6 tips for revamping your reporting dashboards
In their new book ‘Make the numbers count‘, co-authors Chip Heath and Karla Starr explain that our brains did not evolve to easily understand large numbers.
We really only have an instinct for small amounts – like in five and under.
Beyond that, it’s just a vague notion of “lots”.
Data visualizations are used to transform and compare large amounts of data, but most reporting dashboards today still look like websites from the 1990s.
We put up with them, but they’re ugly and awful, and we wouldn’t trust them with our credit cards.
Non-strategic reports, i.e. dashboards that are too cluttered or too scattered to understand, make it difficult for your customers and stakeholders to understand the data and allow you to take intelligent action.
Here’s how to turn those clunky dashboards into useful analytics.
1. Get rid of graphics that have no purpose
Not every chart on your dashboard deserves to be there.
Useless graphics distract and compete for attention with graphics that matter.
They can also derail meetings, encouraging your client to focus on the details and natural variance rather than the essentials.
Not all data distributions are useful. Some are just plain useless, and some are anti-useful.
Make every chart earn its place in the dashboard by deleting anything that isn’t:
- Link to goals.
- Provide context.
- Help with understanding.
2. Get rid of “unnecessary ink”
Statistician and dataviz pioneer Edward Tufte Explain,
“… clutter and confusion are design failures, not attributes of information.”
Tufte presented the “data-to-ink ratiowhich tells us to remove any decorative or extra “ink” from the graphics until we are left with only the essentials.
Improve your data-to-ink ratio by minimizing or eliminating:
- Any beveled or 3D effects.
- Redundant chart legends.
- Chart borders and shadows.
- The background color fills.
Arrays are by nature busy, displaying a lot of data at once.
To make your tables easier to read:
- Remove pagination and line numbers.
- Use packed numbers (12M instead of 12,000,000).
- Remove the truncation (“…”) by expanding the column width or wrapping the text.
- Remove decimals (when numbers are > 1).
When you introduce white space and eliminate chartjunk, your reports tell a clearer story.
3. Fix Misleading Axes
Sometimes graphics are so intentionally misleading that they end up the headlines.
More often, however, graphics that mislead do so unintentionally.
Here’s how to find and fix common data visualization errors.
A common mistake is to use a “truncated graph”, where the y-axis does not start at 0.
Truncated charts are so common that Google Data Studio uses them by default in some of its chart options.
The solution to this is easy.
Just set the “axis minimums” from automatic to zero.
Although less common, charts can sometimes have an inappropriate maximum.
This can happen when you have hard coded the max axis based on a previous data set and forget to update it when it uses a different data range.
Also a very easy solution.
Another problem is to use a “logarithmic scale” for your graphs.
When you tried to get a graph a certain way and nothing else worked, you might have switched to the logarithmic scale for better visualization.
Unless you’re really working with logarithmic data, that’s not correct.
Change it to linear.
4. Fixed wrong chart selection
Chart selection is not as simple as simply changing an axis. But it’s probably more important and easier to get wrong.
Have you ever tried using a chart selection guide, only to wonder if your data is nominal or categorical?
If you’re not comfortable with data visualization, it may be easier to stick with trial and error until you come up with something that looks right.
Marketing Crash Course in Graphics Selection
It’s not a complete guide, but it covers many dashboard errors:
- Use dashboards for your big KPIs, even if the same data is in tables and other charts in the report. It focuses on what is most important.
- Use line graphs to show trends over time. If your x-axis is anything other than a time series (continuous data), don’t use a line chart.
- Use only pie or donut charts to show the composition of a whole, ideally with five categories or less. Need to compare pie charts to show a change in composition? You probably need a different type of chart. A stacked bar chart might be a good choice.
- Map charts are a good way to visualize data between regions, and customers seem to like them. Make sure, though, that you’re not just mapping demographics, which isn’t usually helpful for making business decisions.
- Bar charts work well for comparing category performance for a single metric. Think about the sales generated by (campaign, landing page, etc.).
5. Add contrast
Removing “unnecessary ink” from your graphics puts you on the right track.
This next step is to overlay the “necessary ink” that grabs your reader’s attention and makes your graphic even easier to interpret.
These three graphs all use an identical data set:
Chart A is unfocused and looks “noisy”.
Charts B and C vary the thickness and color of the lines to draw your attention to a single line.
Even if you don’t know the actual measurements or dimensions of charts B and C, you immediately know what to focus on.
This is an example of using “pre-attentive attributeswhich our brain instantly processes on a subconscious level.
When you want to emphasize a key point, you can increase the contrast with pre-attentive attributes such as:
Don’t let your audience ask “what am I looking at?”
Help them with contrast and pre-attention attributes.
6. Add context
Context is another kind of “necessary ink” that clarifies the meaning of your visualizations.
As a marketer and subject matter expert, you know what your charts are about.
You can analyze all your dashboards and quickly identify trends and outliers.
For your customers and stakeholders, that’s probably not the case.
The recipients of your reports are probably not intimately familiar with the acronyms and shortcuts that seem obvious to you.
They need more context in the form of:
- Chart titles and descriptions.
- Acronyms spelled and defined.
- Annotations and microcopy.
Your audience also needs to better understand the factors driving trends and data changes in the report.
The metric is the “effect”, but what is the “cause”?
Look beyond the measurements themselves to find the narrative.
- What are the internal and external forces that contribute to performance?
- What backstory could they miss (history, seasonality, competition, buyer preference)?
- Given current and projected trends, what should happen next?
Finally, don’t assume your audience knows the targets, even if they set them.
Help them by comparing performance to goals and not just to previous periods.
Garr Reynolds, author of “Introduction to Zen”, said:
“…you can achieve simplicity in designing effective graphs, charts, and tables by remembering three basic principles: restrict, reduce, underline.”
Cut out unnecessary, fix remaining issues, and add context and meaning to make your charts and dashboards as powerful as possible.
Featured image: Saklakova/Shutterstock