The number of users who wear smartwatches is on the rise. As of 2020, Pew Research Center estimated that 1 in 5 Americans had a smartwatch or fitness tracker of some kind.

Users prefer brief and simple interactions on smartwatches, largely because of the tiny screens. Users do not expect to access complex information on the watch, so designers must be extremely judicious about what they try to fit into a smartwatch application (if it’s worth building one at all).

Smartwatch Interaction Types

In a 5-day diary study, we asked participants to send us examples of things they do with their watches (Apple Watch, Samsung Galaxy Watch, and Pixel Watch). We collected over 200 examples of individual interactions from 11 different participants. Across all the interactions submitted, we encountered 6 main types:

  1. Receiving: Asking for the user's attention to present information
  2. Referencing: Checking information that is constantly available
  3. Recording: Capturing data as it is generated in the world
  4. Controlling: Manipulating an ongoing process or separate technology
  5. Communicating: Connecting with other people via calls or messages
  6. Guiding: Providing in-the-moment direction during an activity
Depictions and descriptions of the 6 types of smartwatch interactions described in this article.
Each type of interaction describes broad groupings of user intentions and behaviors. One application can make use of multiple types of interactions.

Below, we explain each interaction type with examples and recommendations.

Receiving Interactions

Receiving interactions occur when the watch asks for users’ attention — most often in the form of notifications. Receiving interactions are by far the most common interaction type — a finding that aligns with third-party research. The receiving-interaction category includes some of the most positive and negative interactions reported.

People prefer receiving information on a smartwatch more than on a smartphone for two reasons:

  • Unobtrusive. A notification on the smartwatch generally comes silently (many people keep their watch in silent mode to avoid bothering other people). It is also more socially acceptable to quickly check something that appears on the watch than to pull out a phone.
  • Hard to miss. A smartwatch is attached to the body and generally vibrates when a notification is received. People in our study frequently commented that they would have missed all kinds of information if they had relied solely on their phones.

Types of Receiving Interactions

Updates

Updates are notifications that keep users in the loop about events happening outside of their control. Although users may have initiated or authorized the updates at one time, updates are not triggered by current user actions. In our study, examples of updates included notifications about the number of views of a product on eBay, likes on an Instagram post, and matches with another person on Tinder.

A notification on an Apple watch showing stock market activity from Robinhood
This update about the performance of a particular cryptocurrency was seen as unhelpful and distracting while the participant was praying.

Reminders

Reminders are notifications that are scheduled by users. Many users rely heavily on reminders for events like a due date for a school assignment, a medication that needs to be taken, or an alarm to wake them up from a nap. They don’t always trust that they will notice the reminder on the phone, but they are confident they’ll notice it on the watch.

A reminder on an Apple watch to attend an open house at 1:00PM
A participant commented: This morning, when I woke up, the first thing I did was to set a reminder to remind me that I had an open house. And I did this because I have kids [...] and sometimes, you know, life happens, and you get busy doing one thing or another. And I wanted to make sure I didn't forget.
An announcement on a Samsung Galaxy watch to submit an upcoming school assignment
This reminder helped one participant remember to complete a school assignment before it was due.

Feedback

Feedback is information about a user’s recent behaviors (e.g., activity levels or screen time). Feedback is often sent at predetermined intervals — like the beginning of a day or week. Feedback often reveals the user’s progress toward a goal, such as the number of steps taken in a day. While sometimes it can be demotivating instead of motivating, feedback is generally considered valuable because it is highly personalized and relevant.

An Apple Watch interface summarizing the week's physical activity levels
A participant commented: My week started with my move goal and my horrible performance last week, so that's very demotivating. But it's at least something to work on this week, and I do find it helpful in keeping me active.
An Apple Watch interface with an encouraging message to drink more water
A notification from Water Llama provides the user with feedback regarding their progress toward a daily water intake goal and prompts them to take another drink.

