Kinderspot
Problem: Military members and their families need an easy and trusted way to coordinate childcare during temporary duty assignments
Role: Research, Interaction Design, Usability Testing, Visual Design
Result:
144% adoption increase in 2 years
(Launched Q3 2021 at 9 US-based Child Development Centers; is now offered at all Air Force CDCs worldwide)
5-star rating on FB and Apple app store, more than 1,000 downloads on Google Play
Overview
Our mission with Kinderspot was to bring to life Maj Jacque Vasta’s vision of an app that would help military members and their families coordinate childcare by subleasing currently held slots at the DoD Child Development Centers. Maj Vasta took first place at the 2020 AFIMSC Innovation Rodeo for her idea and a SBIR contract was awarded to a digital agency based in Washington DC. I worked with two other designers over the course of 16 weeks.
Research
Our work on Kinderspot began with sending out a survey via Google Forms. We received 580 responses from patrons of the Air Force Child Development Centers, and 70 responses from employees. The surveys served as a jumping-off point for us to conduct interviews with both patrons and employees of the CDCs, of which we completed 18 and 6, respectively. The biggest frustrations were: finding a person to sublet/lease and not having a trusted, sanctioned way to exchange payment for such a transaction. The CDCs were mostly hands-off in these areas, leaving patrons to use social media or Craigslist to find and facilitate these arrangements.
Task-Based Personas
Two of the four task-based personas we identified: TDY Tawny and TDY Tina. Tawny holds a permanent spot at a Child Development Center and is looking to sublet it during an upcoming temporary duty assignment. Tina is on the other end of this transaction, trying to find a spot to rent during her own temporary duty assignment. Important to note, it is often the case that a single user would be represented by both personas, as they could be looking to sublet their spot at their home location, as well as find a spot near their TDY location.
Journey Maps
How we came to understand the current, pre-Kinderspot experiences of leasing your permanent spot for rent versus finding a spot to rent for a short period, while also moving and changing jobs.
First Iteration
The biggest challenge with Kinderspot was that the Child Development Centers would only allow booking of complete weeks (M-F). We needed a clear way to convey this to users since it is not a common feature of a scheduling interaction in most products. Early iterations had a note to that effect and required the user to just select their start and end dates, with the app then asking whether they’d like to include the incomplete weeks.
We soon evolved to a two-screen process: the user enters their start and end dates on one screen and on the next, they’re shown which bases could accommodate all or part of that request. On the second page, they could learn more about the CDCs and select, via radio buttons, which complete weeks they wanted to lease/let. This would later change to not include the radio buttons but have the weeks offered outlined in the selection green and function as push buttons.
Usability Testing Round 1
For this first round of usability testing, we were mostly interested in knowing how desirable Kinderspot was and if users would understand how to select their rental period. We gave participants two different prototypes for the spot renter and spot owner and had them go through just one scenario for each. We tested this on seven participants, with the first five doing task completion, and the other two doing a cognitive walkthrough, describing the design in detail – what they expected to be able to do with each screen, how they would do it, and what they would expect from each element on the screen. Participants were drawn from the pool of interviewees who said they’d be interested in helping us later on.
The contracted team I was working with had never done usability testing before and was somewhat surprised by the notion of testing something before the code is written. They were initially unsure it was even necessary because the design made sense to them. I finally convinced them to let me run one round of moderated testing with just five participants and if they didn’t see the value, I wouldn’t push for it again. We started by doing a guerrilla test with a team member who had just been added to our team, just to make sure the test would work on “real” participants. The test went well until it was time to select weeks to rent/sublet, which our participant found confusing. This was an ‘aha’ moment for the team: if someone from our own team is having trouble, how will outsiders fare? By the time the session was over, they were all convinced of the value of usability testing, so much so that they began making notes about what needed to be tested in subsequent iterations. I counted this as a major win, not just for the project, but for helping other designers see the importance of testing.
Our guerrilla testing session gave us additional questions to pay attention to and probe deeper as necessary. Participants were overall very excited about the prospect of this becoming a live product, and many echoed the same confusion and frustration from our guerrilla session, which told us we had some work to do to make the app more intuitive.
“Love the concept, the design is great! Really excited for this to be released!”
Second Iteration
In addition to making the date selection mechanism more intuitive, we also set to work on the onboarding process, so users could establish their household profile, select a payment method, and we started working on an “explore” flow, allowing users who might want to take a vacation sometime in the future to look at what might be a feasible time, based on availability (or demand, if subletting). We decided also to start using Google’s Material Design System. For one, our early iteration already bore a resemblance to it. For two, more importantly, we needed this app to feel familiar and give users a sense that it was mature and legitimate and that they could trust it, both to arrange care for their children and to make and receive payments.
Usability Testing Round 2
To find participants for the second round of usability testing, we arranged to have these flyers posted at CDCs. We were able to net eight participants, with them again broken into two groups: 5 doing task completion, and 3 doing cognitive walkthrough.
Feedback was, again, generally positive but there was some confusion about pricing/payment, whose spot they were taking, and where the paperwork lived. Users also asked about the following functionality:
Clearer instructions around password requirements
Improved search functionality when browsing a longer list of CDCs
More information about the CDCs.
Better support for multiple child rentals
Alerts when specific offers are available
“It looks nice and I think it will be super useful once further developed...Keep up the great work. We appreciate it.”
Third Iteration and Visual Design
Following usability testing, we got to work implementing the features our participants felt were missing, many of which were already in the works but not ready for testing.
Our design system, based on Google Material, had grown and been refined from our first iteration to include different styles for mobile, calendar components, and a desktop grid.
There was also a significant effort made toward getting the CDC admin portal ready to launch concurrently with the mobile app, so all parties would stay in sync. I do not have sufficient artifacts to tell that story.
A/B Testing – Create an Account
The final round of testing was to A/B test two different flows for account creation. The main difference between these two flows was how much information would need to be entered on a single screen. Option A was a single screen that would expand to fit each new piece of information, whereas option B was broken up into multiple screens: username/password > user information > child information > household income. Over 42 unmoderated sessions, option B was the clear favorite with 28 votes versus 14 for option A. Some users voiced concern over having to use their military email address and would prefer to also include their personal email.
Previous: