Training Transcript:: Planning For A Successful Dashboard
Training Transcript:: Planning For A Successful Dashboard
Training Transcript:: Planning For A Successful Dashboard
When you're planning a dashboard, make sure that you clearly understand its purpose,
who its audience will be, and the questions this audience wants the dashboard to help
them answer. In this dashboard, for an executive VP, the goal is to give a high-level view
of the company's sales and profit totals over a broad timeframe. To get a picture of
business health, this audience wants to answer questions such as: What our overall
customer sales by location, and also broken down by product category and customer
segment? And how do the data look when we filter by location?
In another example, let's say our audience consists of product analysts, and they need
more specific data. They want to evaluate profitability across the product catalog, to
decide which items to keep in the inventory. They need to drill pretty deeply into the data
asking: Which products had highest sales, and most or least profit, month by month?
And for each product type, what are the specific sales details? If you have any uncertainty
about the questions your dashboard needs to answer for your intended audience,
determine that before you go further.
Once you're certain about a dashboard's purpose and audience, make a plan for the
dashboard's design. Your first consideration should be: How will the dashboard be viewed
and shared? Will it be dynamic and interactive, shared on Tableau Reader or Tableau
Server? Or, will it be viewed in a static format, as a PDF file, for example?
If the dashboard is shared on Tableau Server, you'll have more flexibility for your views.
You'll also have more sizing considerations. Will people view the dashboard on a mobile
device, for example, or an iPad? The more you can customize the dashboard to the
viewing device, the more you can enhance users' experience. If some will view it on
tablets, and others will use desktop monitors, consider creating custom versions for
those devices so that information displays correctly.
2
If the dashboard will be viewed on Tableau Reader, your audience won't be able to edit or
view underlying data, and you'll need to be more explicit in providing answers with the
data. If the dashboard will be static for its users, design your views to work without
interactivity.
Next, in your dashboard planning, decide on the chart types and data you want to use for
the dashboard. Think of each chart's purpose and how you'll want to relate the views to
each other. Then, organize the views in a logical way, and ask yourself whether they
answer the questions this dashboard needs to answer. Is every view essential? Is every
view relevant to your audience?
As the last part of your planning, think about how you want users to interact with the
dashboard. Do you need to create filters, actions, or parameters? How can you help
people get at the data and answer their own questions?
You've developed a solid plan for the dashboard. Now, as you put it together, apply the
principles of good design. Think of the design both in terms of the dashboard as a whole
and the details of individual worksheets.
Considering the dashboard level: Don't overcrowd the dashboard. You can always build
more dashboards and cover other aspects of the data in coherent subsections or even
using story points. Streamline your use of legends and filters. For legends, remove those
that aren't essential. For filters, where possible, have one filter act on multiple views.
Better yet, use actions or parameters to reduce the number of filters. Place important
controls, such as global filters, on the upper-left, and arrange views logically, generally
following a left to right, top to bottom order. Evaluate color and font choices for clarity
and consistency. Remember to use a limited number of colors and carry the meaning of
key colors throughout the views.
Considering the worksheet level: Review titles on views, filters, and legends for clarity
and helpfulness. Include instructions so people know, at a glance, how to interact with
the views and what the data represent. Within a view, remove unneeded headers, labels,
gridlines, and tick marks. And remember tooltips: Edit them for appearance, clarity,
consistency, and helpfulness.
3
Once you have a working version of your dashboard, test it for usability and
performance. Start by applying the 5-second rule: In 5 seconds, could users tell the
dashboard's purpose and get its relevance to them? Is it easy to use, so that a person
would be enticed to interact with it?
Now, try the 5-second rule on some targeted users of the dashboard. Observe them in a
usability test. Does the data make sense to them? Can they find what they're looking
for? Do they need more guidance on how to interact with the dashboard? Do they want
data that isn't shown? Revise the dashboard based on their input.
Last, check the dashboard's performance. Does it have an acceptable load time? When
you interact with it, does it respond as you'd expect? If you have a sluggish workbook,
you can run a performance test on it using the performance recorder on the Help menu
in Tableau. This traces your actions and shows which components are causing the lag.
You can then determine ways to address the cause of the lag in your workbook. You can
find more about evaluating performance issues on Tableau.com.
We've covered four essential steps to creating a successful dashboard: Know the
dashboard's purpose and audience. Plan its design thoroughly. Build it using design best
practices. And test the result.