Suggestions

Suggestions propose an action in response to triggers such as time of day, recent physical activity, or current geographical location. For example, the watch may suggest recording a physical activity such as a walk.

Suggestions are extremely powerful when the timing and context are right because they will seem personalized, responsive, and aware of the user’s current needs and activities. However, they are practically useless if their timing is off. Thus, when the suggestion to record a workout appeared during the workout, its perception was positive, but when it appeared after the workout or during a different activity, it was useless.

Suggestions often prompted recording or guiding interactions (see below).

An notification from Google Maps suggesting a route to walk home
One participant found this suggested route from Google Maps helpful because it came to her right as she got off a bus and was heading in that direction.

Promotions

Promotions are almost always negative. Users do not want promotional content on their watches — particularly when it is generic, seems random, and has no personal relevance. Examples from the study included special offers by Uber Eats and flash deals from Costco.

A notification encouraging the user to gamble
This participant had no interest in this prompt that was received early on a Monday morning and that was luring him back to a rarely used gambling application.

Attributes of Successful Receiving Interactions

Receiving interactions are most helpful when they are:

Informative. When notifications do not have enough useful information, they are annoying because the user must go to the phone to find out everything they want to know.

Glanceable. The most relevant information needs to be easy to read and digest in 2–3 seconds.

Personalized. When notifications are generic or promotional, they are seen as intrusive and annoying on the watch. However, those that contain personally relevant information are seen as helpful.

Timely. When notifications arrive at the optimal moment, they can be incredibly valuable. When they come at the wrong moment, they are generally annoying and useless.

A notification which allows a user to delay the reminder for 15 minutes
How We Feel acknowledges that its prompt might not have come at the optimal moment for the user and allows them to delay it for 15 minutes. This is one method of increasing the user’s autonomy and the chances that they will interact with the notification at all.

Information received on the watch has enormous potential to be helpful and valuable but must be chosen carefully. Users seem to think about the smartwatch as a filter: they are less likely to tolerate irrelevant information coming to the watch than to any other device. It’s like the difference between a salesperson trying to stop you on a busy street (a promotional email) and a salesperson approaching you while you are at a restaurant (a promotional notification on the smartwatch). Both are annoying, but the second feels much more intrusive.

Because the smartwatch is a personal companion, people assume that notifications received on the watch will be relevant; they are very annoyed when they are not. One participant commented, I got this notification while I was working and it literally distracted my thought [...] and I felt, ‘okay, if I've received a message, it might be important.’ And then I realized it was not a message that was helpful at that moment of time.

Receiving interactions can also become annoying when they come regularly (such as the reminder to stand when the watch senses you’ve been sitting too long). Some participants commented that they knew they could disable certain regular notifications, but they simply hadn’t done so. The interaction cost of disabling notifications was too high, so the frequent notifications continued bothering the user and hurting their perception of the sender. Making it easy for users to stop notifications that they don’t want.

A notification which allows users to disable further similar notifications
Google Maps makes it easy to turn off the prompts to review new locations that users visit.

In many cases, receiving interactions precede another type of interaction, such as controlling or recording interactions. For example, receiving a text message might trigger a response, or seeing what’s on the security camera might prompt you to unlock a door.

Referencing Interactions

Referencing interactions occur when users check the information available for reference on the smartwatch. Practically, everything that can be referenced on the smartwatch is also available on the smartphone, but the watch serves as a convenient and accessible medium for quickly checking information. In some cases, referencing interactions are prompted by receiving interactions.

Attributes of Successful Referencing Interactions

Referencing interactions are most helpful when they are:

Simple. Since people often just glance at their watches to see the status of something, they do not expect to think deeply. Prioritizing information critical for decision making and highlighting it prominently is more helpful than providing lots of nuanced details — save the details for the phone.

A Pixel Watch interface displaying minimal information about the current weather in San Diego
Rather than trying to squeeze all manner of weather data into one screen, this Pixel Watch interface prioritizes basic weather information. More detailed information is available as the user swipes up or down. The order of the screens prioritizes the most referenced information.

