When an IVR puts a caller in a queue, that caller must wait for service. People are not equally sensitive to the perception of lost time when waiting for service (Kleijnen, de Ruyter, & Wetzels, 2007). Those who are sensitive experience a buildup of anxiety and stress due to a sense of waste and uncertainty inherent in the waiting situation (Osuna, 1985). Findings from psychological studies of waiting for service (Durrande-Moreau, 1999; Unzicker, 1999) indicate:
When placing callers into a queue, to manage the caller experience, it is important that callers:
Provide an expected wait time
Most IVRs provide callers with an expected wait time, especially if the estimated wait is over some criterion, such as 30 seconds. Knowing an expected wait time decreases anxiety during the wait. Furthermore, knowing the expected wait time helps the caller decide what to do while waiting. Although the system should provide audio to fill time while the caller waits (next section), many callers on hold pay minimal attention to the system while waiting, and typically pursue secondary time-filling activities such as talking with others in the room, working on email, watching television, etc. (Kortum & Peres, 2007). So, as long as the system can provide a reasonably accurate expected wait time, it should do so.
Optionally provide a Web address, but only if it will be of use to the caller
We generally advise against providing Web addresses in the introduction to an IVR, but when the caller is in the process of going on hold, it seems less onerous to remind them of the enterprise's Web presence. Note, however, that we do not know of any published research on this topic. After all, many of the arguments against playing a Web address in an introduction also apply to this point in the interaction – callers have chosen the phone channel over the Web for a reason and already know that the enterprise has a Web site.
Consider a situation where the IVR offers a reasonable chance of self-service, but for whatever reason the caller needs to transfer, and the call center is closed. This would be a good place to let them know about self-service available on the Web, especially if it is known that their task could be done online.
Use music to fill the wait
There has been quite a bit of research on the effect of music on the waiting experience (for a review, see Lewis, 2011, pp. 62-71). The research generally supports the importance of playing music during the wait. Music does not appear to reliably reduce the perception of how much time has passed, but it can make the wait more pleasant, which then increases satisfaction with the wait and attitude toward the service provider.
Callers prefer music to simple audio tones (e.g., ticking) or natural sounds (e.g., pouring water) (Fröhlich, 2005; Polkosky, 2001; Polkosky & Lewis, 2002). Regarding the type of music to play, the only research on this topic of which we are aware is Ramos (1993), which suffered from a critical flaw in its experimental design making its results uninterpretable (Lewis, 2011). The Lewis (2011) reanalysis of the data from Ramos (1993) suggests that the different styles of music studied (classical, popular, relaxation, country, and jazz) had no apparent effect on lost call rates.
There is some evidence that the hold experience is better when music plays that the caller likes. In general, callers prefer music in a major key that matches their expectation of how hold music should sound. Without going overboard, try to match the style of hold music to your typical caller. “The complexities of music and changing styles will keep the selection of hold music more of an art than a science” (Lewis, 2011, p. 70).
If it is possible for callers to experience unusually long hold times, there should probably be some stylistic variation in the music. We know of no published studies, but it seems reasonable that callers will tire of any short musical piece (say, 3 minutes in duration) looped continuously for 15-20 minutes. If you have to deal with this as part of your design, make sure you have an idea of how long the hold times can be and plan accordingly for a long hold.
Generally, avoid advertisements and other verbal messaging once the music starts
Callers appear to find commercials irritating, interpreting them as holding up progress while the ad plays (Fröhlich, 2005). Simple apologies appear to at best provide no measurable benefit to the hold experience and at worst may annoy callers (North et al., 1999). Each time a verbal message begins to play, it potentially tricks the caller into thinking an agent has taken the call, only to immediately disappoint the caller when this turns out not to be the case. See the next item for an exception.
For long expected wait times, consider playing verbal messages that provide a sense of progress in the queue
Verbal messages that provide a sense of progress in the queue appear to be effective in reducing abandonment rate and increasing caller satisfaction (Munichor & Rafaeli, 2007). To date, we know of no research on the optimal timing of such messages, and they seem likely to apply primarily when expected wait times are fairly long. There is no data on how long an expected wait should be for these messages to be effective rather than disruptive or annoying.
There are two ways to provide verbal information about progress in the queue – the number of people ahead of the caller or the time remaining before service. No published experiments have directly compared these two methods. Rational analysis, however, suggests that the better choice is to tell callers how much longer they should expect to wait. For callers to effectively use information about the number of people ahead of them in line, they would need to know how many service representatives are available to take calls and how long it typically takes to provide service. Customers have this information when they are standing in a line at a bank, but not when they are on the phone.
Avoid auditory progress bars
Several researchers have experimented with auditory progress bars (APBs), which are “tonal cues that continuously convey information about the time left in the hold queue” (Kortum et al., 2005) – such as pitch or volume. There is some evidence that some types of APBs can reduce the perceived duration of a wait, but caller reaction tends to be strongly against them (Fröhlich, 2005; Kortum et al., 2005), making them unsuitable for use in IVRs (at least, in their current forms).
Kortum and Peres (2006) conducted a study of an alternative to APBs, matching known musical selections from the Billboard Top 40 to estimated wait times so the song fit in the expected wait. Participants in that experiment estimated wait times as accurately as with the best APBs (about 50% overestimation) and rated satisfaction with a musical filler significantly better than any of the tonal APBs.
Once on hold, callers typically wait until they receive service or until they get tired of waiting and hang up. Most systems do not include any interaction with the caller while on hold, and are therefore unable to respond to requests for an agent. In fact, it’s very likely a request for an agent that got the caller on hold in the first place.
For this reason, design your IVR so it does not mysteriously place callers on hold. Follow the preceding guidelines to make the on-hold experience as pleasant as possible, given that no one wants to wait for service.
Durrande-Moreau, A. (1999). Waiting for service: Ten years of empirical research. International Journal of Service Industry Management, 10(2), 171–189.
Fröhlich, P. (2005). Dealing with system response times in interactive speech applications. In Proceedings of CHI 2005 (pp. 1379–1382). Portland, OR: ACM.
Kleijnen, M., de Ruyter, K., & Wetzels, M. (2007). An assessment of value creation in mobile service delivery and the moderating role of time consciousness. Journal of Retailing, 83(1), 33–46.
Kortum, P., & Peres, S. C. (2006). An exploration of the use of complete songs as auditory progress bars. In Proceedings of the Human Factors and Ergonomics Society 50th annual meeting (pp. 2071–2075). Santa Monica, CA: HFES.
Kortum, P., & Peres, S. C. (2007). A survey of secondary activities of telephone callers who are put on hold. In Proceedings of the Human Factors and Ergonomics Society 51st annual Meeting (pp. 1153–1157). Santa Monica, CA: HFES.
Kortum, P., Peres, S. C., Knott, B. A., & Bushey, R. (2005). The effect of auditory progress bars on consumer’s estimation of telephone wait time. In Proceedings of the Human Factors and Ergonomics Society 49th annual meeting (pp. 628–632). Santa Monica, CA: HFES.
Lewis, J. R. (2011). Practical speech user interface design. Boca Raton, FL: CRC Press, Taylor & Francis Group.
Munichor, N., & Rafaeli, A. (2007). Numbers or apologies? Customer reactions to telephone waiting time fillers. Journal of Applied Psychology, 92(2), 511–518.
North, A. C., Hargreaves, D. J., & McKendrick, J. (1999). Music and on-hold waiting time. British Journal of Psychology, 90, 161–164.
Osuna, E. E. (1985). The psychological cost of waiting. Journal of Mathematical Psychology, 29, 82–105.
Polkosky, M. D. (2001). User preference for system processing tones (Tech. Rep. 29.3436). Raleigh, NC: IBM.
Polkosky, M. D., & Lewis, J. R. (2002). Effect of auditory waiting cues on time estimation in speech recognition telephony applications. International Journal of Human-Computer Interaction, 14, 423–446.
Ramos, L. (1993). The effects of on-hold telephone music on the number of premature disconnections to a statewide protective services abuse hot line. Journal of Music Therapy, 30(2), 119–129.
Unzicker, D. K. (1999). The psychology of being put on hold: An exploratory study of service quality. Psychology & Marketing, 16(4), 327–350.