{"id":20491,"date":"2021-10-19T14:58:57","date_gmt":"2021-10-19T18:58:57","guid":{"rendered":"https:\/\/www.bloomberg.com\/company\/stories\/designing-the-terminal-for-color-accessibility\/"},"modified":"2025-06-03T08:31:20","modified_gmt":"2025-06-03T12:31:20","slug":"designing-the-terminal-for-color-accessibility","status":"publish","type":"post","link":"https:\/\/www.bloomberg.com\/company\/stories\/designing-the-terminal-for-color-accessibility\/","title":{"rendered":"Designing the Terminal for Color Accessibility"},"content":{"rendered":"
\n \n\t\n\t\n\t
\n\t\t\n\t\t\t
\n\t
\n \n

We often think we see user interfaces the same way as everyone else. However, for users with color vision deficiencies, certain color combinations can have detrimental effects on their use of the Bloomberg Terminal. We estimate that 20,000 of our users have Color Vision Deficiency (CVD), and we must not overlook their needs as we design and roll out new interfaces.<\/p>\n

What is Color Vision Deficiency?<\/h2>\n

Some people know it as \u2018color blindness,\u2019 but that is actually a misnomer, since people who have CVD aren\u2019t necessarily \u2018blind\u2019 to colors. Rather, they have a deficiency in which one or more cone cells in their retinas (the cells that detect colors) have reduced or no sensitivity to specific wavelengths of light (Image 1).<\/p>\n\n<\/div>\n

\n \"Image\"Image\n
Image 1: A comparison of the same image simulating Normal Vision and Color Vision Deficiency. Left: Colored markers shown with Normal Vision have distinctive hues; Right: The same image simulating CVD shows the markers in less distinctive hues.<\/figcaption>\n<\/figure>\n
\n \n

The scientific terms are dichromacy<\/strong> (a type of color deficiency where one cone cell is completely absent – for example, if the green cone cell is absent, it is called Deuteranopia<\/em>) and anomalous trichromacy<\/strong> (a type of color deficiency where one cone cell has a reduced sensitivity to certain wavelengths of light – for example, having reduced sensitivity to green light is called Deuteranomaly<\/em>). These color vision deficiencies can take many forms (Image 2). The most common form of CVD affects the red and green wavelengths of the color spectrum.<\/p>\n

This deficiency certainly affects people\u2019s everyday experiences. Imagine not knowing if your clothing matches, accidentally eating unripened (green) bananas, or being unable to tell one sporting team\u2019s jersey from another when watching a game on TV.<\/p>\n\n<\/div>\n

\n \"Image\"Image\n
Image 2: A comparison of the same image simulating Normal Vision and the three main types of Color Vision Deficiency. Clockwise from Top Left: Normal Vision, Red-Green Confusion, Monochromacy (can\u2019t perceive color at all), and Blue-Yellow Confusion.<\/figcaption>\n<\/figure>\n
\n \n

There are varying prevalence rates of CVD<\/a> across ethnicities (due to its genetic component) and genders (with CVD predominantly affecting males). However, even by conservative estimates, 20,000 or more of our Terminal users could have the most common form of CVD, Red-Green.<\/p>\n

Why is this an issue for our product?<\/h2>\n

In the financial industry, red and green are key visual indicators, because they express positive and negative semantic meanings in Western conventions (Image 3). For example, green generally indicates an \u201cup\u201d market status, while red generally indicates a \u201cdown\u201d market status.<\/p>\n\n<\/div>\n

\n \"Image\"Image\n
Image 3: Example of a Bloomberg Terminal Launchpad screen showing the system\u2019s default color scheme. Red colors are associated with \u201cdown\u201d market status and green colors are associated with \u201cup\u201d market status.<\/figcaption>\n<\/figure>\n
\n \"Image\"Image\n
Image 4: Terminal visualization that shows green and red color indicators next to an overlay simulating Deuteranopia. The Deuteranopia simulation exposes that the red and green colors appear very similar in hue, and makes it challenging to quickly parse the two different color indicators.<\/figcaption>\n<\/figure>\n
\n \n

Software programs can simulate the approximate colors that a person with CVD might see. Using this CVD simulator (as shown in Image 4), we can see what an example from the Terminal looks like from the perspective of a user with Deuteranopia. While there still are individual differences between different people, it’s clear that these reds and greens are difficult for these users to distinguish.<\/p>\n