Visual. Information can often be processed more quickly and efficiently when it is represented visually rather than as numbers or words. Visuals help people quickly gauge the current status of something in relation to a goal, for example.

A Pixel Watch interface showing a visualization of progress toward activity goals
Providing a visualization of how close the user is to achieving their goals is very easy to understand and quick to interpret.

Accessible. Smartwatches are not well-suited for deep information architectures. Users do not want to go looking for information they’d like to reference, so they appreciate it when it lives close to the surface within flat information hierarchies. It’s even better when it is available as a widget that can be added to the home screen.

A Samsung Galaxy Watch home screen with lots of added widgets
The more quickly and easily users can check the information they want to reference, the more valuable it is. This user placed key pieces of information she liked to reference on her home screen.

Personal. Although the smartwatch can track many kinds of data, users are interested in only what’s personally meaningful to them. For example, allowing users to create custom goals and check their progress toward them is one of the most meaningful referencing interactions.

A watch interface showing when someone's next menstrual period will be
Multiple participants found it helpful to check the estimated date for their next period.

Recording Interactions

Recording interactions occur when users choose to track specific events or actions, such as exercising, drinking water, or meditating.

Most of what gets recorded with a smartwatch in this way would have occurred whether the user owned a smartwatch or not. However, it’s likely that watch-wearing users choose to perform certain trackable behaviors more frequently because they can track them and because the watch prompts them. Many behaviors tracked with smartwatches are well-being behaviors that are beneficial for one’s physical or mental health.

Recording interactions describe only the nonautomatic tracking of data (e.g., tracking each individual workout). Users must explicitly perform the recording. Recording interactions do not include constantly tracked data such as heart rate, although users did need to grant the watch permission to track this data initially. Constantly tracked data is simply referenced as needed (see referencing interactions).

Data recorded on smartwatches can be rich or simple, depending on the activity being tracked. For example, the data associated with tracking a workout is a lot richer than that corresponding to tracking water intake.

Recording rich data is helpful for making in-the-moment decisions during the activity, such as adjusting one’s running pace. It can also provide a detailed review of the activity once finished —for example, recording a night’s sleep can be used to understand sleep quality in the morning.

Watch interfaces showing data about a run as it occurs, and a summary of a run after it's concluded.
Users commonly record rich data during exercise sessions to be able to reference real-time data throughout (left) and to see a summary of their performance at the end (right).
Two interfaces of data summarizing a night's sleep quality
Users can track their sleep to see their quality of rest each individual night. While many applications will automatically track the data, users must make sure to charge the watch and wear it to bed to make sure the sleep is recorded.

Recording simple data is most helpful for understanding behaviors over time. For example, a smartwatch is a convenient way to record calories consumed at each meal. Each recording interaction involves inputting a number that says little about overall health, but the aggregated data over time can be very informative.

Interfaces allowing users to track calories consumed and water intake
A calorie tracker (left) and water intake tracker (right) make it simple and quick to input data throughout the day. Their value is found in the aggregated data at the end of longer periods.

Attributes of Successful Recording Interactions 

Recording interactions are most helpful when they are:

Contextually prompted. Users are delighted when the watch recognizes that they are performing a particular behavior. Even when it is difficult for the watch to confidently determine what the user is doing, it can still prompt users to record a behavior if it seems they might have forgotten to do so (see the suggestion type of receiving interaction above).

An Apple Watch suggestion to record a workout
When the smartwatch detected activity characteristics of specific workouts and prompted users to record them, users were grateful. They did not want to miss recording these types of events.

Accurate (or at least perceived as such). Users are often skeptical about how accurately the watch tracks their activities. While the watch may provide data that is not accessible in any other way, most users take it with a grain of salt. One of the biggest downfalls of recording interactions is when the data captured feels too inaccurate, and the user loses trust in it — sometimes because the user has the wrong mental model of how the recording works.

