Communications of the ACM
Communications of the ACM
Does Color in Email Make a Difference?
From inauspicious beginnings as a means of communication in the earliest days of computer networks some 40 years ago, email has become a killer application helping drive the Internet's momentum. Surging in tandem, they have given rise to a plethora of information in all areas of enterprise operations while vying for the attention of end users. In terms of marketing strategy, these developments mean the time has come for emotionally evocative email.
The email marketing industry was expected to grow from $164 million worldwide in 1999 to $7.3 billion in 2005; meanwhile, the 200 billion marketing email messages in 2005 were expected to surpass the amount of marketing material sent by traditional postal mail in 2002. This means a huge potential business impact from email advertising and marketing but also a major lack of attention, as most recipients ignore many of their email messages. The problem of getting people's attention and making sure they open and read their email can be overcome by making the messages personalized and emotionally evocative [8]. In this context, color can help increase response rates by grabbing attention and evoking positive feelings.
As email messages become more sophisticated in both organization and format, color is likely to play an increasingly important role in their presentation. Unlike color in print, color in digital documents is inexpensive and increasingly accessible on popular platforms (such as on handheld devices like PDAs and cell phones). At the same time, however, little is known about how to use color in email messages.
Here, we examine the impact of background color on recipients' responses to email messages that try to tempt them to travel to a designated link providing more detailed information. We explore the common scenario of people receiving a list of incoming email messages (their inbox) and the decisions they make as to how to react to a particular messageignore it, read its displayed information (such as sender and title), defer treatment, delete, or just click and open. Our concern is with the process, given that a message has been opened and the recipient must decide whether to act in a way that's been suggested in the email. When the message is open, the format of the email message, including its background color, is evident to the recipient, and the recipient is asked to click on a link within its text. Our research question concerns the effect of color on a recipient's decision to travel or not to travel to the recommended link.
Components of Color
To discuss the effect of color, we must examine its three dimensions: hue, chroma, and value. Hue corresponds to the normal meaning of colorits pigment (such as red, green, blue). Chroma (saturation) is the relative amount of pure light that must be mixed with white light to produce the perceived color; colors low in chroma appear dull in comparison to the richness and depth of high chroma. Value (brightness) is the level of lightness of the color; colors high in value appear "whitish," while colors low in value appear "darkish."
We can use color of high value and high chroma to create a message context in which the recipient is more inclined to respond positively to the message.
Color influences our emotions. It has been 75 years since the claim [11] that yellow and pink are superior background colors for questionnaires, most likely to evoke a higher response rate and willingness by survey participants to invest time and cognitive effort. In marketing, too, color has been tied to consumer reactions to advertising in research defining consumer behavior as a function of the value and chroma of the color being used [4] and establishing a link between color and consumer behavior through two dimensions: boredom/excitement and relaxation/tension.
It should be noted in this context that one can be both excited and relaxed at the same time, as the two dimensions are defined separately. On the one hand, we know that people are more inclined to act as suggested in an ad when they are excited rather than bored and when they are relaxed rather than tense. On the other hand, colors high in value (more whitish colors, like pastels) make one feel more relaxed, and colors high in chroma (brighter colors) make one feel more excited. We can thus use color of high value and high chroma to create a message context in which the recipient is more inclined to respond positively to the message.
Research into the link among color, information, and computers is not new but has concentrated on cognitive rather than affective aspects of human behavior. Color is commonly available and easily applied in information and communication technologies to broaden their scope from mere information processing to communication and presentation applications. For example, color in business communications has been shown to enhance reader comprehension by presenting different categories in different colors [12].
Almost 20 years ago, a study of color and information processing [7] concluded, "Color is a subtle variable that can significantly enhance a decision maker's ability to extract information." Going a step furtherbeyond facilitating information extraction and manipulationcolor today is intentionally used to influence emotions. Indeed, one of the potentially most important areas of using color in email is in motivating consumers to take time to read more detailed information about or to outright purchase a product.
How are consumers motivated to access more information? Our working assumption is that they are already overloaded with information, precisely the feeling many of us have when surfing the Web. We need to be tempted to invest more time in reading online material; without such temptation we might skip to the next item, ignoring the message or deleting it without opening it. In the Internet context, an email designer must grab consumers' attention, perhaps by being emotionally evocative.
