Design consistency


Welcome back, as I continue my slow, alphabetical wade through Universal Principles of Design.

According to the principle of consistency, systems are more usable and learnable when similar parts are expressed in similar ways.  Consistency enables people to efficiently transfer knowledge to new contexts, learn new things quickly, and focus attention on the relevant aspects of a task. (page 46)

That's right: creating report templates is not just busywork, because inconsistency is actually distracting and counterproductive.  When you see a new report, you shouldn't have to search for the run date, and when you flip the pages of a dashboard or analytical interface, you shouldn't need to regain your bearings and relearn the layout.  Consistency allows end users to jump right into consuming the data, and this principle applies equally to the likes of reports, dashboards, analytical applications, and even slide decks and your company's web site.

So pick an approach, given your audience and medium, and stick to it.

  • Replicate your layout on all pages of a report or application, with components aligned and intuitively arranged. Selectors (dropdown boxes, radio buttons) for the same dimensions should be in the same place, the data date should be in the same corner, etc.

By the way, here's a QlikView enhancement idea to make layout consistency a natural part of building applications: create an option on the Layout tab of an object's properties that allows you to choose from a series of checkboxes on which sheets an object should appear, rather than just copying it once per tab.  I imagine it would even save 1) memory, because there would be fewer overall components and 2) calc time, because it wouldn't need to update copies of those components on new tabs when you open them.

  • Have not just consistency of colors, but consistency of meaning of colors. For instance, darker colors might indicate recency or the degree to which a data point is an outlier.


That, of course, precludes you from making psychedelic bar charts, where every slice is arbitrarily assigned a color for the sake of variety.  (This example commits more offenses than just that.)
Junk Charts

    • Pick a standard font, colors and sizes. Choose based on readability and the ability to create some emphasize using contrast.  For instance, everything data-related might be black, while everything else, like help text or metadata, might be a lighter color.
    • For charts, pick a standard alignment for titles and a standard place to display the units and granularity. Making a habit of this improves not only consistency, but provides end users with the context necessary for understanding the data correctly.


  • Apply a standard alignment for tabular data, e.g. left-justified text and right-justified numbers with the same number of decimal places, to facilitate visual comparison.


  • Sort orders should be consistent, particularly if several charts or tables are using the same dimensions.
  • Each page should have the same footprint. In a dashboard, the components on all tabs should fit into a space with the same dimensions, e.g. 1024x768, and reports shouldn't have objects running off of the page.

Some of those sound like common sense, but it's shocking how often they are violated.  Once you have finished your thoughtful planning and are happy with your work, document your choices as the standards upon which future work will be based.

TAGS: Presentation, Reporting, Visualization

Related News

A great dashboard embodies all the same strengths as a great business intelligence platform,...

Read More

A dashboard is often what we envision when we think of business intelligence, and with good...

Read More