For example, while trying to record her blood-oxygen levels, one user commented, I've had scores saying that I'm at like 82%, which is basically dead, so not super accurate. I don't really like this [functionality]. […] I feel like it's more annoying than anything else because it isn't accurate.

Informative. Users often record data so they can use it to make real-time decisions. They should be able to easily reference the real-time rich data being recorded during an activity.

Interfaces showing workout data while in process, and a timer counting down
Users like to see real-time information such as time spent and calories burned during a workout (left) or time remaining on a timer (right). This information can be used to make decisions about subsequent actions.

Easy to initiate (and subsequently access). Sometimes users need to initiate a recording interaction at a moment’s notice. They cannot wait for the watch to detect what is happening, nor can they waste time searching for the button to initiate the recording. Allowing users to add widgets to their watch face or use intelligent assistants to initiate a recording interaction makes these interactions as useful as possible.

Regarding the ease of tracking water intake, one user commented, All I have to do is swipe screen to the left and it comes right up. It's very easy to use. [...] I like that I can keep track of that very easily with two simple [gestures:] one swipe and a tap.

Additionally, users cannot be bothered to navigate the watch interface to find the real-time data that is being recorded during an activity. This information must live close to the surface while the activity is ongoing.

An Apple Watch home screen with multiple exercise widgets
One participant added all her activity and fitness applications to the home screen because she uses them daily and wanted fast accessibility to them.

Controlling Interactions

Controlling interactions are used to start, stop, alter, or otherwise take control of an ongoing process or of another technology — for example, skipping a song, locking a door, disabling a camera, or setting an alarm. Controlling interactions included some of the most positive smartwatch functions captured in our diary study.

Users often seem to think of their watch as a remote control that allows them to easily make things happen from anywhere — in many cases, avoiding the need to locate or unlock their phone. This isn’t surprising given what we know about device inertia, which is the tendency to complete a task on a suboptimal device because users don’t want to spend the effort to find and switch to a device better suited to the task. The tradeoff between accomplishing tasks on mobile phones and computers is not new, but this study revealed a third level of device inertia: watch vs. phone.

Controlling interactions are valuable because they:

  1. Don’t require users to have their phones handy
  2. Offer quicker, more convenient actions than the phone
  3. Allow users to take actions while doing something else

Participants frequently commented positively on controlling interactions: [I love that I can] turn down the temperature super easily without even having to pull out my phone and [unlock it] and go to the app. I could literally just scroll down. So, it's definitely super convenient. Majority of the time, I don't have my phone on me as I'm working, so this definitely comes in handy.

Attributes of Successful Controlling Interactions

Controlling interactions are most helpful when they are:

Prioritized. The interface needs to prioritize frequently performed functions and remove less common ones. For example, the Spotify smartwatch app prioritizes common actions such as skipping and pausing/playing a song rather than less common ones such as fast-forwarding or viewing the lyrics, which are available in the mobile app.

The Android Spotify interface on the watch and phone. The watch interface is simpler than the phone.
The Spotify smartwatch app prioritizes a few basic functions out of those available in the mobile app.
The MyMazda smartwatch app next to the smartphone app. The watch app is simpler.
The MyMazda smartwatch app prioritizes the start/stop engine function but could also include other common functions such as locking/unlocking the doors in the same interface.

Synchronized. One frustration with controlling interactions is when the smartwatch application isn’t synced with other devices — for example, if the user takes an action on the phone, but that action doesn’t immediately show up on the watch, or if the user begins a process on the watch but must go to the phone to sync information.

The Delta smartwatch app tells the user they must sync with their iPhone.
The watch application lags behind the phone; this lack of synchronization makes the smartwatch app far less useful.

Accessible. Participants made it clear that controlling interactions were particularly useful when their hands were busy. [Controlling my music] is helpful when I am busy with both my hands, like, you know, cleaning, cooking, whatever it may be. [...] This one I find very, very, very helpful. It’s particularly helpful when an application opens automatically and remains available right on the watch face when the user raises their wrist, to provide easy access to important controls while they may have full hands or be occupied with another activity.