Background color is an age-old technique for manipulating emotions. Folklore based on anecdotal research tells us that restaurant walls painted in color (such as red) boost one's appetite. Is the same true for online commodities? Although not conclusive, the evidence on the impact of color on consumers' willingness to act seems to indicate that if color does have an impact the most effective colors excite us without making us tense. As noted earlier, the colors (such as pastels) most likely to do this produce warm yet relaxed feelings. Indeed, representative work (see Table 1) provides evidence to support this conclusion, though it also provides evidence to the contrary. Part of the problem of sorting these contradictory conclusions may be the relatively small sample size of studies on response rates, which themselves tend to be low relative to the sample size.
Field Experiment
Our study examined the use of background colors in email messages to motivate recipients to access a more detailed description of a product via a hyperlink. The field experiment tested the effect of background color on a short email message distributed to a sample of 1.4 million consumers, derived from the customer base of an email host service company called IncrediMail (www.incredimail.com). IncrediMail provides mailing capabilities designed to enrich communication; for example, users can choose preferences in animation, voice, background, and e-card text and design and integrate them into their messages.
The IncrediMail customer base consisted of more than four million subscribers in more than 70 countries, of which the major sources were the U.S. (28%), Canada (9%), and Western Europe (30%) (see Figure 1). The age distribution included children up to age 19 (11%); 2029 (27%); 3039 (26%); 4049 (19%); 5059 (11%); and 60+ (5%); 34% of the subscribers were women. The overall population is similar to the general patterns of Internet user populations worldwide, as reported in general user surveys [5]. The results of our field experiment can therefore be viewed as highly generalizable.
In a pre-test we conducted to explore distribution and monitoring mechanisms, we distributed a message to 50,000 IncrediMail subscribers and monitored them for seven days. In that message the distributor wished the recipients a "Merry Christmas," inviting them to view an e-card collection. This paved the way for the main experiment, which was included in an issue of the monthly newsletter (with news items, updates, and links) distributed to all IncrediMail subscribers. We sent the email message in Figure 2 to the 1.4 million subscribers a day after Christmas, then monitored their response for 14 days, up to the beginning of the second week of January. This personalized greeting wished them a "Happy New Year" and invited them to download an attached movie. Each subscriber received the same message with a different color background.
We randomly divided the sample into five groups of 280,000 subscribers each. Each group received the same message with a different background color: white, green, pink, yellow, or blue. Figure 2 lists the definition of the colors we used in terms of hue, chroma, and value. Note they are all relatively high in terms of both value and chroma and are indeed the pastelssoft colors that evoke calm and relaxed feelings. Note, too, that yellow and pink, the traditional favorites, are included. Our strategy was to elicit the superior colors compared with white in the context of email marketing.
Earlier statistics from IncrediMail on its own activities indicate that up to 55% of email recipients open their mail, with over 50% clicking on the link to read a detailed message. We experienced similar rates in our experiment. The overall percentage of experimental recipients who opened their mail was around 30%. The response rate on the first day of distribution was around 4% of the entire sample of 1.4 million subscribers; the next day, the rate peaked at 10%, then gradually declined to less than 0.5% on the 12th and 13th days after distribution. Figure 3 plots the response distribution, which is typical of online surveys. The percentage of experimental recipients who clicked on the recommended link varied from color to color: yellow, 52.80%; green, 52.77%; blue, 51.98%; white, 51.08%; and pink, 50.81%. Yellow was first, but white (the basis for comparison) was not worst (see Figure 3).
We used a statistical test (chi-square) to compare each color with white. Even though the absolute differences in response rates were not significant, the statistical differences were all significant at 0.001. Three of the four preferred colors increased the response rate by up to 1.72% compared to no color at all. Assuming a 5% purchase rate among those who voluntarily traveled to the detailed description, the increase attributed to color represents a 0.2% increase in sales, which can mean a great deal of business through the emailing lists typically employed in email-based marketing.
Conclusion
Color is easy to implement, inexpensive, and increasingly used in email. But how well does it influence a recipient's responsiveness? Color has two main functionsattract attention and set the right moodfor responding positively to a message or request. And because of our increasingly short attention spans and the relatively quick interaction speed we expect in today's electronic world, it must do both at the same time.
