Goal

Reduce cancellations by 1%-4.5%, or 2800 to 8000 shifts annually.

User Problem That We’re Solving

HCPs frequently late cancel/NCNS because of car issues: their car broke down, their ride didn’t show, etc. This causes an undue burden on our users; they have to either cancel their shift and face the consequences or pay out of pocket to get to the facility.

How are we solving the problem?

Let’s subsidize the cost of a Lyft or Uber in-app if an HCP is cancelling due to a transportation-related issue. I envision two ways to do this.

• We provide a 100% direct subsidy of the cost and keep track of how many times an HCP has used the subsidization.
  • Pros: This service comes at no cost to the user and is something that Wei did early on to help guarantee our first shifts were going to be worked.
  • Cons: This is a pretty gameable solution, and one that eats at our margin quite significantly (The average shift provides $72 in revenue and rideshare costs could easily eclipse that). We would have to institute some sort of cap or policy to ensure that HCPs don’t use this for every shift.
• We provide a pay advance to subsidize the rideshare cost; if the HCP receives the payment but still fails to show up, we suspend their account.
  • Pros: Our margin is immune from this solution, and there are built-in balances to ensure an HCP tries to get to the shift themselves before resorting to this option. An HCP can use this solution as many times as they want, and we’ll be able to automate the whole process (instead of requiring policy).
  • Cons: Taking on of a policy like this might not be as high. HCPs probably don’t want to spend out of their earnings. There is some risk to a pay advance if an HCP still doesn’t show up to their shift.

Investment Hypothesis

HCPs late cancel because of transportation issues and are often in a financial situation that does not allow them to advance money out of pocket to get to a shift. We create a solution to help them get to their shift using existing transportation infrastructure. We’ll reduce the number of cancellations, thus improving the overall experience on both sides of our marketplace.

When HCFs get a shift booked through CBH, they’ll be more confident that it will be worked. When HCPs have a car-related life issue, they won’t have to go through our policy team to ensure they can continue to work, and we’ll improve their overall experience on our platform.

Summary & Context

About 2000 of a total of 45000 cancellations in the last 3 months were due to car issues (4.5%). HCFs also report unreliability in HCPs as a primary issue they have with using our platform to get their shifts filled. We also have anecdotal evidence from Wei’s experience founding this company that helping to pay for transportation costs improves the chances of an HCP working a shift.

What’s Next

• First, I’ll find out which MSAs might benefit the most from this kind of solution (i.e. Which MSAs have the highest share of cancellations due to transportation issues). We’ll run an experiment in those MSAs using a 100% transportation subsidy for cancellations due to transport failure.
• Experiment details:
  • Next, we can scrappily add an intervention to alert us when HCPs are cancelling/late cancelling due to a transportation failure.
  • We’ll send the HCP a message that we will pay for them to get to their shift using Lyft or Uber.
  • We’ll ask them to send a screenshot of their ride details (with date and cost) and use Textract to identify how much we should pay them, then automate a payout to their Stripe account if the following conditions are met:
    • The address matches the address of the facility they’re going to.
    • The date matches the date of their shift.
    • They have only used this service once in the last X days (to avoid cheating).
  • We run this experiment for 30 days to see how it impacts the cancellation rate in the MSA(s). I think I could execute on all of this by myself but would need to do some more planning to be sure it wouldn’t require additional resources.
• If the experiment is successful (i.e. it does actually reduce the cancellation rate) we productize this by investing product and engineering resources. I would probably hand off to a PM at this point, or borrow engineers.
  • We would choose a direction to go in regarding direct subsidies versus payment advances.
  • When the HCP attempts to cancel we throw a pop-up saying that we can help provide rideshare transportation (copy would be different for either solution).
  • If an HCP clicks “yes, help me”, we configure a payout using Textract + Uber/Lyft receipt.