The objective of this research study¹ was to see if mHealth apps followed best design practices for notifications, engaging users in health promoting activities toward meeting their health goals.

I hoped to get some insights into what, if any, content design considerations I should keep in mind when writing these types of components. These were my thoughts after reading the study.

Key finding take-aways

Notification frequency

“For notifications, prior work suggests that more is not necessarily better and to consider the user’s readiness for change when determining frequency, such as increasing notifications if the user is more willing to change. Most of the notifications we received were sent multiple times a day and none of the apps we surveyed considered the user’s level of inclination to change.”

Even though this finding is not about content per se, I want to call it out anyway.

I think the idea of a user’s level of inclination to change can be powerful when developing a notification strategy to inspire and support users. And to some extent I think it is being done.

The onboarding flow in many mHealth apps include some get-to-know-you questions to tailor the way a user experiences the app. Often this tailoring extends to notification messaging. This notification strategy may not always be explicitly rooted in the framework of the Transtheoretical model (aka Stages of Change)², but it reflects the model nevertheless.

I think this could be a space for more pointed collaboration between health education professionals and UX writers.

Content purpose

“The majority of notifications (89.6%) were aimed at reminding the user to do something. The next more common notification purposes included providing insight (6.6%), motivation (4.4%), user reflection (4.1%), and tips (3.2%). … Although the majority of the notifications were reminders, prior studies have found that reminders hinder the habit development since the user becomes reliant on the notifications. Also, prior work states that notification content should be varied to keep users engaged.”

First, it’s good to know what research indicates: reminder notifications can hinder rather than help habit development. But also, I think it depends on the purpose of the app and the intended audience. Habit formation won’t mean the same thing to someone who has memory loss as with someone who has a strong mental acuity or with ADHD. Something to keep in mind for sure. I think it’s worth really thinking about when this type of messaging can be downplayed and others can be leveraged.

Second, the idea that variety keeps users engaged. When they become too routine, they can also become ambient, barely registering in our field of attention. Reminders can easily fall into this pattern. “Time to do X!” repeatedly stops having any gravity. A simple solution might be to have an array of a few simple but varied reminder statements. Another might be including some sense of accomplishment: “You’ve done X for 28 days in a row. Keep it up!” Still, I could see both eventually falling into the same attention void. So, again, playing up those other types of messaging can have an impact in ongoing engagement.

Goal type

“Our results show that 93.4% of the notifications targeted short-term goals, 5.1% did not contain a goal, and only 1.5% targeted long-term goals. Even though prior work has found that setting short-term goals can be more motivating, we do not know of any recommendations on whether to include short-term or long-term goals in mHealth app notification.”

My thoughts here relate back to the researcher’s findings on notification frequency. This idea might play into the idea of a user’s inclination for change. I could see the case that reminding users of their short-term goals, especially if they are at an earlier stage in adopting a new behavior, can be motivating. Taking one small step at a time until they’ve built momentum toward the goal. If they are at a later stage, established routine or maintenance level of a skill, maybe a message about their long-term goal might be more reaffirming of their ongoing practice.

Or perhaps the pattern oscillates depending on where the user is within each stage as they progress…

Interactivity

“Including interactive elements has been found to be important for mHealth notifications in increasing user engagement and the likelihood of notifications being attended to. However, the majority of the notifications in our survey were not interactive.”

In this study, only ~41% of notifications included interactive elements. ~28% of those were acknowledgement actions (i.e. tapping “OK”) and ~1% included a link to another page or application.

It’s also worth mentioning that ~76% of all notifications received were push notifications. I wondered if they considered the push notification itself (tap to open the app (to a relevant screen, in especially good design)) as an interactive element.

There’s a time and place to ask the user for extra interaction, and the user’s context matters. I think there’s got to be a threshold where more interactivity can draw better engagement, until it reaches a point where it just becomes a nuisance and the user just turns the notifications off. I would agree that there could be room for more interactivity in notifications, just not in every case. There has to be a really good reason to include it or a really good reason to not include it.

Tailoring

“Tailoring notifications has been shown to increase engagement and persuasively motivate users.
…
Our data shows that only 32.6% of the notifications included tailoring. User data was used in 25.2% of notifications, whereas 10.1% of the notifications included the user’s name.”

Again, I’m sure there’s an optimal threshold or range here.

I also think this is compounded by lumping the different types of notifications together. Push notifications, for example, should tread carefully (probably conservatively) in including personal information in notification messages.

I another compounding factor is space. Devices can get small, making every character count. Each bit of copy should have purpose, hold weight. So there is a fine line to walk here between space and priority of each piece of a message. I think this is where seeing the big picture of the content strategy as a whole is important for that give-and-take in the granular spaces the components occupy.

Again, I could buy into that there is room for more tailoring, but there has to be a really good reason for why it matters for the user to receive it in this way at this time. As a UX writer, it’s decisions like these that I need to be able determine and articulate back to a product team.

Personalization

“Although prior work has found that people rarely change default notification settings, we are not aware of any recommendations on including defaults or manual notification settings, and whether they affect user engagement and behavioral change.”

My main take-away here is that the ultimately, the user is in control (as they should be!). So every notification should count. They need to be designed as helpful and unobtrusive as possible so that a user not only keeps them on, but continually finds value in them.

Final thoughts

Overall, it was an interesting article, both as a research study, and as a summation of relevant studies on this topic. Like an absolute nerd, I circled several citations that I’d be interested to dive deeper into.

I was a little disappointed that there wasn’t more of a tie between best design practices and health behavior models. Nevertheless, there were several references that indicate that there could be a benefit to having a health researcher on or in collaboration with a UX design team.

If you’re interested, you can read a pdf version of this article online.

References

1. Woodward, J. B., Chen, Y., Jurczyk, K., Ross, K., Anthony, L., & Ruiz, J. (2021). A Survey of Notification Designs in Commercial mHealth Apps. CHI EA ’21: Extended Abstracts of the 2021 CHI Conference on Human Factors in Computing Systems. https://doi.org/10.1145/3411763.3451639
2. Prochaska, J. O., & DiClemente, C. C. (2005). The transtheoretical approach. In Oxford University Press eBooks (pp. 147–171). https://doi.org/10.1093/med:psych/9780195165791.003.0007

---

Interested in partnering? Let’s talk to see how I can help you create a comfortable digital health experience. Drop me a note at evy.haan@gmail.com.

Photo by James Orr on Unsplash.