UI Response Times
This article is intended as a reference for defining and characterizing response times in a user interface. The classifications are derived from cognitive psychology, neuroscience, and human factors research into human perceptive abilities in combination with first principles of UX design.
Response Time is only one factor involved in how a UI unfolds over time. Two other factors are Latency Indication, and UI Animation. This article offers a cursory recommendation for indicating latencies. A full treatment of Latency Indication, however, includes classification of latencies across more dimensions that just duration, and makes recommendations for appropriate indicators and how to escalate the indicator as the latency progresses and crosses duration boundaries.
UI Animation — distinct from animation for art and storytelling — is a design device for maintaining context across UI display changes, controlling visual attention, and providing direct-control feedback. While timing is a major factor in UI animation, it too includes other dimensions such as the relationship among UI elements, object permanence, movement gestalts, etc., and demands its own treatment.
Look for separate articles covering Latency Indication, and UI Animation.
Some of the longer response times discussions are a work in progress, and will be updated over time.
Note: The primary reference for response times less than 10 minutes is Jakob Nielsen’s Alertbox, as indicated by the embedded quotes from that source in the appropriate sections.
Content for response times 10 minutes or greater is original.
0.1 second
The threshold of human perception.
“0.1 second is the response time limit if you want users to feel like their actions are directly causing something to happen on the screen… Thus, to create the illusion of direct manipulation, a user interface must be faster than 0.1 second.”
– Jakob Nielsen, Alertbox, www.useit.com
Special feedback required: None. The system response itself is the feedback.
Special interaction considerations:
Cases where response times this fast may are absolutely required:
- mouse movements
- keypresses
- touch interactions (selection, drag movements, gestures, etc.)
Cases where response times this fast may not be necessary, or appropriate:
- When the illusion of direct manipulation is not required or expected of the interface being designed; when it’s important to make it clear the computer is responsible for completing an action
Cases where response times this fast or faster should be avoided:
- When a human is expected to comprehend or react to interface changes or content (e.g. status messages, scroll rate, progress meters, etc.)
1 second
The limit for experiencing the interaction as being one continuous flow.
“…it feels like the computer is causing the result to appear. Although users notice the short delay, they stay focused on their current train of thought during the one-second interval… during 1-second response times, users retain the feeling of being in control of the interaction even though they notice that it’s a 2-way interaction (between them and the computer).”
– Jakob Nielsen, Alertbox, www.useit.com
Special feedback required: None. Note, however, the feeling of direction manipulation is absent.
Special interaction considerations: n/a
10 seconds
The limit of short-term memory. Also the limit for keeping attention focused on the ongoing interaction.
“…users get impatient and notice that they’re waiting for a slow computer to respond… after about 10 seconds, the average attention span is maxed out. At that point, the user’s mind starts wandering and doesn’t retain enough information in short-term memory to easily resume the interaction once the computer finally loads the next screen.”
– Jakob Nielsen, Alertbox, www.useit.com
Special feedback required:
- 2–5 secs. — simple busy indicator
- >5 secs. — true percent-done progress indicator
Special interaction considerations: Interactions in the 2–10 second timeframe represent the boundary between a “responsive” system and a “slow” system.
At the short end of this spectrum (2–5 seconds) the latency is experienced like a pause in the ongoing dialogue: noticeable, but not long enough to break concentration, nor to fill in the idle gaps with other tasks. Attempts to shorten the response time should be made, of course, but latencies in this range are acceptable for interactions that proceed in the form of an ongoing dialogue and do not require direct manipulation characteristics (see 0.1 second.)
At the long end of the spectrum (5–10 seconds) the latency is starting to be experienced like a delay: long enough to break concentration, but still not long enough to fill in idle time. This is the worst experience in the entire spectrum of response times! Attempts should be made to either shorten the response time into the range of a “pause”, or modify the interaction model to make longer “breaks” in the interaction both possible and expected.
All latencies greater than 10 seconds in response to interface actions should be treated as “breaks” in the dialogue. The interaction model should allow other tasks to fill in the idle time, and the original task should support interruption and resumption without relying on the person’s short-term memory.
1 minute
The upper limit for completing simple tasks. Also the limit for video content that is supplementary to the primary task.
“Users should be able to complete simple tasks in about 1 minute. Likewise, most Internet videos should last no more than 1–2 minutes because people don’t like passively watching something for much longer than that while they’re in the active frame of mind induced by Web surfing.”
– Jakob Nielsen, Alertbox, www.useit.com
Special feedback required:
- Step-wise progress indicator (for an activity comprised of a series of tasks)
- Percent-done progress indicator (for a single task that is effectively a “break” — see 10 seconds)
10 minutes
Discussion TBD
Examples TBD
Special feedback required:
- Check-list progress (for an activity comprised of a series of tasks)
- Percent-done progress indicator (for a single task that is effectively a “break” — see 10 seconds)
1 hour
Discussion TBD
Examples TBD
Special feedback required:
- Check-list progress (for an activity comprised of a series of tasks)
- Percent-done progress indicator (for a single task that is effectively a “break” — see 10 seconds)
- Notifications
1 day
Discussion TBD
Example:
Collaboration workflows among actors.
Additional examples TBD
Special feedback required:
- Check-list progress (for an activity comprised of a series of tasks)
- Notifications
- Status display
1 week
Discussion TBD
Example:
Collaboration workflows among actors.
Additional examples TBD
Special feedback required:
- Check-list progress (for an activity comprised of a series of tasks)
- Notifications
- Reminders
- Status display
1 month
Discussion TBD
Example:
Collaboration workflows among actors.
Additional examples TBD
Special feedback required:
- Check-list progress (for an activity comprised of a series of tasks)
- Notifications
- Reminders
- Status display
1 year
Discussion TBD
Examples TBD
Special feedback required:
- Check-list progress (for an activity comprised of a series of tasks)
- Notifications
- Reminders
- Status display
References
Nielsen, Jakob, “Powers of 10: Time Scales in User Experience,” Alertbox, October 5, 2009, www.useit.com, (referenced on 1/15/2010 at http://www.useit.com/alertbox/timeframes.html)
Nielsen, Jakob, “Response Time Overview,” Excerpt from ch. 5 in Usability Engineering, (referenced on 1/15/2010 at http://www.useit.com/papers/responsetime.html)
Wikipedia, “Reaction time,” (referenced on 1/15/2010 at http://en.wikipedia.org/wiki/Reaction_time)
Wikipedia, “Human timescales,” (referenced on 1/15/2010 at http://en.wikipedia.org/wiki/Human_timescales)