Communicating Interactions

Communicating interactions include anything involved in communicating with another person. Communication notifications are among the most important things received on the watch. As one large-scale study by Sahami and his coauthors concluded, “[smartwatch] notifications are perceived as important if they notify about communication with other persons, inform about other persons’ actions, or about events.”

One of the most valuable things that the smartwatch affords is comprehensive awareness of all incoming communications. Smartwatch users seem to rely on the assumption that, if they are wearing their watch, they will not miss incoming communications. Our study participants made it clear that they did not always have their phones nearby and, thus, missed many incoming communications in the moment. However, the watch allowed them to be away from the phone but still respond to, or at least be aware of, all incoming communications so they would not miss something important.

A user is receiving a call on their Pixel watch
A participant commented: I was getting a call while I was in the staff room on my watch. However, my phone was at my desk in my cubicle, which was down the hall. Thankfully, I was able to utilize this and pick up an important phone call that I was literally waiting on to get.

Types of Communicating Interactions

The word messages” below refers to text messages from native message applications, third-party messaging applications, and emails.

Receiving Messages

Users hate missing incoming communications. They consider checking communications on the watch more discreet and polite than pulling out a smartphone. The watch makes the user aware that a message has arrived and gives enough information for them to decide whether to respond. They can also decide whether that response is simple enough to be sent from the watch or whether they need to pull out their phones.

This is one reason why incoming messages should display as much content as possible. As soon as the watch vibrates, users expect to be able to see most, if not all, of the message they have just received. However, it can be challenging for users to go back through multiple messages in one conversation on the watch; if that type of interaction is needed, they will most likely use the phone.

Unfortunately, incoming email notifications — particularly stylized emails from businesses — often waste valuable space with irrelevant information.

The smartwatch notification for an email disproportionately emphasizes info that the smartphone deemphasizes.
The text August 18, 2023 View Online | Sign Up | Shop 10% Off Morning Brew… is visually deemphasized and can easily be ignored on the phone (left). On the watch, however, it occupies more than half the screen. The information displayed is not enough for users to decide whether the email is worth opening (right).

To optimize message notifications for smartwatches:

  1. The Sender should be a short, meaningful name that users will recognize.
  2. Subject lines should be frontloaded with meaningful information.
  3. The first words in the email (which will display below the subject line) should have a strong information scent, indicating the value of the message.

Sending Messages

Typing on the smartwatch is possible, but very inconvenient and difficult. Users avoid it whenever possible by either not responding to a message, dictating what they want to say rather than typing it, or responding on the phone.

A user with long nails shows the tiny keyboard on a Pixel watch
Any form of typing on a smartwatch is tedious and challenging. It is almost guaranteed that users will make mistakes.

The interaction cost of moving to the phone means that, in many cases, users will choose to respond to a message later or not at all. However, when they do decide that a message received on the watch merits an immediate response, they will almost always attempt to dictate rather than type. Dictation is convenient but becomes extremely frustrating when it introduces mistakes, because correcting text on the watch is cumbersome and error-prone. Dictation is even more likely to cause typos when the user has an accent or is in a noisy environment.

Some participants in our study expressed the value of premade response suggestions that can be sent with one tap, but these are obviously not suited to every case. Participants did not seem to condemn the smartwatch for being a difficult input method, but they were all familiar with the challenges it presents.

Calling

Taking a call on the watch can be very convenient if the phone is not nearby or if the user’s hands are busy. In many ways, calling on the watch is equivalent to the speaker functionality on the phone, but even more convenient because the user’s hands remain free. However, sometimes users hesitate to take a call on the watch because they are in a public place and do not want to disturb others.

Guiding Interactions

Guiding interactions provide users with in-the-moment direction during an activity. Examples include workout routines, breathing exercises during meditation, navigation directions, and even a metronome. The user initiates these interactions, and they generally provide guidance for the whole duration of the activity.

