The Origin of WICster
Participants were interested in the resources PA WIC had but found them inaccessible. Our team believed that proactively providing participants these resources could increase the value and effectiveness of PA WIC’s educational offerings and keep participants in the program longer.
Our initial version of WICster was a “Wizard of Oz,” in other words, we manually sent tips through text messages. We conducted a 4-week longitudinal study to test how to most effectively get participants these resources.
What Makes a Good Tip?
“Tailoring content and functionality to a particular user's need based on his/her characteristics increases the efficacy of the system”
– Alqahtani et al, 2019
The content of our nudges were just as important as the personalization aspects included. The criteria we used when creating “good tips” were to include:
- What health information is being shared
- Why the information is valuable
- How to implement the information
Our user research revealed 2 findings:
- There is a high correlation between personalization and implementation: Participants who received more personalization had a higher average rate of tip implementation.
- There is a high correlation between personalization and user engagement: Participants who received more personalization had a higher average response rate, meaning they engaged more with the tips we sent.
To further personalize WICster's content to caregivers, participants can adjust their health and nutrition topics of interest at any time. WICster also adapts to a family's situational context by changing age-appropriate health information sent to users.
Designing WICster's Personality
We aimed to make the WICster companion supportive and non-judgemental.
We want participants to feel empowered and confident in the health choices they make for their families.
To ensure that our tone and voice is encouraging, we asked participants about the tone of the daily texts and found that most participants found the tone encouraging and informative, as we had hoped.
“It doesn't belittle you or make you feel stupid like, ‘oh duh, you should know this’... It’s kind of like when you get a Snapple, and you look at the top, there's a Snapple fact. It's like, when you open the text message...there's a Snapple fact inside. You know there's going to be some type of good fact that you could actually use or pass on to somebody that you know...”
– PA WIC Participant
The Right Amount of Proactivity
We delivered nutritional resources from WIC pamphlets in two different ways.
Proactivity reduces the effort and burden placed on participants trying to learn
We found that participants didn’t actively seek out WIC’s health and nutrition resources, but they do value and use this information when it’s proactively sent to them.
There is a fine-line between effective nudging and nagging.
Within our daily tip group, we increased the amount of texts that some of them received from 1 to 3 messages (a daily health tip, a mid-day reminder encouraging them to try out the tip, and an end-of-the-day check-in message). To keep metrics consistent, we compared the response rate of just the first nutritional tip text they received.
Engagement dropped heavily when we sent more than one text reminder or request for feedback. We, therefore, included the ability for participants to choose the amount of notifications they receive and most opted for receiving one text per day– that being their nutritional tip.
What Impact Does Personalization Have?
After learning about the many challenges caregivers experience when trying to implement healthy and nutritious choices, we realized that the health advice sent must constantly change to match the situational context of a parent.
But how personalized does it need to be? We sent health tips to 2 groups over a 4-week period. The first group received generalized tips that weren’t as tailored to their health interests while the second group received health tips that were highly personalized and aligned with their selected health goals.
Changing Behavior with WICster
We found that participants were interested in approaches like planning and self-monitoring, but only if they were easy to get started and maintain.
After going through multiple rounds of prototyping and testing, we came up with a number of design principles for our solution to abide by (shown on the right).
“As long as I'm getting recognized for some progress for my kids' life and nutrition, that's all I really would worry about. At least someone can see that I'm doing good things for my child and things that [are] supposed to be better [for] my child.”
– PA WIC Participant
Behavioral Design Principles
Reduce open-ended prompts
Through testing a 5 minute journal in our first prototype, we found that participants preferred something that was quicker and had more structure.
Provide clear guidance and ample praise
The efforts of caregivers is often unrecognized. They really value guidance and support as they try to do what’s best for their families.
Lower barriers to entry
To make it easy for participants to get value out of the service, we automatically opt them into topics and meal planning. We provide opportunities for them to take a more active role in planning and monitoring.
Below are 3 different prototype iterations. By testing these with users, we were able to develop the principles stated above.
Prototype Testing: Self-Monitoring and Planning
Three iterations of our prototype: our first iteration(left) testing journaling and and open-ended chat, our second iteration(middle) testing self-monitoring, praise, and planning, and our final prototype(right) which automatically suggests recipes based on the participants' benefits.
In order to address the technical feasibility of this solution, we created an automated version of the SMS daily tips and started running it with participants.
We created a database of over 100 tips and used a number of technologies to enable the personalized tips, and goal check-in through SMS. This allowed us to validate that the automated version can be as valuable as the initial ‘Wizard-of-Oz’ prototype.
Participants stayed enrolled in the program
We allowed participants we interviewed to opt-in to the service and started sending tips in the middle of July. We let them choose the frequency they’d like to receive tips.
Over 3 weeks of running the prototype, we sent 217 tips and none of the 12 users opted out of the service.
Our final prototype deliverable to PA WIC consisted of both a coded, functional SMS chatbot and an app prototype. The chatbot provides a more near-term, limited, version of WICster while the app prototype demonstrates how our solution could extend beyond SMS and integrate more closely PA WIC’s services.
The functional SMS chatbot sends enrolled users daily tips, tip check-ins and goal reminders to help them reach their nutritional goals. It also allows new users to onboard completely through SMS and existing users can change their preference through the menu options. Below are all the flows included in our final prototype.
The mobile app prototype expands on the SMS chatbot to give users even greater agency over reaching their family’s health goals by incorporating more behavior change elements. It also shows how WICster could integrate with the user’s WIC profile to allow for an all-in-one PA WIC app.
Want to see more?
Check out our final design!