Tableau Tricks
|
Tableau Tricks |
10 Dashboarding Tips toward User Satisfaction
Essentials Explained
Platform: Any BI Platform
Leading Thoughts:
Developing relationships with business users isn't always easy, and it’s our job to make sure they can find the right answer they need, the way they want, with the right level of detail. Don’t let your dashboards end up in the recycling bin shortly after delivery. Follow these 10 simple tips to build trust with your stakeholders, starting from the very first second of delivery.
Table of Contents:
The Magic of Look & Feel: Clean Up & Formatting
Don’t Leave Your Users in the Dark: Use Custom Tooltips
Consistency is Fluency: Sync the Filters
Someone has made Styling Guides: Modern Enterprise UI Design
Tell, don’t Show: Use Reference Lines to Help Decision-Making
Life isn’t that Simple: Introduce the Third Dimension whenever Appropriate
End that Export-CSV Battle: Drill Down if You Can
Do They Really? Delete that Mobile View
Data-“Breaks”, it’s Okay: Data Validation
Accelerate the Action Ahead: End with an Action, not Overwhelming Information
The Magic of Look & Feel: Clean Up & Formatting
Cleaning up the charts might sound basic, but it holds the live-or-death moment of the very first impression of your dashboard.
The following can be used as a quick starter, but you should develop your own manual!
Does your chart have a meaningful title?
Does the color palette match the rest of the dashboard?
Does EVERY grain of ink have a meaning? If not, remove those chart borders, marks, grid lines, axis rulers, and even axis labels.
Did you set up appropriate padding (gap) so your charts and texts are not falling off the border?
Is the size of the mark appropriate to view on a preferred screen (in many cases, a small laptop screen for business users)?
If using a painted background, does the filling color look natural / blend in?
Don’t Leave Your Users in the Dark: Use Custom Tooltips
The default tooltip works, but it is just as annoying as an error message without saying what to look for.
Use your best judgment and knowledge to decide how much information should be included in the tooltip on hover.
If the users come from a purely business practice, just as finance or operations, bullet-point style tooltip should be implemented
If the users are using the dashboard for a niche scenario that involves heavy context, a natural-language style tooltip should be used to help interpretation.
It is ultimately how you build your user persona.
Consistency is Fluency: Sync the Filters
Imagine how many times you click on something on your phone and it takes you to a completely different place than you thought of? Almost zero!
Technical proficiency is assumed and should be carefully tested on filters. They are as dangerous as they are helpful. Make sure your dashboard flows smoothly!
Filters should be (mostly) synced for all visuals on the same view to avoid confusion, and in some cases across views.
Yes, there are complex use cases where filters should not be synced. Just make sure there are enough instructions or visual cues to instruct the user on how the dashboard functions.
Be extra careful when applying filter actions and filter logics that apply to multiple views. They are the root of the user frustration and bugs most of the time.
Someone has made Styling Guides: Modern Enterprise UI Design
Exactly as it sounds, someone in the organization probably has a styling guide for internal communication and branding.
Following the guideline makes your dashboard “less foreign” and more “embedded” into the decision-making workflow.
If this hasn’t been a guide yet, following the look & feel of the company website grants you a high chance of a “WOW that looks nice (polished and paid attention)”.
Use rounded visual design whenever appropriate, yet not necessary.
Rounded bar charts, buttons, and icons match the modern UI design we see across different platforms. Leverage that clean and minimalistic design to give users a smoother transition between your dashboard and other productivity tools in their workflow.
Tell, don’t Show: Use Reference Lines to Help Decision-Making
A dashboard can be overwhelming when there are millions of records behind the charts. And our goal is to help business users find what they need ASAP and not miss anything.
When possible, always use a reference line to show a sense of “slicing the data” and comparison.
Some chart types are naturally attracted to reference lines:
A scatter plot can be transformed into a four-quadrant plot to remind users of the business context and narrow their attention to exactly what they need to focus on.
Bar charts are a great use case for reference lines to describe a sense of target hit-or-miss, distance to average, or else.
Life isn’t that Simple: Introduce the Third Dimension whenever Appropriate
“Can’t I just put something together in Excel?”
“You can, but we would be answering different questions, thus different answers.”
Our quality of work should demonstrate an in-depth level of analysis that looks completely different than my first giant Excel workbook when I first learned it, made with flushed thoughts and excitement.
Effective dashboards are built to answer questions that can’t be answered by X and Y axes alone:
Use color, size, or dual axis to bring in relevant information or context that aids decision-making.
Innovate your own chart design that addresses a specific demand from a user.
Always step back to assess the feasibility of implementation, maintenance, and adoption.
End that Export-CSV Battle: Drill Down if You Can
“This chart doesn’t quite answer my question, so let’s bring 3 million rows of data into an Excel Pivot Table.”
It is not the stakeholder’s job to provide a step-by-step checklist on what chart and information they need to see. If they can’t find their answer, we must have missed it during requirement gathering.
Building drill-down logic and dynamic zones helps users to start from the externalized indicator of the issue and dig into the finer-grained details to find the leading performance indicators of these issues.
Drill-downs are meaningful in expressing the following relationships:
Categories and sub-categories
Enterprise level to regional level/division level
Any hierarchy that exists in the data
Metrics presented at various levels of detail
Do They Really? Delete that Mobile View
Let’s make it clear: Some dashboards are specifically designed to be used on a mobile phone. But most are not.
The mobile view that comes by default in Tableau can be difficult to navigate. The more complex your workflow looks on a desktop, the worse it will get on a mobile phone screen.
Mobile dashboard view should either be specifically crafted or deleted before publishing. This way, we are confident that our users can find the right answers they need wherever they are.
Data-“Breaks”, it’s Okay: Data Validation
It’s the human behind the screens and keyboards who developed your data pipeline in production. It’s okay when it doesn’t refresh correctly on an abnormal day, and sometimes your users catch that before you do.
Add a data validation element to your dashboard that describes the most recent date of record, total count of records, or other indicators that our user is familiar with and confident about.
This “safeguard” builds trust between you and the user. It can be done in many ways:
A descriptive sheet in the corner of the dashboard
A dynamic zone pop-up window that requests validation of these indicators upon opening the workbook
An IF logic that triggers a warning if the refresh test of the date or record count doesn’t match.
Accelerate the Action Ahead: End with an Action, not Overwhelming Information
What should the user do after they look at the dashboard? If they head off to lunch, we did not do a good job.
It can be tempting to think that dashboards are the final destination of our users, simply because that’s where we hand them off. But it’s not.
An effective (hopefully less recycled) dashboard should prompt the user to execute an action:
An URL action that redirect users to
A dynamic zone pop-up window that requests validation of these indicators upon opening the workbook
An IF logic that triggers a warning if the refresh test of the date or record count doesn’t match.