Attributes of Successful Guiding Interactions

Guiding interactions are most helpful when they are:

Independent. Once initiated, the guidance on the watch should be self-sufficient and not require the use of another device or any additional interaction from the user. Setting the phone aside and following the independent, reliable directions coming from the wrist makes the watch a useful tool for providing guidance.

An Apple Watch and Pixel Watch interface showing navigational steps
Navigational guidance on the Pixel Watch (left) and Apple Watch (right) both allow users to initiate the guidance and then follow it from their wrist without the need to look at the phone or touch the screen again.

Haptic. Most guiding interactions should utilize haptic feedback to help the user notice what comes next. This is particularly important for two reasons: (1) users often keep their watches on silent mode, and (2) they are often focused on performing some other behavior and, thus, less likely to notice things change on the watch.

A Samsung Galaxy watch metronome application
One participant commented that he preferred the metronome app on his watch to vibrate rather than beep while he played his guitar.

Focused. Because users are likely to be occupied by some task while guidance is provided, the interface should focus the user’s attention on the next step. Providing an overview of all steps or a comprehensive view of information is more confusing than showcasing what comes next.

Guiding interactions can drain the watch battery relatively quickly because they use so many features of the device at once, but this is unlikely to be problematic for most users.

Less Useful Types of Interactions

Interactions other than the 6 types described in this article do exist on smartwatches; however, participants in our study showed no interest in them. Some of those interactions include:

Consuming: The act of passively taking in content on a device — for example, watching a video or movie, looking at photos, reading an article, and so on. These types of interactions are extremely common on both desktop and mobile, but not on tiny smartwatch screens.

Creating: The act of combining inputs to craft or build something original — for example, writing a document, creating a visual, or mixing music. It is hard enough to write simple text messages on the watch that users will not attempt much more than that.

Browsing: The act of casually looking through information with no goal in mind — for example, window shopping. People don’t go to the watch  to look around; they use it to get things done.

Searching: Seeking out specific information on the watch is cumbersome for two reasons: (1) it’s difficult to accurately input the information required to return useful results (i.e., typing or dictating, as mentioned earlier), and (2) the screen is too small to meaningfully display many search results. Whether popular hiking trails or favorite songs, users tend to search on the phone, not the watch.

Unfortunately, many applications built for smartwatches center around these types of less useful interactions. While these apps will receive some downloads because some users might hope to do things more conveniently on the watch, we predict that they will ultimately struggle and see quick usage dropoff because the watch is too small for these behaviors.

Conclusion

Smartwatches have progressed beyond simple step counters — they can facilitate rich communication, control devices around us, and keep us on track in life. The 6 interaction types described in this article reflect progress in wearable-tech design and underscore the importance of making designs well-timed, personally relevant, and extremely simple. Smartwatches have the potential to make many tasks easier, but, just like any other channel, they still have limits.

References

Marta E. Cecchinato, Anna L. Cox, and Jon Bird. 2017. Always On(line)? User Experience of Smartwatches and their Role within Multi-Device Ecologies. Proceedings of the 2017 CHI Conference on Human Factors in Computing Systems (2017), 3557-3568. DOI:https://doi.org/10.1145/3025453.3025538

Alireza Sahami Shirazi, Niels Henze, Tilman Dingler, Martin Pielot, Dominik Weber, and Albrecht Schmidt. 2014. Large-scale assessment of mobile notifications. Proceedings of the SIGCHI Conference on Human Factors in Computing Systems (2014), 3055-3064. DOI:https://doi.org/10.1145/2556288.2557189

Steven Schirra and Frank R. Bentley. 2015. "It’s kind of like an extra screen for my phone" Understanding Everyday Uses of Consumer Smart Watches. Proceedings of the 33rd Annual ACM Conference Extended Abstracts on Human Factors in Computing Systems (2015), 2151-2156. DOI:https://doi.org/10.1145/2702613.2732931