet it is going to make them uneasy.
We have a variety of tools we employ to help ease this process and make it as clean and repeatable as possible. If the data preparation isn’t too complex, this might even be able to be done in the Data Source interface of Tableau Desktop. If that’s not possible we generally turn toward some combination of Custom SQL, Tableau Prep and Alteryx. The biggest determining factors here are experience/tools at our clients’ disposal and the complexity of the data preparation.
4. Show off the completed dashboard, with the right data.
You now have the data right and built a functional Tableau dashboard out of it, great! Now it is time for a final review with your stakeholders. Any minor tweaks or any business needs that have changed can be addressed at this stage. One again you are seeking buy-in and sign off to proceed with launching the content to a wider audience.
5. Schedule and complete training.
Do you want your dashboards to get used? This is going to seem like an obvious piece of advice but make sure people know how to use them! We spend so much time developing and interacting with content in the world of data that we forget for a lot of people this is a small sliver of their day.
You can build training content directly into your dashboards with images and videos, but it’s also a generally good idea to schedule some training sessions with users to demonstrate how the content works and be able to address their questions. Sometimes it’s the little things that make all the difference too. For instance, with some audiences we’ve seen encouraging them to bookmark the Tableau Server dashboards in their browsers has led to 200%+ growth in engagement.
6. Check in on actual usage.
Source: tableau.com
Want to know how much your dashboards are actually being used? Don’t believe what people say, believe what they do! We’ve come across situations where stakeholders pay dashboards great lip service, “Oh yes, those dashboards are game changers. We use them all the time.” only to find out they are hardly being used.
Tableau Server tracks usage and with the right permissions, you can access that data. You can also check to see if anyone has set up subscriptions to receive regular screenshots of your dashboards in their inbox (not a bad thing to demonstrate in training!).
If you find they aren’t being used, it’s a good idea to check in with users. The reason may be innocuous (e.g. a manager sends a PDF of the dashboards to their team every Monday) or it might be that they truly aren’t being used. If that’s the case, ask the audience why. Ask them what kinds of actionable insights they need included for these dashboards to be relevant to them. Ask them how their needs have changed since the initial Design Sprint session.
7. Keep iterating, great dashboards constantly evolve.
“No dashboard is ever truly finished.” That’s one of our favorite mantras at OneNumber. Not because we are sadists, but because we are realists living in the fast-moving 21st century! :)
One of our clients told me that she sets up a recurring meeting every three months to check-in with different stakeholder groups. In those meetings she checks in to see if the content she created is still meeting their needs. She asks if there have been any changes to their business processes or data requirements.
I love that approach. You don’t know what you don’t know. If you aren’t proactively reaching out and continuing to iterate, you will likely find that usage in your content drops over time. Because your users’ needs continue to change, the Tableau product continues to evolve, and even the kind of data you are capturing may be growing, keep on iterating!
Thanks for following along! Did we miss something you find to be a crucial step? Let us know in the comments below!