Availability in Onboarding • Elemy (Tilly)

Iterating on a complex task in customer onboarding to balance cognitive load with usable data

While I oversaw the automation of all of family onboarding for our pediatric autism company, this case study focuses on the most challenging piece — collecting the family’s availability. As the designer on this project, I researched, shipped, and iterated on ways to collect family availability.


Tilly was formerly known as Elemy. This project was under the Elemy brand.

Outcomes

💸 Reduce spend on customer support

We were able to run admissions with 20 fewer customer support agents by the end of the year.


📵 Removed need for phone calls

It used to take 25 minutes on a phone call to collect availability for each potential new client.


⏱️ More flexible schedules to optimize

The quantity of available hours to schedule with increased when we introduced my final time bucket solution

Contributions

UX design, concept generation, user interviews

Team

Eng & product leads

Dates

Nov 2022 - Feb 2023

Problem overview


In 2022, Elemy began trying to grow. To do that, we needed to automate more of family onboarding and reduce the cost of bringing in each new family. A tricky part of the existing onboarding phone calls was sorting out the schedule. Autism ABA therapy is unique, because it is typically 25+ hours/week. This is sometimes a surprise to families, and can also lead to funnel churn. Plus, on our end, we want to schedule one therapist with 2 families to create a fulltime workload, but the timing and location has to be just right, with as much as possible during standard business hours. I had a few challenges to tackle in this project:


  1. How might we collect the maximum amount of time a family is available, so we can find a therapist who matches their schedule?

  2. How can we do this as early as possible, so we know whether we need to hire a new therapist to accommodate this family?

  3. How do we filter out families who are not ready to commit to ABA therapy?



My approach

Led from interviews to concept directions

I planned, conducted, and synthesized interviews, culminating in a readout and ideation workshop. Supported by the workshop ideas, I identified 3 potential solutions to explore.



Rapid MVP due to shifting timelines

I had been juggling this project with other projects thinking that it was a bit further down the roadmap. Unexpectedly, my PM asked me to build high-fidelity designs within a week. Armed with insights from the first round of research, I used some high level principles to launch something grounded in user insights, even if it wasn't perfect.



Self-initiated additional testing, improving final solution

Anticipating that there would be no time to test once we inevitably needed to iterate, I took it upon myself to preemptively launch a quick round of user tests to understand whether time buckets would be a better solution in the long run. Ultimately, I was able to apply these learnings quickly to defend and finalize designs for a final solution.


Solution

Based on the data we received from onboarding families & the learnings from my user testing, a new PM and I decided to do another iteration. The availability we were receiving with the MVP solution still was not sufficient to effectively match the families to a therapist; agents were having to call and follow up with 6% of families.

The product manager proposed that we simply update the existing validation logic & errors to increase (and complicate) the minimum availability requirements. I was highly concerned this would result in a clunky, confusing UX early in the onboarding funnel.


I leveled up the project by:

  • Exploring additional low-lift ways to guide the family earlier

  • Getting Product/Ops alignment to adopt the time buckets

  • Getting Engineering alignment to apply our new design system