Color can be a prime attention grabber when and where people's attention is scarce. Communication on the Web must therefore be emotionally evocative for it to be able to direct attention to the message in question, bearing in mind that color can move people by increasing excitement and lowering tension at the same time.
The right color helps generate feelings that have been found to induce people to respond favorably to requests and marketing communications (such as questionnaires and ads). The pastelsthe class of colors most effective at doing thisare all relatively high in value and chroma (see Table 2). Extending previous research demonstrating the effectiveness of pastels, especially yellow, our study found that color can indeed make a difference, albeit a small one, but one that can be judged statistically significant due to the large sample we used. Moreover, in the age of Internet-based mass communication, small yet reliable differences in response rates translate into substantial business gains.
Designers of mass online communication systems, including advanced forms of email, Web sites, chats, bulletin boards, and pop-up ads, should aim to be emotionally evocative. Color is perhaps the first and most obvious example of how to make a message emotionally evocative, but other elements of design (such as type font and size) may be equally important. A 2003 study of emotional reactions to Web homepages found that several design factors, including graphics, color, and aesthetics, may all be emotionally evocative [9].
Caution is also called for. Affective qualities of human-computer interaction are especially susceptible to cultural differences. For example, the color red in certain Asian cultures and the color green in Islam are usually associated with sacred objects; because inappropriate use of these colors is offensive, we deliberately avoided them in our study, but designers must be aware of these sensitivities [12]. As we begin to understand the factors that help attract attention while generating positive feelings, it is also time to capitalize on this understanding.
References
1. Blythe, I. and Essex, P. Variations on postal theme. In Proceedings of the Annual Conference of the Market Research Society (Boca Raton, FL). The Market Research Society, 1981, 3551.
2. Buttle, F. and Thomas, G. Questionnaire color and mail survey response rate. Journal of the Market Research Society 19, 4 (1997), 625626.
3. Fullerton, S. and Dodge, R. The impact of color on response rates for mail questionnaires. In Proceedings of the Academy of Marketing Science 11. Academy of Marketing Science, Akron, OH, 1988, 413415.
4. Gorn. G., Chattopadhyay, A., Yi, T., and Dahl, D. Effects of color as an executional cue in advertising: They're in the shade. Management Science 43, 10 (1997), 13871400.
5. Graphic, Visualization, and Usability Center's WWW User Survey. Georgia Institute of Technology, Atlanta, GA, 1999; www.gvu.gatech.edu.
6. Hartley, J. Obtaining reprints: Does color help? Science Communication 22, 2 (2000), 212218.
7. Hoadley, E. Investigating the effects of color. Commun. ACM 33, 2 (Feb. 1990), 120125.
8. Jobber, D. and Sanderson, S. The effects of prior letter and colored questionnaire paper on mail survey response rates. J. Market Res. Society 25, 4 (1983), 339349.
9. Kim, J., Jooeun, L., and Dongseong, C. Designing emotionally evocative homepages: An empirical study of the quantitative relations between design factors and emotional dimensions. Int. J. Human-Comput. Stud. 59, 6 (Dec. 2003), 899940.
10. LaGrace, R. and Kuhn, L. The effect of visual stimuli on mail survey response rates. Industrial Marketing Management 24 (1995), 1118.
11. Reed, V. Planned Marketing. Ronald Press, New York, 1929.
12. Te'eni, D., Carey, J., and Zhang, P. Human-Computer Interaction: Designing Effective Organizational Information System. John Wiley & Sons, Inc., Chichester, U.K., 2006.
Figures
Figure 1. Distribution of the customer base by country.
Figure 2. The message sent in a variety of background colors.
Figure 3. Distribution of responses for each email message sent. Colors indicate the background color of the message.
©2006 ACM 0001-0782/06/0400 $5.00
Permission to make digital or hard copies of all or part of this work for personal or classroom use is granted without fee provided that copies are not made or distributed for profit or commercial advantage and that copies bear this notice and the full citation on the first page. To copy otherwise, to republish, to post on servers or to redistribute to lists, requires prior specific permission and/or a fee.
The Digital Library is published by the Association for Computing Machinery. Copyright © 2006 ACM, Inc.
No entries found