This could potentially have large financial repercussions, when using our system to make a decision to trade millions of dollars of securities could be affected by these important visual indicators.<\/p>\n

What about other products?<\/h2>\n

Operating systems, such as Microsoft Windows or Apple macOS, have enhanced contrast or increased text legibility support to aid in accessibility. These settings cater to a more general approach in an attempt to cover multiple types of vision deficiencies. Apple iOS also includes a small set of preset colors that can be selected for specific CVD types, but this has disadvantages in that it changes specific colors that are crucial in the financial industry. Because of this, we needed to take a researched approach to make our product more inclusive.<\/p>\n

Our research process<\/h2>\n

To better understand how to implement CVD themes in the Bloomberg Terminal, we first needed to evaluate the user experience of our clients with CVD.<\/p>\n

Once we found clients who had self-reported CVD, our next step was to assess which type of CVD they had, and how mild or severe it was. We did this by administering the Ishihara test for color deficiency<\/a>, a well-known color differentiation test, over the internet.<\/p>\n\n<\/div>\n

\n \"Image\"Image\n
Image 5: Example of an Ishihara color test plate. The number “74” should be clearly visible to viewers with Normal Color Vision. Viewers with Deuteranopia or Protanomaly may not see a number at all or they may see a \u201c71\u201d.<\/figcaption>\n<\/figure>\n
\n \n

It\u2019s important to note that this color differentiation test is not perfect, and that test results can vary from what users initially report. To understand how CVD affects the Terminal user experience, we invited these clients into our usability lab and had them evaluate a series of different Terminal image types (e.g., graph lines, colored text, buttons, bars, and heat maps). Clients\u2019 task performance was measured, in addition to self-reported confidence scores of their performance, and overall preferences. Linear mixed effects model analyses showed statistically significant differences between the different color sets, such that clients were more accurate, reported greater confidence, and preferred the alternative and high contrast color sets compared to the current color set.<\/p>\n

We concluded from this study that clients with CVD can distinguish alternative or high contrast colors on the Terminal better than the current color set.<\/p>\n

Once we understood that certain color sets affected our clients\u2019 task performance and confidence, we wondered, \u201cWhich colors would make the Terminal easier to use for clients with CVD?\u201d To answer this question, we needed to go outside the usability lab and interview clients at their own desks to understand how their perceptions of Terminal colors are influenced by the context of their environments.<\/p>\n

We interviewed clients using the same image types we evaluated in the previous study. However, this research differed from the previous study in that we enabled clients to run several alternative and high contrast color sets on functions within their Terminal windows, and to provide feedback on these alternative and high contrast color sets in real-time, and in their actual work environments.<\/p>\n

Based on this feedback, we found that clients\u2019 perceptions of Terminal colors were affected by how mild or severe their CVD was. For instance, those who had a milder form of CVD tended to have fewer issues differentiating colors, but did have minor issues with differentiating between thinner red and green lines.<\/p>\n

Those who had moderate CVD experienced more difficulties with colored images, particularly those with reds and greens blending together. They also had more difficulty distinguishing between darker colors on dark backgrounds.<\/p>\n

Finally, those with severe CVD had issues differentiating between different hues of colors like yellows, purples, pinks – in addition to standard greens and reds.<\/p>\n\n<\/div>\n

\n \"Image\"Image\n
Image 6: The severity level of CVD results in escalating challenges parsing different types of visual communication with color. For example, Mild CVD complicates interpreting the color of small text and thin lines, Moderate CVD additionally makes darker shades of red and green harder to disambiguate, while Severe CVD encompasses all of the former issues, in addition to muddying the distinction between more hues.<\/figcaption>\n<\/figure>\n
\n \n

Our research also uncovered that some colors had different meanings in different contexts.<\/p>\n

While users with CVD may have trouble distinguishing between different color combinations, their semantic color associations persisted. For example, greens and blues tended to be self-reported as denoting an \u201cup\u201d market status, while reds, yellows, and oranges tended to denote a \u201cdown\u201d market status.<\/p>\n

Our solution<\/h2>\n

Once we gathered these learnings, we knew that our design solutions needed to be:<\/p>\n